GEO Environment and Landscape Evolution II PDF

Geo-Environment and
Landscape Evolution II
Evolution, Monitoring, Simulation, Management and
Remediation of the Geological Environment and
Landscape
WIT Press publishes leading books in Science and Technology.

Visit our website for the current list of titles.
www.witpress.com
WITeLibrary
Home of the Transactions of the Wessex Institute.
Papers presented at Geo-Environment and Landscape Evolution 2006 are archived in the
WIT elibrary in volume 89 of WIT Transactions on Ecology and the Environment
(ISSN 1743-3541).
The WIT electronic-library provides the international scientific community with
immediate and permanent access to individual papers presented at WIT conferences.
http://library.witpress.com
Transactions Editor
Carlos Brebbia
Wessex Institute of Technology
Ashurst Lodge, Ashurst
Southampton SO40 7AA, UK
Email: carlos@wessex.ac.uk
WIT Transactions on Ecology and the Environment

Editorial Board
Y N Abousleiman A Aldama
University of Oklahoma IMTA
USA Mexico
D Almorza Gomar A M Amer

University of Cadiz Cairo University
Spain Egypt
M Andretta J M Baldasano
Montecatini Universitat Politecnica de
Italy Catalunya
Spain
J G Bartzis A Bejan
Institute of Nuclear Duke University
Technology USA
Greece
J Boarder B Bobee
Cartref Consulting Institut National de la
Systems Recherche Scientifique
UK Canada
H Boileau C A Borrego
ESIGEC University of Aveiro
France Portugal
A H-D Cheng C-L Chiu
University of Mississippi University of Pittsburgh
USA USA
A Cieslak W Czyczula
Technical University of Krakow University of
Lodz Technology
Poland Poland
M da Conceicao Cunha M Davis
University of Coimbra Temple University
Portugal USA
A B de Almeida K Dorow
Instituto Superior Tecnico Pacific Northwest National
Portugal Laboratory
USA
C Dowlen R Duffell
South Bank University University of Hertfordshire
UK UK
J P du Plessis A Ebel
University of Stellenbosch University of Cologne
South Africa Germany
D Elms D M Elsom
University of Canterbury Oxford Brookes University
New Zealand UK
D Emmanouloudis J W Everett
Technological Educational Rowan University
Institute of Kavala USA
Greece
R A Falconer D M Fraser
Cardiff University University of Cape Town
UK South Africa
G Gambolati N Georgantzis
Universita di Padova Universitat Jaume I
Italy Spain
F Gomez K G Goulias
Universidad Politecnica Pennsylvania State
de Valencia University
Spain USA
W E Grant C Hanke
Texas A & M University Danish Technical University
USA Denmark
A H Hendrickx S Heslop
Free University of University of Bristol
Brussels UK
Belgium
I Hideaki W F Huebner
Nagoya University Southwest Research Institute
Japan USA
W Hutchinson D Kaliampakos
Edith Cowan University National Technical
Australia University of Athens
Greece
K L Katsifarakis H Kawashima
Aristotle University of The University of Tokyo
Thessaloniki Japan
Greece
B A Kazimee D Kirkland
Washington State Nicholas Grimshaw &
University Partners Ltd
USA UK
D Koga J G Kretzschmar
Saga University VITO
Japan Belgium
B S Larsen A Lebedev
Technical University of Moscow State University
Denmark Russia
Denmark
D Lewis K-C Lin
Mississippi State University of New
University Brunswick
USA Canada
J W S Longhurst T Lyons
University of the West of Murdoch University
England Australia
UK
U Mander N Marchettini
University of Tartu University of Siena
Estonia Italy
J D M Marsh J F Martin-Duque
Griffith University Universidad Complutense
Australia Spain
K McManis C A Mitchell
University of New The University of Sydney
Orleans Australia
USA
M B Neace R Olsen
Mercer University Camp Dresser & McKee Inc.
USA USA
R O'Neill K Onishi
Oak Ridge National Ibaraki University
Laboratory Japan
USA
J Park G Passerini
Seoul National University Universita delle Marche
Korea Italy
B C Patten M F Platzer
University of Georgia Naval Postgraduate School
USA USA
V Popov H Power
Wessex Institute of University of Nottingham
Technology UK
UK
M R I Purvis Y A Pykh
University of Portsmouth Russian Academy of
UK Sciences
Russia
A D Rey A C Rodrigues
McGill University Universidade Nova de
Canada Lisboa
Portugal
R Rosset J L Rubio
Laboratoire d'Aerologie Centro de Investigaciones
France sobre Desertificacion
Spain
S G Saad R San Jose
American University in Technical University of
Cairo Madrid
Egypt Spain
J J Sharp H Sozer
Memorial University of Illinois Institute of
Newfoundland Technology
Canada USA
I V Stangeeva E Tiezzi
St Petersburg University University of Siena
Russia Italy
T Tirabassi S G Tushinski
Institute FISBAT-CNR Moscow State University
Italy Russia
J-L Uso R van Duin

Universitat Jaume I Delft University of
Spain Technology
Netherlands
A Viguri Y Villacampa Esteve
Universitat Jaume I Universidad de Alicante
Spain Spain
G Walters
University of Exeter
UK
SECOND INTERNATIONAL CONFERENCE ON EVOLUTION,
MONITORING, SIMULATION, MANAGEMENT AND
REMEDIATION OF THE GEOLOGICAL ENVIRONMENT AND
LANDSCAPE
Geo-Environment & Landscape Evolution 2006

CONFERENCE CHAIRMEN
J. F. Martín-Duque
Complutense University, Spain
C. A. Brebbia
Wessex Institute of Technology, UK
D. E. Emmanouloudis
Technological Educational Institute of Kavala, Greece
U. Mander
University of Tartu, Estonia
INTERNATIONAL SCIENTIFIC ADVISORY COMMITTEE

A. Cecioni
M. E. Conti
W. S. Fyfe
A. E. Godfrey
G. Lorenzini
C. Miller
M. Noormets
J. Pedraza
Organised by
Sponsored by
Municipality of Rhodes, Greece
Technological Education Institute of Kavala, Greece
WIT Transactions on Ecology and The Environment
This page intentionally left blank
Geo-Environment and
Landscape Evolution II
Editors
J. F. Martín-Duque
C. A. Brebbia
D. E. Emmanouloudis
Technological Educational Institute of Kavala, Greece
U. Mander
University of Tartu, Estonia
J. F. Martín-Duque C. A. Brebbia
Complutense University, Spain Wessex Institute of Technology, UK
D. E. Emmanouloudis U. Mander
Technological Educational Institute of Kavala, University of Tartu, Estonia
Greece
Published by
WIT Press
Ashurst Lodge, Ashurst, Southampton, SO40 7AA, UK
Tel: 44 (0) 238 029 3223; Fax: 44 (0) 238 029 2853
E-Mail: witpress@witpress.com
http://www.witpress.com
For USA, Canada and Mexico
Computational Mechanics Inc

25 Bridge Street, Billerica, MA 01821, USA
Tel: 978 667 5841; Fax: 978 667 7582
E-Mail: infousa@witpress.com
http://www.witpress.com
British Library Cataloguing-in-Publication Data
A Catalogue record for this book is available

from the British Library
ISBN: 1-84564-168-X
ISSN: 1746-448X (print)
ISSN: 1743-3541 (on-line)
The texts of the papers in this volume were set

individually by the authors or under their supervision.
Only minor corrections to the text may have been carried
out by the publisher.
No responsibility is assumed by the Publisher, the Editors and Authors for any injury
and/or damage to persons or property as a matter of products liability, negligence or
otherwise, or from any use or operation of any methods, products, instructions or ideas
contained in the material herein.
© WIT Press 2006
Printed in Great Britain by Cambridge Printing.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted in any form or by any means, electronic, mechanical, photocopying,
recording, or otherwise, without the prior written permission of the Publisher.
Preface
This book contains papers presented at the Second International Conference on
Monitoring, Simulation, Management, and Remediation of the Geological
Environment and Landscape, called Geo-Environment and Landscape Evolution,
held in Rhodes, Greece, in June of 2006.
The success of the first Geo-Environment conference held in Segovia, Spain
(2004), prompted the organizers to reconvene a new meeting in Rhodes in 2006. The
basic reason for this second Geo-Environment Conference is the perception that
the information that all the geosciences can bring to environmental management is
still being underutilized by politicians and developers. We strongly believe that a
through understanding of all the elements of natural systems and processes is
necessary for a proper use of the land. To get a broader perspective on this issue,
other multidisciplinary studies in the fields of geoecology, earth surface processes,
and landscape ecology are also included by the addition of ‘landscape evolution’ to
the Conference name.
In this regard, we believe that the group of geosciences is the most complete of
the physical sciences. This is because to be able to practice them properly, one
must have a mastery of mathematics, physics, chemistry, and biology. In addition,
they require an understanding of time, measured in terms of thousands and millions
of years. It is the tools provided by these areas of knowledge that a geoscientist
brings together to study the processes, rates and durations that form the landscapes
and conditions that present opportunities and hazards to society. From a pragmatic
point of view, environmental geo-science is simply basic earth sciences applied to a
process or condition that society perceives important. On the other hand, there
have been innumerable geologic and landscape studies that started out as intellectual
delights, but later turned into valuable information for humanity’s use of the land in
a sustainable manner.
Based on these principles, this book contains important contributions by
geologists, geographers, soil scientists, ecologists, engineers, planners, and other
professionals interested in the contribution of geo-information to sustainable
development or suitable use of the land.
The editors of this book are grateful to all the authors for their outstanding
contributions. We especially would like to thank all the distinguished scientists
who supported this effort by serving on the International Scientific Advisory
Committee, by promoting the conference and reviewing the submitted abstracts
and papers. We thank the Complutense University of Madrid, Spain, for supporting
the conference, through the collaboration of its Faculty of Geology. We are also
indebted to the Municipality of Rhodes and the Technical Institute of Kavala for
their support as well as the University of Tartu in Estonia. The Wessex Institute of
Technology and WIT Press coordinated all the work, organizing the meeting and
producing this book to their usual high standards.
The Editors
Rhodes, 2006
Contents
Keynote address
Towards a truly sustainable future and a high quality of life for all
people on our planet: the role of the Earth Sciences
W. S. Fyfe ..............................................................................................................1
Section 1: Environmental planning and management
Ethics, geological risks, politics and society

A. Cecioni & V. Pineda .........................................................................................7
Environmental impact assessment and environmental management

plans: an example of an integrated process from the UK
M. A. Broderick & B. Durning............................................................................15
Geological factors in monitoring and planning nature trails at

tourist centres in northern Finland
K. Lehtinen & P. Sarala......................................................................................25
The acoustical improvement plan as a process to re-establish

acceptable acoustical conditions
M. Clerico & G. Soffredini..................................................................................33
Estimation of the on-farm-costs of soil erosion in

Sleman, Indonesia
A. Möller & U. Ranke .........................................................................................43
Section 2: Environmental modelling and monitoring
Large scale soil erosion modeling for a mountainous watershed

P. Zhou, J. Nieminen, T. Tokola, O. Luukkanen & T. Oliver..............................55
Decreased deposition of sulphate and the responses in soilwater at
Estonian integrated monitoring sites 1995–2004
J. Frey, K. Pajuste, K. Treier, Ü. Mander, P. Kask & T. Frey ...........................69
SAKWeb© – Spatial Autocorrelation and Kriging Web Service

J. Negreiros & M. Painho ...................................................................................79
Section 3: Environmental pollution and remediation
After treatment of landfill leachate in peat filters

M. Kõiv, M. Kriipsalu & Ü. Mander...................................................................93
Studies on the distribution of heavy metal Cd in contaminated soils

of various particle sizes and removal efficiencies of heavy metal
using acid washing
S. F. Cheng, C. Y. Huang & L. S. Hsiao ...........................................................105
Mitigation of alkaline mine drainage in a natural wetland system

J. Kelly, P. Champagne & F. Michel ................................................................115
A sequential aerated peat biofilter system for the treatment of

landfill leachate
P. Champagne & M. Khalekuzzaman ...............................................................125
Are crop and range lands being contaminated with cadmium and
lead in sediments transported by wind from an adjacent
contaminated shallow lake?
H. O. Rubio, T. R. Saucedo, M. R. Bautista, K. Wood, C. Holguin
& J. Jiménez .....................................................................................................135
The sorption characteristics of PAHs onto soils in the presence of

synthetic and bio surfactant
J.-H. Chang, Z. Qiang & C.-P. Huang .............................................................143
Carbon dioxide sequestration in coal: implications for CO2 disposal

and CH4 displacement from coal seams
M. Mirzaeian & P. J. Hall.................................................................................151
Section 4: Climatological processes
Response of dissolved organic carbon in a shallow groundwater

ecosystem to a simulated global warming experiment
K. P. Wilson & D. D. Williams .........................................................................163
Regional analysis of climate and bioclimate change in South Italy
A. Capra, P. Porto & B. Scicolone ...................................................................175
Section 5: Geo-environment in urban settings
The urban geo-science model: an essential tool to support planning

and sustainable development
D. McC. Bridge, B. L. Morris & J. R. A. Giles .................................................187
Developing design-oriented strategies to combat regional scale

climate change
B. Stone, Jr ........................................................................................................195
Land subsidence evolution and controlling mechanisms near

Mexico City: environmental planning and management
M. A. Ortega-Guerrero .....................................................................................205
Stability of slopes of municipal solid waste landfills with

co-disposal of biosolids
M. Chopra, D. Reinhart, M. Vajirkar & B. Koodhathinkal ..............................215
Section 6: Geoecology
Net photosynthetic O2 evolution and calcium precipitation in

Chlamydomonas reinhardtii
Y. Y. Wu, P. P. Li, B. L. Wang & C. Q. Liu.......................................................225
Dynamics of suspended sediment concentration and the impact on

a lake-inhabiting bivalve (Corbicula japonica) in the
Abashiri River basin, Hokkaido, northern Japan
S. Yanai, Y. Nishihama & R. Tamura ................................................................231
Potential of poplar plantation for enhancing Polish farm

sustainability
A. Czarnecki & A. Lewandowska-Czarnecka ...................................................241
Section 7: Hydrological studies
Modified quantitative estimation model of erosion and degradation

in three mountainous watersheds
D. Emmanouloudis & M. Kaikis .......................................................................253
Bathymetric curve (75 years old) validation, using the soil
erosion transportation at Cuitzeo Lake Watershed
J. Lafragua, A. Gutiérrez, A. Bahena, G. Moriel & S. Férnández....................263
Efficient watershed modeling using a multi-site weather generator

for meteorological data
M. Khalili, R. Leconte & F. Brissette................................................................273
Hydrological modelling for river basin management in a highly

hydro-geological conditioned environment
D. Guida, A. Longobardi & P. Villani ..............................................................283
Hydrological modelling of snow cover in the large upper Po river

basin: winter 2004 results and validation with snow cover
estimation from satellite
D. Rabuffetti, A. Salandin & R. Cremonini.......................................................293
Section 8: Landscape analysis
Spatial correlograms and landscape metrics as indicators of land

use changes
R. Aunap, E. Uuemaa, J. Roosaare & Ü. Mander ............................................305
The role of geosciences and landscape in the management of

Natural Parks of Guadalajara (Central Spain): in search of
suitable applications
A. García-Quintana, M. P. Abad, M. Aguilar, L. Alcalá, I. Barrera,
M. Cebrián, M. C. Fernández de Villalta, J. F. García-Hidalgo,
R. Giménez, A. E. Godfrey, J. A. González-Martín, A. Lucía,
J. F. Martín-Duque, M. Martín-Loeches, E. Quijada,
J. M. Rodríguez-Borreguero, R. Ruiz López de la Cova & A. Solís .................317
Exploring the effect of demographic elements on the evaluation

of the scenic beauty of various landforms – preliminary results
A. Tsouchlaraki .................................................................................................329
An approach to the landscape analysis

B. Badiani..........................................................................................................341
Evaluation and analysis criteria of the environmental risk factor

of the anthropic perturbation in the infrastructure works
G. Gecchele & G. Pizzo ....................................................................................349
Section 9: Natural hazards and risks
M3 (Monitoring, Management and Mapping) – tool for the solution

of the conflict: “people and environment”
Z. Boukalová, V. Beneš & P. Kořán..................................................................361
Gas hazard: an often neglected natural risk in volcanic areas

W. D’Alessandro ...............................................................................................369
Section 10: Remote sensing
Identification of As-bearing minerals associated with mine wastes

from former metalliferous mines in France using laboratory
reflectance spectra
V. Carrère .........................................................................................................381
Rapid mapping with remote sensing data during flooding 2005 in

Switzerland by object-based methods: a case study
Y. A. Buehler, T. W. Kellenberger, D. Small & K. I. Itten ................................391
Using spatial technology for analyzing disturbed areas and potential

site selection in Chihuahua, Mexico
V. M. Tena, A. C. Pinedo, A. H. Rubio, P. de L. G. Barragán,
A. A. Pinedo, M. V. Hernandez & C. Velez .......................................................401
Section 11: Soil and rock properties
The pedoecologial conditions of natural and opencast peat fields

in Estonia
M. Noormets, T. Köster, T. Tõnutare, K. Kauer, R. Kõlli, T. Paal
& M. Oder .........................................................................................................413
Monitoring programme for underground rock characterization

facility
K. Lehto & J. Lahdenperä.................................................................................423
Geotechnical evaluation of Stabilized Dredged Material (SDM)

from the New York/New Jersey Harbor
A. Maher, A. Sarmad & M. Jafari.....................................................................431
The impact of soil quality on cocoa yield in Nigeria

O. A. Amusan & F. O. Amusan .........................................................................443
Section 12: Vulnerability studies
Influence of ground water extraction in the seismic hazard of

Mexico City
J. Avilés, L. E. Pérez-Rocha & H. R. Aguilar ...................................................457
Predicting favourable areas for landsliding through GIS modelling

in Aparados da Serra (Brazil)
A. J. Strieder, S. A. Buffon, T. F. P. de Quadros & H. R. Oliveira...................467
Author Index ...................................................................................................477

Geo-Environment and Landscape Evolution II 1
Keynote Address
Towards a truly sustainable future and

a high quality of life for all people on our
planet: the role of the Earth Sciences
W. S. Fyfe
Department of Earth Sciences, University of Western Ontario,
Canada
Abstract
We are living in a new period of earth history. I was impressed by the writing of
Vladimir Vernadsky, 1929, in his book The Biosphere. He said we were in a
transition to a new geological era which he termed the psychozoic era. As he
stated, mankind as a whole would become a new and powerful geological entity
able to transform the planet. I very much liked the words of Sir Crispin Tickell
in his British Association lecture of 1993 “I was recently asked if I was an
optimist or pessimist. The best answer was given by someone else. He said that
he had optimism of the intellect but pessimism of the will. In short, we have
most of the means for coping with the problems we face, but are distinctly short
on our readiness to use them. It is never easy to bring the long term into the
short term. Our leaders, whether in politics or business, rarely have a time
horizon of more than five years”. I have been fortunate that I have walked in
over 60 nations. When you walk you see, smell, the problems and you meet
people of all types.
Keywords: the Psychozoic era, education needed for all people, diversity, energy
and waste reduction, soil quality.
WIT Transactions on Ecology and the Environment, Vol 89, © 2006 WIT Press
www.witpress.com, ISSN 1743-3541 (on-line)
doi:10.2495/GEO060011
2 Geo-Environment and Landscape Evolution II
1 What are the great world problems today?

When we look at many regions of our planet there is no question we face the
problem of too many people. An excellent data source is provided by the
publication by The Economist [1], pocket world in figures, which is published
every two years. I am looking at the edition for 2003. For example, we see that
the average female in Niger has eight children, Ethiopia 6.75. This can be
compared with advanced nations like Spain 1.13 children, Italy 1.20, Sweden
1.29.
But there is no doubt that related to such problems is quality of education
for every male and female. For example, in Niger adult literacy is only 15.9%,
compare Spain 98%, Italy 99%, Sweden 99%. I have noticed when I have
visited some nations in Africa, most females do not go to school! It is interesting
that the Scientific America (September 05), recently reviewed such problems
including a section on the Population Peak.
I was born on a farm in New Zealand. The main activity was sheep
farming. One thing we all knew: Every year the climate, rainfall, can be
different. If there is too little rain, there is too little food and the animals die.
This is also true for humans.
There is also no doubt today that our climate is changing - Human
activities have changed the chemistry of our atmosphere and the convection
patterns in the oceans and atmosphere. There is a fundamental feature of all
fluid convecting systems. As the temperature increases, so do chaos and the
unreliability of predictions. Today we see new deserts, new floods, etc. etc. As
the recent UN meeting in Montreal, Canada, showed, even our leaders in politics
are beginning to accept the reality of climate change and the technologies
causing such changes.
Our life support systems include climate, air to breathe, water quality and
quantity, and quality food. At this time millions of people suffer from serious
malnutrition as in the nations of Africa, India and many more.
As the famous French scientist and philosopher, Montaigne, said in 1581,
“diversity is security”. There is no doubt that good nutrition requires great
diversity.
Food diversity again depends on climate, water and soil quality. Soil
quality is not well appreciated by many. In many of the most fertile regions of
the world the soil quality and quantity is related to recent tectonic events,
including volcanism and mountain building. Good soils need complex mineral
components, quantity and diversity. Recent studies in the NW of the U.S.A.
show that crop yields increase following many volcanic eruptions. Work in many
nations has shown that the use of pesticides, etc., can decrease quality and
increase health problems. As they say in the U.K., such chemicals are
BIOCIDES. As with all animal species, we require a very wide range of
chemical elements in our food. We have studied soils of laterite soils, and water
chemistry in many parts of the Amazon region. Temperature is warm, rainfall
can be extreme. The key nutrients in the soils can be washed away leaving soils
not productive and not producing quality food. The great trees live because of
complex root systems and dust. We have shown that some parts of the Amazon
River water systems are very deficient in mineral components. For a general
discussion, see Van Straaten, 2002 [5], Brown and Wolf, 1984 [1].
Without our present energy technologies we would not have our present
population. Imagine Canada, where I now live, if there was no electricity,
natural gas or oil.
One of the most urgent problems on our planet is to improve energy
technologies. The largest consumers of energy at the present time are: the
U.S.A., China, and Russia. Major components include sources from coal, oil,
nuclear, hydro, etc. All lead to environmental problems.
Given our present knowledge, can we produce adequate clean energy for
all nations? The first problem we must consider is how to reduce waste by
intelligent construction technologies, and use of fossil fuels. I have noticed the
potential use of underground construction in some nations.
Do we all need a big automobile?
Work in many nations has shown how we can use natural energy
resources such as the Sun, Wind, tides, and beneath our feet, geothermal energy.
As the temperature rises about 30o C every km beneath the surface, there is vast
potential for geothermal heating. Even more potential is available where we
have recent active volcanism. A few nations have made use of such resources.
And there is active volcanism in certain rock types, the hot fluids are rich in
hydrogen, a clean gas energy resource (see Fyfe, 1999) [2].
We must improve our management of waste products. In this respect,
‘Europe leads the world! First reduce the use of materials which cannot be
recycled and reused, as with plastic bags, etc.
In most nations that use nuclear power, there is a problem as to where do
we put nuclear waste where it will be safe for millions of years (Fyfe, 1999) [2].
But we also have many other examples of waste problems, as with materials
which involve elements such as arsenic in mine wastes.
Recent work we have done on the volcanic basalt rocks of Hawaii show
that these rocks will absorb carbon dioxide. When we burn coal and oil, why not
put the gas products underground? I have discussed this problem with many
people in China and India.
Do we design cities intelligently? Again, in this respect Europe leads the
world. As I mentioned above, storage facilities, etc. can be built underground.
Don’t waste the surface! To live in a city, do you need a car? My daughter lived
in London, U.K., for several years. As she once said, a car is a nuisance. Public
transport is excellent. Can you walk to work, etc. etc.?
I have been in many cities in Europe. You do not need a car. Many
hotels are designed to save energy and water needs. It is possible!
We all use massive quantities of materials derived from our mining
industries. How much iron do we need? Do we recycle what we have used? Do
you need a gold ring? How much rock was mined to get that gold? Was it in
your nation or imported? Most gold ores are rich in arsenic and there are many
examples of deadly arsenic pollution related to gold mining. The same is true for
many of the materials used in the fertilizer industries of our agri-technology.
2 Concluding thoughts
I always remember discussions during a meeting of the International Union of
Geological Sciences in Florence, Italy. There was agreement that the two basic
sciences are astronomy and earth science, from which all others are derived.
Also, there was concern about how few scientists were working on the urgent
world problems. As was said, too many scientists spend their life finishing their
Ph.D. research projects.
World data are clear. The quality of life in nations is related to the quality
of education for all people, all ages. And another feature of education is clear.
Numerous studies of animal life including our relatives, the apes, show that in
general the females are more intelligent than males. There are many reasons
why this should be. When we examine nations and their problems, such as
AIDS, there is a clear relation between problems like this and female education.
We also need more females in governments as with the situation in N. Europe. I
recently attended a wonderful conference in England. It was based on the need
for education for all. We had about 10,000 young people of all ages, and about
1,000 school teachers from all Europe. The discussions were wonderful (see
Moody et al, 2000) [3].
We need new systems, new groups of experts to integrate knowledge and
the planning of our world for future generations. Such groups must include
scientists, sociologists, economists, politicians and, particularly, citizens of both
sexes. As Sir Crispin Tickell stressed, we must plan for now and for future
generations. That is why the focus of this meeting on Landscape Evaluation is
so important.
My final question to all experts at this meeting. When you leave the
planet, is it better for all people than when you arrived? My answer is NO, it is
not better and a major cause is overpopulation in many nations related to a poor
quality of education for all people. We must improve our systems and plan for
the next generations.
References
[1] Brown, L.R. and Wolf, E.C. Soil erosion and quiet crisis in the world
economy, WorldWatch paper 60. The Economist 2003. Pocket world in
figures. 236 pp. 1984
[2] Fyfe, W.S. Clean energy for 10- billion humans in the 21st century - Is it
possible? Coal Geology 40:85-90, 1999.
[3] Moody, R.J.J. et al, Earth Alert, the past, present and future of our planet.
Geologists Association, London, 149 pp. 2000.
[4] Crossroads for Planet Earth, Scientific American, Special Issue, September
2005.
[5] Van Straaten, P. Rocks for crops. Department of Land Resources, University
of Guelph, Canada. 338 pp, 2002.
[6] Vernadsky, V. The Biosphere. Synergetic Press Inc. 1986. 82 pp. 1929.
Section 1
Environmental planning
and management
Ethics, geological risks, politics and society

A. Cecioni & V. Pineda
Departamento Ciencias de la Tierra, Universidad de Concepción, Chile
Abstract
Since humans live in communities, the moral regulation of behavior has been
necessary for the collective well-being. From the antiquity, the Greek
philosophers discussed moral behavior, which led to the later development of
ethics.
Some geological processes, such as earthquakes, tsunami, landslides,
volcanism, and geological structures such as faults, are potentially hazardous for
the inhabitants. Some geological processes can affect entire cities and active
faults can damage buildings constructed on them.
In general, the urban and urban development areas are sustained on numerous
components. Among these components can be pointed out the economic one that
regulates the urban already consolidated zone and the areas of urban expansion;
the information provided by geologists about the possible occurrence of
geologically hazardous or risky processes; and the political authorities that must
take the difficult decisions based on the economic component of the private
investors, the geologically dangerous processes and the commitment of the State
to provide inexpensive accommodation to low social-economic people and the
responsibility of granting safety to the society.
Analysis and questions about professional ethics, investors’ ethics and
political ethics of the State, taking as a goal the information that must be
provided to the citizenship and to the civil protection, are discussed.
Keywords: ethic, geological risks, land planning.
1 Introduction
Ethics (gr. êthos) is the support of the principles of the human behavior, called
also morality (lat. moralis).
Ethics is the natural norm of the human conduct. Since the peoples live in
community, the moral regulation of the conduct has been necessary to obtain the
collective well-being.
doi:10.2495/GEO060021
Among the classic civilizations, the Egyptian developed the ethics combined
with the religion; in classic China culture Confucio’s doctrine was accepted as a
morality code. From the VIth century B.C., the Greek philosophers, thought
about the moral conduct, developing the ethics as a way of personal life related
to the respect for the social community.
Socrates did not partake the ideas of the sophists. His philosophical position
it is represented in the dialogues of his disciple Plato. The Socratic vision can be
summarized in the following concepts: the virtue is knowledge; the people will
be virtuous if it has instruction and culture, and the vice (evil) it is the fruit of the
ignorance. According to Plato, the human soul is composed by three
components: the intellect, the determination (“the want”) and the emotion. The
virtue of the intellect is the knowledge of the intentions of the life; the
determination has the virtue of the value in the aptitude for acting; and the virtue
of emotion it is the prudence. Plato thinks that the real reality is the purification
of all the passions, which originates knowledge, temperance, justice and virtue
[1] talents that he reiterate in his works. Plato has an absolute conception of the
“good” and of the virtues of the ethical and moral attitudes.
Christianity extended the application of the moral values to the whole
society, including the slaves. This conception of equality in agree to the ethics
and to the morality, is represented in several references and concepts, such as
loves your neighbor as to you itself [2].
It is interesting to observe the masterly classification of the faults that Dante
Alighieri describes in the Chapter of the Hell of the Divine Comedy [3]. He
places the more condemnable attitudes between 7th and 9th circle of the Hell; the
violent ones against God, against the nature and against the society imprisoned in
the seventh circle; and the traitors to their benefactors condemned to the ninth
circle. Then, according to Dante, we have that the treason and the lack of ethics
and of morality, are located in the maximum circles of punishment.
For need of living together, there were formed groups of persons that
constitute the society. The city represents the symbol of the social collectivity.
From the first cultures and men’s groups, the city constitutes a living space in
which the society feels protected. In the city there develop the economic, cultural
and family activities. Consequently, to be able to coexist, the persons must act
with ethics and be mutually respectful. Leaders and scientists are not exempt
from the ethical duty towards the society.
The society must be regulated in the ethics and morality conditions.
Institutions and laws were created. But the laws become anachronistic, since they
are applied to a way of life and of attitudes that happen in a certain temporary
and circumstantial episode. If this social situation is overcome, the law remains
obsolete. It is very difficult to be able to establish laws on events that might
happen in certain future circumstances, in conditions that we do not know. The
serious problem is not to legislate and not to take decisions related to situations
that we know, that can happen and that are hazardous for the safety of the
society.
In this respect, no person has the right to damage other one, understanding
the damage in all the possible forms. The duties and the rights exist. One of the
principal duties is the honest and complete information, since giving only a part
of the truth is a falsehood. One of the principal rights is the attribution of living
safely, and this is an obligation of the State.
Politics, science and society must constitute a harmonic whole founded on
the ethics.
2 Geological hazards and some related question: Concepción

area as a pattern
In Concepción (South-Central Chile) area, that could be a case of other similar

regions of other countries, there occur constructions that are founded on faults, in
zones of flood for tsunami and/or that can be affected by landslides. This
situation causes a decrease of the quality of life and a dangerousness situation for
the inhabitants.
It is thus convenient to design and apply procedures for determining
responsibilities. At this point, we must consider some questions related to (A) all
ready constructed areas, and (B) use of expansion urban areas.
(A) For already constructed areas, some questions are:
Who is the responsible to notify to the inhabitants of a building that it is
constructed over a fault that could damage the building in case of an
earthquake?
Who is the responsible to notify to the inhabitants that the coastal zone in
which they live or work is dangerous for tsunami hazard?
Who is the responsible to notify to the inhabitants that the area where they
live or work could be destroyed by a tsunami?
Who is the responsible to notify to the tourist skiers that the volcano where
they are is not monitored?
(B) For expansion urban areas:
Who is the responsible to approve the use of a certain hazardous areas for
general social benefit?
Who is the responsible to consent the construction of a building on a fault?
Knowing the dangerousness of some area, who is responsible to accept the
proposal of an investor, which have, obviously, an economic interest to
develop constructions?
Who is the responsible to request information of hazard natural processes,
vulnerability and risk of some areas?
In the area of Concepción there are active faults, some zones with high
seismic risk (Fig. 1), landslide risk (Fig. 2) and tsunami risk [4] (Fig. 3).
3 Some answers
Answers to these questions could be very complicate for some ones. But
applying ethics, the solution is simple.
(A)
(B)
Figure 1: (A) Main faults in Concepción city. (B) Zones of major seismic
dangerousness, considering faulting, soil, underground waters and
other geological factors.
(A)
(B)
(C)
Figure 2: (A) Landslide in zones of housings. (B) Slide in mountain highway.

(C) Crown of new landslides (retrogresives) – hazardous zone.
Figure 3: Zone of flood by tsunami indicated with light gray colour,

extending across the whole city of Talcahuano.
The occupation of an investor is to make investments to develop social

activities, for which it is normal to expect economic benefits. Scientists must
don’t have conflict of interests and its role is to study neatly the territory, analyze
in an objective way the hazardousness and associate risks, prepare maps of
vulnerability and to warn the authorities about these dangers. The politicians and
managers of the State, as the scientists, must not have created interests and they
has the obligation to be informed about the geological and natural risks, in
general, before granting the permissions of urban development.
Consequently, the administrator of the State must warn the population about
the dangers and risks of geological processes that could occur in the areas where
they live and work. Additionally, the administrators must evaluate the
convenience of authorize investor to occupies territories for some social
activities. On the other hand, the administrators of the State must request studies
of vulnerability of geological risks that allow to carry out a suitable managing of
the territory and, in the case that some dangerous zone should be occupied for
urbanism, industry and tourism, among others, they will have to request and
finance a appropriate monitoring system for the geological hazardous processes
that affect this territorial space.
4 Thoughts and conclusions

In several countries there do not exist procedure that force to the State to effect
studies of vulnerability, risk and monitoring of geological hazardous processes.
The State provides ambulances and hospitals to care patients who have
accidents; it provides policemen to protect the citizens of the delinquency; it
provides firemen to attack fires and to carry out rescues. Then, why the State
does not place in the same plane the geological hazardous processes and
provides to the society of a strategy of monitoring intended to save lives?
We are strongly sure that an international organization must demand to the
States the application of studies on geological hazardous processes and to carry
out an ethical planning of the territory, in order that the society could feel surer
of living in an environment in which these processes are monitored.
References
[1] Platón. Diálogos Socráticos. Traducción de Patricio de Azcárate. Editorial
Océano de México. S.A. 365 p.
[2] Sagrada Biblia. Biblioteca de Autores Cristanos, de la Editorial Católica,
S.A. Madrid 1970. Cuarta Edición. San Lucas (10:27). 1377 p.
[3] Dante Alighieri. La Divina Comedia. Adapt. Francisco José Fernández
Defez - México, D.F.: Editorial Selector, S.A. de C.V., 2004. 96 p.
[4] Servicio Hidrográfico y Oceanográfico de la Armada de Chile. Bahía de
San Vicente. Carta de Inundación por tsunami, escala 1:10.000. Armada
de Chile, SHOA, 2000.
Environmental impact assessment and

environmental management plans:
an example of an integrated process
from the UK
M. A. Broderick1 & B. Durning2
1
Halcrow Group Ltd, UK
2
Oxford Brookes University, UK
Abstract
Environmental impact assessment (EIA) is a mature process implemented around
the globe to identify significant impacts from development and provide
mitigation measures to reduce these impacts. Increasingly in the UK the process
is being supplemented through the integration of an environmental management
plan (EMP) into the resulting environmental statement. The EMP specifically
aims to manage the impacts during the construction phase of the development.
This paper presents an example of practice from the UK in this integrated
process for the installation of a high pressure natural gas pipeline through open
countryside. It demonstrates the added benefit that the integrated process
provides in managing and reducing environmental impacts from the
development.
Keywords: environmental assessment, environmental management,
construction, pipeline.
1 Introduction
The environmental impact assessment (EIA) process is a well established
method used around the globe in the identification and mitigation of the impact
of developments on the environment. It consists of a series of studies and
discussions which are designed to:
• identify which legislation is relevant to the proposals (screening)
• assess the scope of the project (scoping+consultations)
doi:10.2495/GEO060031
• identify the nature of the existing environment (baseline)

• obtain stakeholders views on proposals (consultation)
• identify the impacts of the proposals and predict the likely magnitude and
significance of those impacts on the environment (environmental assessment)
• allow the formulation of mitigation measures (mitigation).
The outcome of the process is the production of an environmental statement
(ES). Increasingly in the UK the EIA process for developments at varying scales
is being supplemented through the integration of a voluntary environmental
management plan (EMP) into the resulting ES. The purpose of the EMP is to
serve as an operational manual for implementing appropriate environmental
controls and monitoring procedures within the construction phase of the
proposed development. It sets out to ensure that the construction of the works
complies with relevant environmental legislation, licence conditions and
accepted good practice and that measures to mitigate impacts discussed in the
project ES are implemented.
In this paper we present a case study from the UK of the construction of a
high pressure natural gas pipeline, in central southern England, which was
subject to the EIA process with a fully integrated EMP (Environmental
Resources Management [1]). It is based on the experiences of one of the authors
who at the time was employed by consultants Environmental Resources
Management (ERM) and undertook the EIA and implementation to the EMP.
We aim to demonstrate how the combined process ensures that environmental
management procedures during the construction phase limit environmental
impacts and ensure that good quality restoration of the pipeline route limits
environmental damage, to the extent that eight years later there is very little
evidence in the landscape for the presence of the pipeline.
2 Background to case study

The pipeline, from Aylesbury in Buckinghamshire to Chalgrove in Oxfordshire,
a distance of approximately 26km, was constructed by National Grid (called
Transco at the time of pipeline construction in 1999). National Grid is
responsible for the operation of the national gas distribution system in Britain
known as the National Transmission System (NTS). The NTS transports gas at
high pressure along a network of pipelines from the gas production terminals to
major gas users (“40 power stations, a small number of large industrial
consumers”) and a series of “Local Distribution Zones” from which the gas is
then distributed at a lower pressure to consumers (National Grid [2]). The case
study pipeline was constructed to provide additional capacity for the NTS in
order to satisfy an increase in demand for gas in the south of England. It
supports an existing pipeline which runs parallel and continues past Chalgrove.
3 EMP as part EIA process

The requirement for screening of developments which are possibly subject to the
environmental assessment (EA) process has been a requirement in UK law since
the late 1980s (although the legislation was revised in the late 1990s [3]).
Although not all developments are required to go through the process the
requirement for high pressure natural gas pipelines falling within certain criteria
(length, diameter) to go through the EA process is enacted in specific gas
pipeline legislation [4, 5]. There is no requirement in UK or EU legislation for
an EMP to be part of the EIA process. However, it is part of National Grid
corporate procedures that an EMP be included in the contract documents on
which the commercial terms for the construction contract are negotiated.
3.1 The EMP process
The broad purpose of the EMP is to:

• provide a mechanism for ensuring that measures to mitigate potentially
adverse environmental impacts are implemented
• ensure that standards of good construction practice are adopted throughout
the construction of the pipeline
• provide a framework for mitigating impacts that may be unforeseen or
unidentified until construction is underway
To be successful an EMP should evolve throughout the life of the project.
For this project the EMP was issued for consultation to various stakeholders and
was refined as additional information, design changes or comment from
stakeholders becomes available. An EMP can therefore also provide assurance
to stakeholders that their requirements, with respect to environmental
performance, will be met.
Although the EMP detailed the mechanisms through which the issues outlined
above were to be addressed during construction of the pipeline and the
responsibilities for meeting them, it was the contractor who was required to
provide method statements of the details of the actions to be taken, in order to
implement each aspect of the EMP. The method statements had to demonstrate
how compliance with the requirements of the EMP were to be achieved, and
specify the names of the individual people who will be charged with achieving
and monitoring compliance
3.2 Auditing/monitoring during construction
The EMP also provided a framework for compliance auditing and monitoring to
ensure that its aims are being met. As the EMP formed part of the commercial
contract for the contractor during the construction of the pipeline, National Grid
required that inspections and audits were undertaken to ensure that the plan was
being implemented. In addition to any audits the contractor may undertake,
National Grid also commissioned their consultants to undertake periodic site
audits. A checklist pro forma was used which covered the environmental issues
addressed in the ES and the EMP. Where problems were identified corrective
actions were required to be undertaken. These could include further direct
mitigation, changes to procedures or additional training.
4 Aylesbury-Chalgrove pipeline
For natural gas pipelines the EIA process can be divided into a series of
successive components which mirror the overall engineering design and
construction process and also the general approach adopted to EIA for other
developments (see Table 1).
Table 1: Overview of environmental, design and construction of pipelines.
EIA process for natural Pipeline design and General approach to

gas pipelines construction phases EIA required in UK
legislation
Establish need Pre-feasibility Screening
Route corridor Feasibility Scoping
information study
EIA (baseline data Conceptual design EIA (baseline data
collation, impact collation, impact
prediction) and prediction) and
production of ES production of ES
Production of EMP Detailed design, Consenting from
commissioning and Department of Trade &
construction Industry
4.1 Route corridor information study
The pipeline route was to be from Aylesbury in Buckinghamshire to Chalgrove

in Oxfordshire, a distance of approximately 26km. The desk based route
corridor information study was carried out at the pipeline feasibility study stage
and constraint/overlay maps were produced at a scale of 1:50 000, based on:
• geological/ground conditions
• distances
• archaeological sensitivities
• ecological sensitivities
• numbers of roads, rivers, railway, hedgerow crossings
• two route corridors 1km wide
The route corridors passed completely through open countryside, in an area
known as the Vale of Aylesbury. The topography along the route was gently
rolling, although some areas of high ground lay on the fringes of the route. It
also passed through a river valley system (River Thame).
The majority of the route corridors was underlain by clay (Jurassic and
Cretaceous) with some minor areas of Cretaceous sandstone. Very little
superficial deposits occurred along the route and those that were encountered
were clay deposits and some floodplain gravels.
The key environmental issues identified from the route corridor study which
would require detailed evaluation were: archaeology; ecology and land take.
Secondary issues which would require consideration, were impact on the

landscape and drainage as the land use along the route was agriculture.
4.2 EIA and production of ES
The next stage in the process was the undertaking of the EIA, which involved
detailed baseline surveys and the collation of the information into an ES where
impacts were identified and mitigation measures proposed. Information
concerning the project itself, including background, project schedule,
construction techniques, restoration and operation were also included. Important
information and areas of particular concern were identified and plotted on
constraint/overlay maps at a scale of 1:10 000. A preferred route corridor was
brought forward for EIA based on distance crossing and avoidance of sensitive
areas.
4.3 Production of an EMP
The final stage involved the production of the EMP, detailing all environmental
constraints along the final route, and the mitigation measures to be taken. It also
included detailed restoration practices and highlighted areas where aftercare was
necessary, and the nature of the care required. Areas of particular concern were
identified and plotted on constraint/overlay maps at a scale of 1:2500. The EMP
identified eleven activities which may give rise to potential impacts during the
construction of the pipeline and for which mitigation measures were required
(ERM [1]). These are reproduced in Table 2.
Table 2: Proposed mitigation to environmental impacts contained within
EMP.
Activity Potential impacts Proposed mitigation

Pipeline Damage to significant Will be avoided by: re-routing;
construction ecological, archaeological restriction of working width;
(physical sites and species bore underneath site; rescue dig
disruption from for archaeology; translocate rare
clearance of the plants; minimise hedgerow
working width, removal and avoid trees;
pipe stores and carefully reinstate topsoil and
temporary habitat
working areas)
Fuel storage Leakage/ Site stores located >50m from
spillage may give rise to watercourses. Bunded (110%
contamination affecting: capacity) design with
abstraction downstream; impermeable liners for stores
ground-waters; ecology of and refuelling point will be used.
surface waters Use drip trays wherever possible.
Provide local first response
absorbents, booms etc. Training
will be given to all staff. Inspect
the works frequently.
Table 2: Continued.
Activity Potential impacts Proposed mitigation

Prepare and exercise oil spill
contingency plans. Carry out
regular inspections and
maintenance of plant.
Machinery Smoke and fumes Proper maintenance will be
operations maintained
Noise Sitting (pumps, generators etc)
away from dwellings. Provide
adequate silencing. Switch off
when not needed
Dust Restrict vehicle speeds. Spray in
dry weather
Construction of Fisheries (migration and Schedule construction activities
river crossings spawning) to avoid sensitive times or
minimise sedimentation effects.
Stimulation of bank/bed Design adequate emplacements
erosion and protection measures.
Clearance and Sediment run off to Provision of cut-off drainage and
activities within watercourses, ponds and settlement ponds.
the Right of lakes
Way
Effects on fish Discharge of pump outlet to soil
(asphyxiation; indirect surface/crops to promote seepage
effects on mitigation and (subject to agreement).
spawning)
Waste Contamination of soils Effective containment and
management and water from wastes management of wastes.
External pipe Contamination of soils Use mechanical brushing in lieu
cleaning and water by blast grits of grit blasting.
Machinery Topsoil compaction Strip and segregate topsoil.
movement Break up panned sections
(trafficking)
Hydrotesting Disruption from Plan and agree abstraction and
abstraction and disposal of discharge points, rates and
large volumes of water contingency arrangements.
Pollution from additives Minimise/avoid use of dyes,
corrosion inhibitors, oxygen
scavengers
Erosion from failure under Reinstate to pre-erosion
test conditions
Disruption to Water-logging/crop losses Insert header drains
field drains Reconnect severed drains
Following completion of the ES and EMP the document was sent to the
relevant Government department for approval, as required under relevant
legislation. They undertook a consultation process with statutory bodies and

regulators before granting approval for the pipeline.
1. Receiving material
2. Setting out
3. Pre-construction land
drainage
4. Right of way and topsoil
stripping
5. Stringing
6. Welding
7. Excavation/trenching
8. Ditching/lowering and lay
9. Tie-ins
10. Bedding and covering pipe
11. Backfilling
12. Reinstatement
13. Post construction land drains
14. Final trim
Figure 1: Normal sequence for pipeline construction.
Figure 2: Pipeline route at Stage 4 – excavator is passing over a road and

between a gap where hedging has been removed along the field
boundary.
4.4 Construction of pipeline and auditing EMP
Barnett and Jordin [6] provide a useful guide to the pipeline construction process
adapted for the case study. A “spread” method is employed for the construction
which “involves several groups of workers and equipment who collectively
conduct the various stages of the construction operation. Each group completes
an activity which picks up where the last one left off, advancing the construction
process a step at a time and leaving it ready for the next step to begin”. Of note
is that fact that construction work is limited to a “seasonal window which extends
from March/April to October during which time the weather is more predictable
and ground conditions are more favourable”. The normal construction sequence
is given in Figure 1 (taken from Barnett and Jordin [6])
Auditing during construction of the potentially impacting activities listed in
table 2 was carried out regularly by consultants using the methodology referred
to in section 3.2.
Figure 3: Area of restored hedging and field (taken January 2006 from a road
looking along the restored pipeline route).
4.5 Post construction follow-up

Due to the national importance of the NTS there is post construction follow up of
this development: aerial surveys are regularly employed by National Grid to
monitor the route and agriculture liaison officers and land agents maintain
regular contact with land owners. In order to assess whether there was any
evidence of degradation of the landscape caused by either the pipeline
installation or poor restoration a number of points along the pipeline were visited
by the authors in January 2006. The points included a road crossing (similar to
that in Figure 2), a footpath and a stream crossing. The location of the pipeline
was determined by identifying its position on the ES map and locating the
National Grid marker post in the nearest road. In all cases very little evidence
was found of a legacy of the pipeline. In one field the route could be identified
by a darker green swath of grass, suggesting that restoration had improved the
land quality. The most obvious evidence was in the restored hedging (see
example in Figure 3) where the extent of growth is not yet equivalent to that of
the hedging removed.
5 Discussion
EMPs provide a critical link between EIA and project implementation. In effect
EMPs comprise the operational response plan that implements the mitigation and
monitoring programs for the project. The execution of an EMP is increasingly
becoming conditional to project approval or licensing, and/or to project financing
(Equator Principles [7]). The preparation of an EMP acceptable to all
stakeholders is therefore a key part of the project development process.
Preparing an effective EMP requires a balance between what is desirable, what is
affordable, and what can be implemented. In particular it requires:
• all stakeholders to have a common understanding of the objectives of the
EMP and particularly to understand the link between the EMP and any
approvals or conditions that may be applied to the project on its
implementation
• project owners/proponents to have an understanding of the requirements of
relevant permitting processes applicable to the EMP and/or be familiar with
the needs of specific, relevant financing agencies
• EMP costs to be clearly defined and understood by all parties
• provision to be made for sustainability in implementation of the EMP,
particularly in post construction monitoring of impacts.
EMP provides a concrete reassurance that construction/operational impacts
identified in the EIA are addressed and mitigated during construction/operation.
However, the absence in legislation of a requirement for EMPs and follow up
environmental audits is a weakness, a fact which is increasingly being
recognised (Morrison-Saunders and Arts [8]). EIA legislation could be
strengthened and made more credible if EMPs were mandatory, incorporating
environmental auditing during construction/operation.
References
[1] Environmental Resources Management (1998) Aylesbury to Chalgrove Gas
Pipeline: Environmental Review. Report prepared for Transco, UK.
[2] National Grid http://www.nationalgrid.com/uk/Gas/About/
How+Gas+is+Delivered/
[3] Town and Country Planning (Environmental Impact Assessment) (England
and Wales) Regulations (1999) SI 293
[4] Gas Act (1995) c 45
[5] Public Gas Transporter Pipe-line Works (Environmental Impact
Assessment) Regulations (1999). SI 1672
[6] Barnett, J. and Jordin, M, (1998) Pipelines – a worm’s eye view. Transco,
Ambergate, UK
[7] Equator Principles http://www.equator-principles.com/principles.shtml
[8] Morrison-Saunders, A. and Arts, J. (2005) Learning from experience:

emerging trends in environmental impact assessment follow up. Impact
Assessment and Project Appraisal 23 (3) 170-174
Geological factors in monitoring and

planning nature trails at tourist centres
in northern Finland
K. Lehtinen & P. Sarala
Geological Survey of Finland
Abstract
Increasing tourism in northern Finland produces challenges for tourist centres,
their land use plans and sustainable development. Increasing activities in areas
that are sensitive to environmental changes need solutions for sustainable land
use. Equipment to minimize environmental effects, and the planning of
ecologically, culturally, and visually sustainable built up areas at popular tourist
centres on the Ylläs and Levi fells are investigated in the LANDSCAPE LAB
project. The project is partly financed by the EU LIFE Environment and the
Geological Survey of Finland is involved as a partner. Geological factors such as
the quality and composition of bedrock, the maturity of the matrix of surficial
sediments, and geomorphology affect the resistance to erosion in different
geoenvironments. In Fennoscandia, the glacial erosion has been intensive and the
terrain is composed of eroded hill slopes and glacial landforms. Traditionally, in
Finland, erosion resistance of nature trails has focused on vegetation and
trampling resistance. The erosion rate is studied mainly by measuring the width
and depth of the path and the amount of exposed roots and stones. Geological
factors are not studied for nature trails, but geology and geological factors are the
basic elements affecting resistance to erosion and should be included in land use
planning. Careful planning and monitoring are the keys to creating visually
impressive and geologically sustainable nature trails.
Keywords: erosion resistance, geological factors, nature trail, land use.
1 Introduction
Increasing tourism causes pressure on the land use planning and sustainable
development of tourist destinations. Northern Finland tourist destinations have
doi:10.2495/GEO060041
several thousands of kilometres of outdoor routes. These routes are situated in

areas of different status like protected nature reserves, nature parks, state owned
recreation areas and privately owned land. Many outdoor routes have multiple
uses like, for example, nature trails for hikers, skiing and snow mobile routes and
dog sled routes.
Outdoor routes that are used as nature trails can be classified in four groups:
1. Old, naturally formed summer routes, which are historical travelling routes;
2. Old, naturally formed routes that were developed as nature trails after the
construction of guiding services;
3. Natural types of routes or other older routes, which have become nature
trails when taken into multiple, round-the-year use because of increased
tourism (for example some winter routes taken into nature trail use);
4. Other new outdoor routes, originally made for different purpose, but are
afterwards used as nature trails (for example constructed skiing routes,
which are used as nature trail at summer).
Nature trails at tourist destinations are situated in places that have something
special to see. Speciality is often based on the geology of the area. Geological
features like block fields, gorges, lateral channels, esker chains and hanging bogs
are major components of fell landscape (Fig. 1). These features are the key to
creating the perceptions of hikers and other tourists that these areas are unique.
These geoenvironments are sensitive to erosion due to their geological
properties. Furthermore, tourist destinations in northern Finland (Fig. 2) are
situate in a severe climatic zone with four seasons; a short growing season in
summer, frost heaving and frost weathering problems caused by the cold winter,
erosion impacts of melt waters during spring and autumn rain. Thus, nature trails
become sensitive to impact of increasing amounts of visitors. Nature trails used
at spring and summer, are particularly sensitive to erosion problems.
Figure 1: Scenery of highland on the Pallas fell.
Figure 2: The location of study and target areas in northern Finland.
The effects of nature tourism on nature trails have traditionally studied in

Finland using the trampling resistance of vegetation [1, 2, 3]. The erosion status
of nature trails is also estimated by measuring the width and depth of the path
and by the amount of exposed roots and stones. Several visitor studies were
made [4, 5] concerning visitor counting and impact of the amount and the type of
visitors at nature trails. Geological factors on the resistance to erosion in relation
to natural trails have earlier been studied for example in Sweden and North
America [6, 7] but not in Finland. Finnish studies concerning construction work
and road building have been done, but the circumstances and objectives of these
studies were different. However, geology is the basic factor affecting both
vegetation and erosion resistance properties of the areas. For planning,
monitoring and conservation of nature trails, geological factors and the
geological history of the area must be studied. Things to investigate are the
bedrock quality and composition, geomorphology, Quaternary deposits and their
properties like maturity and composition of the matrix of surficial sediments.
Geological factors in land use and nature trail planning are studied by the
Geological Survey of Finland as a part of the LANDSCAPE LAB project. The
aim of the study is to investigate how and at what rate geological factors are
affecting the resistance to erosion in different geological environments at tourist
destinations in northern Finland. The objective is to develop a system to predict
erosion effects and to develop a tool for nature trail planning and land use. Based
on field and laboratory studies and visual examination, a classification of erosion
resistance of different landform and soil types will be created and the equipment
for monitoring erosion will be investigated. Furthermore, one other aim is to
make recommendations for the planning of new nature trails and for suitable
cover material for nature trails in different environments.
2 The Project LANDSCAPE LAB
The project LANDSCAPE LAB (2004-2007) ‘Tourist Destinations as Landscape

Laboratories – Tools for Sustainable Tourism’ is an EU LIFE Environment
supported project situated in Finnish Lapland. The beneficiary is the University
of Lapland and there are nine partner organizations and two municipalities
involved in the project. The main target of the project is to determine solutions to
sustainable land use, and to plan ecologically, culturally and visually sustainable
built-up areas, where disadvantages caused by tourism, would be minimized. The
main research areas are the fells of Ylläs and Levi, popular tourist centres in
western Lapland.
The LANDSCAPE LAB Project consists of five tasks of which one
concerns dissemination and management. The others are:
• The LABLAND task, in which the ecologically, culturally and visually
sound urban structures at tourist destinations are studied. It makes
landscape analyses in which the focus is on aspects like geology,
landscape structure, history of land use, visual landscape, soundscape
and landscape experiences.
• The LABECO task, in which the extent and types of the environmental
impacts are studied to determine bio-indicators suitable for monitoring
environmental effects for tourist destinations.
• The LABSOC task, where functional and social structures and activities
of community are studied.
• The LABPLANT task, where usable and hardy plant species for
sustainable restoration needs and landscaping in northern areas are
sought. The task will select hardy plant species, develop methods of
plant propagation and produce plant material for planting.
By combining all the data produced in the project, recommendations for use
in the planning of tourist areas and implementation of environmental care will be
presented.
3 Study of geological factors

3.1 Geological environment
The study area is situated in northern Finland and consists of nine geologically
variable target areas (Fig. 2). The target areas have variations in bedrock quality
and composition, geomorphology, glaciation history, Quaternary deposits and
deepness of surficial sediments. The bedrock in Finland is part of Fennoscandian
Shield, which is the oldest part in Europe (1.8-3.5 Ga). The bedrock consists
mainly of quartzite, amphibolites, granulites and granites. In the study area, the
rock types are quite resistant to erosion and are nowadays seen as fell areas at
Finnish Lapland. The bedrock in the topographic depressions consists mainly of
schist and greenstones [8, 9].
Finland was repeatedly covered by continental ice sheets over the last two
and half million of years. Finland’s position in the central area of the ice sheet
and the variation in glacial dynamics have formed various Quaternary deposits
and glacioerosional features [10]. Different kind of till beddings, moraine
formations and glaciofluvial deposits were formed [11, 12, 13]. These features
control land use and have an influence, for example, on the location of
settlements and the formation of passages.
3.2 Field and laboratory studies
Field studies started at summer 2005. The status of erosion on nature trails was
mapped by measuring the width and depth of the path and the amount of exposed
stones and roots. In luxuriant areas, where soil is enriched of nutrients,
vegetation is worn and roots are often exposed (Fig. 3). In barren areas, stones
are usually exposed and impeding walking (Fig. 3). The erosion rate caused by
approximately the same amount of visitors was not equal in different
geoenvironments. Different kind of soils and Quaternary deposits are the key to
studying erosion on nature trails. Mire and areas where bog formation is on-
going, and deposits like dunes and deflation areas (Fig. 4), are particularly
sensitive to erosion. Nature trails in these areas need protective structures like
duckboards and stairs to prevent erosion problems.
Figure 3: Exposed roots on nature trail at Oulanka (left); nature trail on

moraine slope at Ylläs (right). Angular quartzite stones are exposed
when fine material of moraine have been eroded.
One of the objectives is to study how suitable different cover materials are in
different kinds of environments and how resistant they are against, for example,
melt water erosion (Fig. 5) and their stability in variable slope steepness. The
aim is to make recommendations for the use of cover material on nature trails in
different areas. For this purpose, present cover materials were mapped and
samples were taken at target areas. In some places, nature trails sensitive to
erosion are protected with covering materials like bark chip, sawdust, gravel,
mineral aggregate and stone ash (industrial product) to prevent erosion or to
conserve the path. Preliminary results show that the bark chip and sawdust seem
to be problematic, because they tend to stay wet for a long time after rain and
melting of snow. Sawdust also is not an aesthetic material on a nature trail.
Gravel and stone ash do not have a water content problem, but coarse gravel
seems to be unstable on slopes (Fig. 5). In some places, stone ash seems to be a
suitable covering material, but its use is not economic due to high transport
distances and costs.
Figure 4: Deflation area on nature trail at Saariselkä.
The objective is also to define indicators for monitoring geological factors

on nature trails and for estimating erosion potential in planning new trails. 3 to 6
sites on nature trails have been chosen from every target areas for soil sampling.
The samples are taken from the surface layer (depth ca. 30 cm) and analyzed to
determine physical characteristics like matrix composition and maturity,
weathering intensity, absorbing properties and frost heaving properties.
To develop a useful method for studying soil properties and their variations
in the field, geophysical methods will be tested. Electrical conductivity will be
measured at all the sites on the nature trail and beside the trails in early spring
and autumn to study adsorption, permeability and compression properties.
Figure 5: Nature trail at the Ylläs fell (left). Study areas are covered by snow
about six months of the year. During spring, melt waters cause
erosion problems on nature trails. Some trails act as drainage
channels, particularly during spring when the soil does not absorb
water because the ground is still partly frozen; example of erosion
problems on nature trail caused by unsuitable cover material (right).
4 Conclusions
Significance of this study is to investigate measurable geological factors
affecting the resistance to erosion on glaciated terrain. As a result, equipment and
recommendations will be created for planning, monitoring and conserving nature
trails at tourist destinations in northern Finland.
Preliminary results show that geological factors seem to have a significant
impact on resistance to erosion on nature trails. The composition and quality of
bedrock, depth of glacial overburden and grain size, stone content and quality of
surficial sediment have a direct impact on absorption and permeability of soil.
Stone content and quality of surficial sediments also affects the rate of wearing
of fine-grained material at nature trails. However, soil and bedrock properties
affect on vegetation and its luxuriance. To protect and conserve badly eroded or
heavily loaded nature trails, it is important to find naturally looking but durable
cover material.
References
[1] Tolvanen, A., Rämet, J., Siikamäki, P., Törn, A. & Orell, M., Research on
ecological and social sustainability on nature tourism in northern Finland,
Working Paper of the Finnish Forest Research Institute 2, 2004.
[2] Jämbäck, J., Aspects of ecological capacity: trampling tolerance and

disturbance, Saarinen, J. & Järviluoma, J. (eds.). Luonto virkistys- ja
matkailu-ympäristönä. Metsäntutkimuslaitoksen tiedonantoja 619, pp.
143-163, 1996.
[3] Hoogster, M., The effect of trampling on vegetation at four cottages in
Torne Lapland, northern Sweden. Report from Kevo Subarctic Research
Station 19, pp. 25-34, 1984.
[4] Erkkonen, J., Jokimäki, J., Saarinen, J., Tuulentie, S. & Virtanen, E.
(eds.). Policies, Methods and Tools for Visitor Management. Proceedings
of The Second International Conference on Monitoring and Management
of Visitor Flows in Recreational and Protected areas, June 16-20.2004,
Rovaniemi, Finland. Working Papers of the Finnish Forest Research
Institute. 2004.
[5] Erkkonen, J. & Sievänen, T., Standardisation of Visitor Surveys –
Experiences from Finland. Arnberger, A., Brandenburg, C. & Muhar, A.
(eds.). Monitoring and Management of Visitor Flows in Recreational and
Protected Areas, Conference Proceedings p. 252–257, 2002.
[6] Bryan, R., The Influence of soil properties on degradation of mountain
hiking trails at Grövelsjön. Geografiska annaler 59 A, pp. 49-65, 1977.
[7] Hammit, W., Cole, D., Wildland recreation, ecology and management.
John Wiley & Sons, 1998.
[8] Mikkola, E. Muonio-Sodankylä-Tuntsajoki. General Geological Map of
Finland 1:400 000, Explanation to the Map of Rocks, sheets B7, C7, D7
(with an English summary). Helsinki: Geological Survey of Finland,
1941.
[9] Lehtonen, M., Airo, M-L., Eilu, P., Hanski, E., Kortelainen, V., Lanne, E.,
Manninen, T., Rastas, P., Räisänen, J., Virransalo, P., Kittilän vihreä-
kivialueen geologia, Lapin vulkaniittiprojektin loppuraportti, (summary in
English), Geological Survey of Finland, Report of Investigation 140,
1998.
[10] Sarala, P., Glacial morphology and dynamics with till geochemical
exploration in the ribbed moraine area of Peräpohjola, Finnish Lapland.
PhD thesis. Geological Survey of Finland, 2005.
[11] Hirvas, H., Pleistocene stratigraphy of Finnish Lapland. Geological
Survey of Finland, Bulletin 354, 1991.
[12] Kujansuu, R., On the deglaciation of western Finnish Lapland. Bulletin de
la Comission Geologique de Finlande 232, 1967.
[13] Johansson, P, Kujansuu, R. (eds.), Pohjois Suomen maaperä:
maaperäkarttojen 1:400 000 selitys (Quaternary deposits of Northern
Finland – Explanation to the maps of Quaternary deposits 1:400 000)
(summary in English), Espoo: Geologian tutkimuskeskus. 2005.
The acoustical improvement plan as a process

to re-establish acceptable acoustical conditions
M. Clerico & G. Soffredini
Politecnico di Torino, DITAG, Torino, Italy
Abstract
The word “Acoustical Improvement Plan” indicates a set of provisions, related to
land management and suitable to reach the targets defined in the planning, with
particular reference to acoustical standards satisfaction.
The acoustical improvement plan could be interpreted as a loosening of the
most critical nodes checked by the comparison between the noise mapping and
the acoustical characterisation of a territory, but this interpretation could not
reply to the most diffused question of acoustical quality. This condition derives
by an approach founded on a multiplicity of actions and provisions, able to
implement new logics in decisional processes that determines the territory
planning and manages the transformation, with a particular attention to
environmental noise problems.
The purpose of this paper is to describe the plan identity, which isn’t
represented by a specific planning action, but it invests and interests in
particular actions of all politics of planning and territorial management,
involving therefore the necessity to coordinate and to interact with the main
instruments of territorial management.
The acoustical improvement plan will not be the design of the intervention
aimed to restore the sound levels limits, but a process, structured as a set of
provisions and principles of urban planning and government of the territory, with
the purpose to re-establish acceptable acoustic conditions for the critical zones,
but also in order to prevent eventual future suffering.
Keywords: noise pollution, acoustical improvement plan, environmental
planning, land management.
doi:10.2495/GEO060051
1 The characteristics due to a healing plan

The word “Acoustical Improvement Plan” indicates a set of provisions, related to
land management and suitable to reach the targets defined in the planning, with
particular reference to acoustical standards satisfaction.
A healing plan could be made of different measures, as administrative,
legislative and through intervention to reduce environmental noise.
The accordance of these procedures indicates a healing plan as a project that
requires an interaction and coordination among the most important instrument of
environmental management.
The most strategic interaction could be with the functions asked to plan and
manage the traffic and related infrastructures.
The plan identity is not recognized in a specific action of a specific project ,
but involves intentions and actions of the whole politics of the territorial
management [1].
This need of coordination, does not remain an internal request of Local
Administrator, but becomes essential when other Subjects have to prepare and
project a healing plan for the acoustical and environmental improvement.
The healing plan is not the intervention project to take the acoustical value
into law limits, but a whole of coordinate intervention for the progressive
improvement. The interventions have to be different as type, time and related to
specific part of territory or specific sources. The plan aimed to these objectives is
not a project describing works, but a process, fixing structure and administrative
conditions that determine a progressive acoustical improvement [2].
2 The contents of the Municipal Acoustic Improvement Plan

(PRAC)
An acoustical improvement plan (PRAC) is the result of the comparison between
the territorial PCA, “acoustic classification plan” (municipal action mandatory
for the Italian law LQ 447/95) and the relevant survey: it must represent a
solution of the most critical issues and the recovery actions have to answer to an
extensive acoustic quality request which can only be the outcome of an
integrated approach leading to the implementation of new specific strategies
focused on the environmental noise and aimed to the territory management and
evolution.
The improvement plan will therefore be based on a wide actions range not
only finalized to the mitigation of specific limits overcoming, but mainly
intended as a coherent project where at the same time mitigations of the most
critical situations, urban development plans, territorial government and
administration will act to restore an acceptable acoustic situation and to prevent
future problems.
The correct approach to the acoustical improvement plan is therefore to put in
place a set of effective activities, extending the application range from the
reduction of the environmental noise intensity from fixed an mobile sources, to
the mitigation of the acoustic impact on the affected people and to the
optimization of the passive protection performances of the buildings where the

human activities are carried out, starting from the most sensitive situations. In
this logic the technical and design actions to be developed in an improvement
plan can be targeted to:
• noise emission reduction (source matrix reduction)

• receivers exposure reduction (propagation path countermeasures)
• buildings passive protection improvement (where sensitive activities are
located) [1].
3 The applied interaction between noise and land

management
The PRAC is a dynamic tool in the acoustic pollution management and control,
which includes design and direct mitigation capabilities always based on
administrative actions, territorial (town planning and acoustics) and mobility
management, through mandatory carry-back on actuation regulations [3].
Therefore administration, legislation and regulations aspects represent the
core of the plan finalized to containment and mitigation of already compromised
situations: such a scope requires interactions and coordination with the other
territorial management tools and with current regulations.
Even if the PRAC is not intended to produce administrative measures or
urban development plans, this instrument must be used as an integration tool to
include the acoustical impacts analysis on the territory, showed into “acoustic
classification plan” PCA, and finalize proposals for future development. This
point can be developed by the identification of procedures, operative tools and
processes for the monitoring of acoustic effects related to different options, or
through the comparison of acoustical effects of alternative scenarios.
This can for example applied to the development of the City (PRGC) and
Mobility plans and to the definition of the acoustical optimization criteria, or
define references to be used in administrative measures finalized to the traffic
management (such as Limited Traffic Zone (ZTL), heavy vehicles traffic, speed
limitation along specific city areas).
Due to the important role the territorial government plays especially on
mobility, the contribution of the improvement plan also accounting for
environmental acoustic quality can be regarded as a significant issue both in
terms of recovery and prevention.
Also the normative aspects can provide an essential contribution, in addition
to the administrative ones, with specific reference to the prevention of further
problems (i.e. the European Directive of 1996 about IPPC “Integrated Pollution
Prevention and Control”, including noise pollution problems).
The normative horizon of the improvement plan is deployed through a
coherent program involving the local administration government, also including
integrative instructions for building codes, health and safety issues and municipal
police regulations about “noisy activities”.
In fact, as far as the building codes are concerned, the building codes refer to
the passive acoustic requirements to be specified as a function of use,
construction characteristics, infrastructural situation, and acoustic impact
documentation to be attached to the authorization request.
The integration of the operative technical specifications into the city plan
plays a major role in territory government, through regulations finalized to the
implementation of the acoustical classification as a tool in urban and building
evolution [2].
Table 1: Scheme for the applied interaction between noise and land
managements.
Administration Laws and Operative

Planning Actions Codes Actions
Reduction of noise • City Plan (PRGC) European, • Impact Noise
emissions in the • Acoustic National and Assessment
environment Classification Plan Local laws • ZTL
(PCA) • Transport
• Mobility City Noise Mapping
Planning (PUT)
• Infrastructural
PRAC
• Private PRAC
• IPPC
Mitigation of impact • City Plan (PRGC) European, • Noise
on exposed people • Strategic Noise National and Conditions
Mapping Local laws Assessment
• Noise Exposure • Private PRAC
Assessment
Improvement of • Building Codes National laws • Passive Noise
passive protection and codes Building
performances on Assessment
receivers
4 The application of the improvement plans: case histories

The national Italian legislation (LQ 447/95 and actuation decrees) establishes the
need of municipal improvement plans and private company improvement plans
for specific sources as transportation infrastructures and production sites under
specified conditions: such conditions are identified when pre-existing problems
are to be recovered.
In particular, for the private company and transportation infrastructure
improvement plans (DM 29/11/2000) [6] the intervention is required when the
emission limits, defined by PCA, are overcome.
Differences are addressed as a function of the operational areas:

• Municipalities: the acoustic improvement plan represents a complete
process, involving different competencies, regarding the whole city
asset, town planning, administration, production and transportation
infrastructures.
• Transportation infrastructures and production sites: in case of
environmental acoustical pollution generated by a source are committed
to restore the pre-existing compatibility acting on the source (when
sustainability is not achievable) with specific containment actions.
As far as the municipal plans are concerned, the authority intervention is
mandatory when the warning limits are overcome (LQ 447/95, art. 2, comma 1,
lett. G) which in long term perspective coincides the emission limits of the PCA
or in case of borderline connection between urban areas having quality levels
with more than 5 dB (A) discrepancy. A case history of the last condition is
showed in the Figure 1, where a discrepancy is caused by an industrial plan into
the residential zone.
Figure 1: Example of the urban area where PRAC is mandatory: case of

borderline connection between urban areas having in PCA quality
levels with more than 5 dB (A) discrepancy, caused by an industrial
plan into the residential zone.
The plans are also managed by different authorities who are required to
deliver a fully exhaustive result homogeneous across the whole territory.
From this point of view, it is easier to check the need and the validity of a
production site improvement plan compared with the plan for a transportation
infrastructure where is far more difficult to define the right counterpart authority:
at municipal level the local government is competent for the whole territory
except the pertinence bands of the transportation infrastructures.
As mentioned, the municipal improvement plan process can be activated by
the administrations in case of need (limits overcoming or acoustically non
compatible boundary conditions) but also when an acoustic quality target is to be
achieved. This characteristic is in line with and anticipates the UE regulation,

which supports the need to preserve the good acoustical areas.
Indeed, the authority intervention is discretional when the local administration
wants to reach a fixed quality level, even if the warning thresholds are not
overcome or he needs to co-ordinate and to manage different private or/and
infrastructural acoustical improvement plans (i.e. Figure 2).
Acoustic classification plan
Industrial improvement plan Railway improvement plan
Figure 2: Case history of a discretional PRAC of an urban area: the

residential part of the town does not need any improvement actions,
but the industrial and the railway noise sources require specific
acoustical improvement plans.
The Italian legislation addresses the general PRAC contents and the activation
procedures and criteria are still undefined. The PRAC is by definition a dynamic
instrument whose effectiveness depends on the checking and updating work to
be carried out according with a pre-defined timing, to allow the efficient revision
of the achieved results (and their validity) and of course to monitor the on going
activities.
The main open points refer to inter-functional relationship with the
transportation infrastructures and private companies and the correlation
mechanisms with the other town planning and territory management tools.
The Italian law deals with the transport noise pollution [4], allowing in the
pertinence band of roads and railways (until 100 – 250 m of wildness from the
centre of the infrastructure line) emission values higher than safety noise limits
as defined in the PCA.
Therefore, from the legal point of view, if the noise does not exceed the legal
infrastructure noise limits, the PRAC is not mandatory, also in case of strong
noise exposure and pollution.
However, if the objective is to reach the safety conditions of people exposure,
by respecting the noise limit values as indicated by WHO, the PRAC is needed
and is a precondition in any situation of noise pollution, even if the legal limits
are formally respected.
In figure 3 is showed a case history of an urban area where one national road,
one highway and one railway generate, in particular in the night period, the noise
values incompatible with the people safety conditions, without exceeding the
Italian law limits for transport noise. In this case PRAC is not legally mandatory,
but is the precondition to reach the required improvement in terms of people
noise exposure.
Road, highway and

Acoustical railway noise
City Plan Classification Plan pertinence bands
Figure 3: A case history of an urban area where one national road, one
highway and one railway generate, in particular in the night period,
the noise values incompatible with the people safety conditions,
without exceeding the Italian law limits for transport noise. In this
case PRAC is not legally mandatory, but is the precondition to
reach the required improvement in terms of people noise exposure.
5 How to integrate the national and UE action plan levels?

It is evident that, in urban areas strongly affected by the transportation noise
pollution (i.e. figure 4), no effective short/medium term prevention are actually
achievable and therefore it becomes mandatory to manage the territory
development together with the infrastructures to and obtain at least long term
improvements in population exposure conditions. This approach had been
adopted by UE requiring jointed actions from the national communities.
Road, highway and railway noise pertinence band

( ____60dB(A); ____ 70dB(A))
Figure 4: A case of an urban areas strongly affected by the transportation

noise pollution where it becomes mandatory to manage the territory
development together with the transport actions.
The tool for acoustic pollution and the associated problems management is
defined in the action plan, which must be developed to account for urban areas
and for the transportation infrastructures, with a timing plan depending on the
dimensions of the involved area.
The UE standards do not contain a detailed specification of a plan, but
underline the priorities whose identification depends on limits overcoming. The
application of such a tool is related with the strategic mapping which is the
knowledge instrument intended to define the acoustic situation.
The relationship between the National and the European tools is realized with the
acknowledgment of the DIR 2002/49/CE [5] through the legislative decree 19
AGOSTO 2005, N.194 [6].
Such a decree defines the competencies and the procedures for the elaboration
of the acoustical mapping and of the acoustical strategic maps, as fundamental
instruments to define the existing acoustic situation.
The Acoustical Strategic Maps is a map finalized to the definition of the
global exposition to noise in a certain area due to various noise sources, or to
define the general forecast for such an area.
The Acoustic Mapping (which is required for transportation infrastructures)

consists in the representation of data describing an existing or forecasted noise
situation in an area, generated by a specific source, as a function of an acoustic
parameter showing the overcoming of the applicable limits together with the
number of exposed people or the number of houses exposed to a determined
level.
From the above definitions it is clear that the acoustical strategic map
represents the noise emissions while the acoustic map describes the acoustical
emission of specific sources.
The acoustic mapping and the strategic acoustical maps finally represent data
describing an existing noise situation and the relevant exposure: both types can
be presented as graphs or diagrams.
The directive application field appears different from the Italian laws: if,
according the national law, it is necessary to design healing plans for all the
Municipalities (consequent to the application of the acoustic classification of the
territory) in the European normative action plans are demanded for the big
agglomerates and for the transport infrastructures.
Differences are minimal for the infrastructures, for which a study of the
acoustical emission (also in term of acoustical map) and a healing plan is
required: the European directive doesn’t include the road infrastructures < 3000
vehicles/year and the rail infrastructures < 30.000 passages/year.
There are important differences between European and Italian law
considering about the environment of life, because the Italian law includes all the
towns in the necessity of the healing plans, but are included in the European
Directive only the agglomeration having population in excess of 100.000 persons
and a population density that the Member State considers it to be an urbanised
area.
The distribution of the Italian population, divided in little or medium
dimension town, should implicate a reduced impact for the European norm, with
many realities excluded by an intervention of acoustical healing.
In spite of what above stated, the action plans defined by UE should
constitute the guide lines about the transport noise analysis for what refers to the
acoustical improvement plan (PRAC) and represent an important integration of
the land management in all urban areas.
6 Conclusions
The aim of the European Directive and of the Italian law shall be to preventing
and reducing environmental noise where necessary and to preserving
environmental quality where it is good.
The analysis of the environmental noise and its representation in term of
maps, the evaluation of the population noise exposure are the novelty in the
Italian regulation.
Because the objective is to reach the safety conditions of people exposure, by
respecting the noise limit values as indicated by WHO, the acoustical
improvement plan (PRAC) is needed and precondition in any situation of noise

pollution, even if there is the formal respect of the legal limits.
The healing plan, beyond the content of the single measure, express the theme
of the acoustical quality in the local administration’s culture and it represents a
systematic instrument of political addresses and concrete actions, to support the
choices of territory management and its activities.
In fact, a territorial planning integrated with the acoustical management
instruments, through a correct urbanite planning, should permit to eliminate from
the beginning the critical situations. The healing plan is not an instrument to
improve compromise situations or a set of passive protection interventions on the
buildings.
These solutions are obviously necessaries where an incorrect urbanite
planning, in acoustical terms, has generated situations solvable only with a
rilocalization, not easily realizable.
References
[1] ANPA Agenzia Nazionale Per La Protezione Dell’Ambiente “Linee guida
per l’elaborazione di piani comunali di risanamento acustico”, a cura di –
ANPA , APPA Bolzano, APPA Trento, ARPA Emilia Romagna, ARPA
Liguria, ARPA Valle d’Aosta, ARPA Veneto, ARPA Toscana, Regione
Lombardia, 1998
[2] Alberto Muratori “Piani di risanamento acustico: dimensione
amministrativa, pianificatoria e normativa”, Convegno Nazionale “I piani di
risanamento delle aree urbane”, Modena 22 – 23 febbraio 1999, in Atti
pp 3- 17
[3] Jacopo Fogola, Rosario Romano “Piani d’azione e piani di risanamento
acustico”, Convegno Nazionale “La direttiva 2002/49/CE: Determinazione e
gestione del rumore ambientale e suo impatto sulla legislazione
italiana”Pisa, 18 novembre 2004, in Atti pp. 71-81.
[4] Decreto del Ministero dell’Ambiente 29/11/2000 “Criteri per la
predisposizione, da parte delle società e degli enti gestori dei servizi pubblici
di trasporto o delle relative infrastrutture, dei piani degli interventi di
contenimento e abbattimento del rumore”
[5] Direttiva 2002/49/CE del Parlamento Europeo e del Consiglio relativo alla
“Determinazione e gestione del rumore ambientale”
[6] Decreto legislativo 19 agosto 2005, n.194 “Attuazione della Direttiva
2002/49/CE relativa alla determinazione e alla gestione del rumore
ambientale”
Estimation of the on-farm-costs of soil erosion

in Sleman, Indonesia
A. Möller & U. Ranke
Federal Institute for Geosciences and Natural Resources (BGR),
Hannover, Germany
Abstract
Soils are non-renewable resources. World-wide in many regions a sustainable
use of soils is endangered through anthropogenic accelerated soil erosion. From
the economic point of view erosion protection is the transfer soil use potential
into the future. However, in developing countries such as Indonesia usually only
short term profit counts and consequently soil resources suffer from accelerated
exploitation. Advising farmers that soil erosion protection measures not only
ensure a prolonged agricultural potential for the future, but can also include
economic benefits is a promising attempt to promote the use of soil erosion
protection measures. The prerequisite for the decision process on a farm level is
the possibility to estimate the costs of soil erosion. In Indonesia mostly reliable
data are missing, and also no expensive surveys can be accomplished. Therefore,
less data intensive methods such as the “replacement cost” or the “productivity
change” method were used to estimate the “on-farm-costs” of soil erosion in
Sleman on Java. The “replacement cost” method resulted in clearly higher costs
compared to the “productivity change” method. This is due to an over-estimation
of the costs by the “replacement cost” method. The use of both methods
comparing costs and benefits of soil protection measures indicate similar
decision guidelines. However, more information is necessary on the additional
effects of soil conservation and political constraints to be a base for sound
decision-making on a farm level, but making information available on the
benefits of conservation measures helps farmers in their decision process to
invest in soil conservation. Beyond this, soil erosion is also a societal problem,
including external costs making up a large portion of the economic effects of soil
erosion.
Keywords: soil erosion, on-farm costs, replacement cost method, productivity
change.
doi:10.2495/GEO060061
1 Introduction
Land conversion within the developing world is occurring at an unprecedented
rate. Expansion of subsistence farming practices in the form of field crop
agriculture and pasture within rural areas is contributing significantly to
ecological alteration in many tropical countries (Uhl et al. [1]; Landa et al. [2];
Lubchenco [3]).
Soil erosion is thereby a major environment threat for the sustainability and
productive capacity of agriculture. During the last 40 years, nearly one third of
the world’s arable land has been lost by erosion and continues to be lost at a rate
of more than 10000 million hectares per year (Pimentel et al. [4]).
In Indonesia more than 50% is mountainous and consequently highly
vulnerable towards soil erosion. On Java about 23.7 million hectares are stated to
be “critical” land concerning soil erosion.
Erosion adversely affects soil quality and productivity by reducing nutrients,
infiltration rates, water-holding capacity, organic matter, soil biota, and soil
depth. Several studies have shown a reduction of soil productivity in the long
term between 2 and 70% for many soils (Wolman [5]). The main reduction of
soil productivity in the short term is thereby due to the loss of soil nutrients and
water availability. In the long term the loss of soil depth, water holding capacity
and organic matter can contribute largely to the loss of soil productivity.
While it is widely accepted that erosion lowers agricultural productivity,
there is little agreement on exactly how productivity is related to erosion or on
the quantitative impact of erosion on yields (Magrath and Arens [6]). Erosion
involves changes in the availability and relative concentration of nutrients for
plant growth and changes in the soil structure which influences root growth and
affects the availability of water.
On the other hand, soils get differently affected by erosion based on their
individual fertility. They range form soils whose natural fertility is accumulated,
along with the soil organic matter, in the top few centimetres only, to soils being
fertile throughout the whole profile. Furthermore, different crops get differently
affected by a potential loss of soil productivity. Demanding crops may react with
high yield reductions, while non-demanding crops like Alfalfa may only be little
affected.
In measuring the on-site costs of soil erosion the main objective is usually to
estimate the present value of net income lost through excessive (i.e. sub-optimal)
soil erosion. According to Barbier [7], to be an economic cost, the onsite costs of
soil erosion must be an opportunity cost, which is defined as the value of a
forgone alternative like the investment in soil conservation. Because soil
conservation is not costless, the on site cost of soil erosion must be the loss in the
long-run net profitability of the farming system not investing in soil
conservation, providing of course that such an investment is an economically
worthwhile alternative. The on-site costs of soil erosion are than the difference
between the net returns of the farming system with soil conservation and the net
returns with erosion (Barbier [7]).
Although the methodology seems to be the best choice to estimate the on-
site costs of soil erosion, it has often proven to be very difficult to implement
empirically. It is not easy to determine an economically viable alternative
conservation investment to current erosive practices. Particularly, in developing
regions such as Southeast Asia, with diverse and heterogeneous small-scale
farming systems, the data constraints are often enormous, whereas simplifying
assumptions and generalizations may be misleading.
Therefore, the On-site costs of soil erosion were estimated in this study
based on alternative empirical models determining changes in soil productivity,
or the costs to replace the lost nutrients, water, eroded topsoil or organic matter
(Kim and Dixon [8]; Magrath and Arens [6]; Dixon et al. [9]; Gunatilake and
Vieth [10]; Krausse et al. [11]).
These approaches may by less reliable or even second-best from an
economic perspective, but they were the only implementable choices based on
the data available. Especially the data required for the replacement cost approach
are easier to generate in developing countries.
2 Site description
The Kabupaten Sleman is situated in the northern part of the province of
Yogyakarta (Figure 1). It is located at the southern flank of the Merapi volcano
starting almost from see level in the south to the top of the volcano at an altitude
of 2986 m.
The Climate in Sleman is humid tropical with a distinct dry season from Mai
until October. The average annual rainfall rages between less than 1500 mm and
3300 mm. Heavy rainfalls with more than 100 mm per day or within a period of
three days are common. These storms are a major driving force of heavy soil
erosion and can trigger Lahars (mud flows) at the upper slopes of Mt. Merapi.
Figure 1: Topographic map of the Province of Yogyakarta on Java.
Topographically Sleman can be divided into two major areas, the upper and
medium slopes of the Mt. Merapi and the foothill area below. The upper slope
represents the young Merapi cone with devastated to bare land and some thin
gullies. The medium slopes comprise a complex of old and young Merapi
products forming deep valleys, mainly in pyroclastic deposits. Smooth slopes
and shallow valleys characterize the foothill area consisting of fluvial deposits.
The major soils according to US soil classification in the area are Andisols
and Mollisols at the upper slopes of the Mt. Merapi and Inceptisols at the
medium slopes and the floodplains of Mt. Merapi. Some Entisols and Vertisols
can be found in the mountainous areas in the south of Sleman.
Land use in Sleman is stamped by a vertical zonal distribution according to
the increasing slope of the Merapi volcano. At lower slopes from about 0 to 5%
paddy fields dominate the mostly agricultural used area. From about 5 to 10%
slope mainly rain feed agriculture of vegetables and cash crops can be found.
Weather at slopes greater than 10 % the land use is dominated by forest, shrubs
and grassland.
3 Estimating soil erosion costs

To estimate the economic significance of soil erosion its physical dimension has
to be determined, and linked and valued to changes in crop production and
farming systems.
In this study soil erosion was determined using the Universal Soil Loss
Equation (USLE). The model is widely used in science to estimate soil erosion at
multiple scales (Renard et al. [12]; Turnage et al. [13]), and also commonly used
with some adoptions under tropical conditions (Millward and Mersey [14];
Wiriosudarmo and Bisri [15]; El-Swaify [16]).
3.1 Replacement cost method
The so-called replacement cost approach estimates the forgone input which is
necessary to overcome the negative effects of soil erosion (Kim and Dixon [8];
Dixon et al. [9]; Gunatilake and Vieth [10]; Krausse et al. [11]).
Usually only the fertilizer replacement as major costs is considered. Thus,
the replacement costs can be seen as the costs to replace the lost nutrients and the
additional energy, maintenance and labour work to apply the extra fertilizer to
the fields. In the ith land use of an area it can be presented as eqn (1):
RCi = ( St − S (t +1) )∑ N ij Pj + Cil + Cir (1)
i = 1...n, j = 1...k
where:
RCi is the replacement cost of nutrients in ith category of land use, Rp/ha
St − S ( t +1) ist he soil loss from time t to t+1, t/ha
N ij is the quantity of jth nutrient in ith land use type, kg/t

Pj is the price of jth nutrient, Rp/kg
Cil is the cost of labor in spreading fertilizer, Rp/ha
Cir is the cost of repair and maintenance of damages due to soil erosion Rp/ha
Included in the calculation of the replacement costs are the major nutrient
Nitrogen, Phosphorous and Potassium. Nutrient levels of eroded soil are usually
not available. Therefore it is assumed that the nutrient level in the eroded soil
and the farm soil are the same. Based on the assumption the amount of nutrients
lost is calculated using representative soil nutrient analyses from the farm soils.
3.2 Productivity change method
The change in productivity approach determines the difference in crop yields

with and without erosion, multiplied by the unit price of the crop, and less the
variable costs of the production (Magrath and Arens [6]; Gunatilake and Vieth
[10]). Although this seems straightforward and simple, in practice the
quantification of the effect on crop yield losses is conceptually difficult. In this
study the approach of Magrath and Arens [6] introduced for Java was used to
quantify the changes in productivity.
They assumed that if output falls farmers adjust variable inputs in
production to yield declines and that fixed costs remain fixed. Percentage
productivity declines are denominated base on the response of sensitive and less
sensitive crops. The result of this procedure is a linear decline in profits as
productivity falls. To account for possible adjustments in cropping systems, farm
budgets for a variety of representative dry land cropping systems across Java
were constructed, and then used to estimate the effects of the yield losses from
erosion on net farm incomes. This was done comprehensively for a single year.
They have estimated an average yield reduction on Java between 4–7%,
depending on soil type and crops planted. Using their results the costs for
reduced yields were calculated by relating the lost yields to the average cross
margin of the agricultural production in the region, based on prices of important
cash crops and vegetables in the District Yogyakarta in 2002.
4 Results and discussion
4.1 Soil erosion
On most of the cultivated area in Sleman bench terraces or at least raised bed
terraces are used. This is reflected by low erosion rates < 5 t ha-1 yr-1 (Figure 2).
At the upper part of the Merapi volcano at areas with steep slopes and badly
maintained or no terraces erosion rates exceed by fare a sustainable level. Here
erosion rates with more than 100 t ha-1 yr-1 can be found at cultivated areas.
Figure 2: Map of the average annual soil erosion in the District of Sleman.
Areas with particularly high erosion rates at the top of the volcano with up to
500 t ha-1 yr-1 are not cultivated, but consist of loose pyroclastic deposits
triggering debris flows during heavy rains in the rainy season.
Similar erosion rates were found in other studies carried out on Java.
Magrath and Arens [6] has estimated for steep slopes on Java erosion rates up to
500 t ha-1 yr-1 with an average of 123 t ha-1 yr-1 for “Tegal” land use (rain fed
agriculture) and 87 t ha-1 yr-1 for degraded forests. Kusumandri and Mitchell [17]
found soil erosion rates for the Citarik watershed on West Java to be about 100 t
ha-1 yr-1.
Although at most of the agricultural land in Sleman already some soil
conservation is practiced, some areas at the upper slopes of the volcano and at
some hilly parts in the south have significant soil erosion problems with more
than 15 t ha-1 yr-1. The area affected is 2737 ha with an average soil erosion rate
of 30 t ha-1 yr-1 (Figure 3). These areas took center stage in the estimation of the
costs of soil erosion in Sleman.
4.2 Soil erosion costs
The costs to replace lost nutrients in these agricultural areas were estimated to be
14100 Rp t-1 soil, taking into account an average soil nutrient content of N = 1.1
kg t-1, P = 0.8 kg t-1, and K = 3.7 kg t-1 and prices of common fertilizers in 2003
(Urea = 1000 Rp; TSP = 1000 Rp; KCl = 1200 Rp). Thus, the replacement cost
with an average soil erosion rate of 30 t ha-1 yr-1 is 423000 Rp ha-1 yr-1. The
additional costs for energy, maintenance and labour work were estimated based
on results from literature and expert experience to be 85000 Rp ha-1 yr-1 or about
20% of the replacement costs. Thus, the total costs were estimated to be 508000
Rp ha-1 yr-1. However, the approach overestimates soil erosion costs, based on
the conceptual assumption estimating the difference between erosion and “zero”
erosion and the assumption that all nutrients lost would be available for plants in
the long term, which is in reality in agriculture not realizable. On the other hand
other effects like the loss of organic matter or water holding capacity are not
considering.
On the bases of the change in productivity approach of Magrath and
Arens [6] the average productivity loss at agricultural areas with significant
erosion was calculated to be 160000 Rp ha-1 yr-1. This seems to underestimate
the actual costs of soil erosion. However, Magrath and Arens [6] “capitalized”
the one year cost of erosion by a factor of 10 to obtain a total present value of
current and future losses, assuming that one year loss in net income recurs over
each successive year. On the other hand considering only plant available
nutrients lost (N = 1.1 kg t-1; P = 0.45 kg t-1; K = 0.8 kg t-1) within the
replacement cost approach the average costs are comparable, with 250000 Rp ha-
1
yr-1 (Figure 3). These costs are equal to ≈ 17 % of the average farmers net
income per ha agricultural land.
A similar order of magnitude of soil erosion costs was reported by Krausse
et al. [11] and Gunatilake and Vieth [10] for agricultural soils in New Zealand
and Sri Lankan high land soils, respectively. Krausse et al. [11] estimated the
actual costs for soils suffering significant erosion in New Zealand to range
between Aus$ 8 and Aus$ 25 ha-1, with an average erosion rate of 10 t ha-1 yr-1.
Considering the average erosion rate of eroded soils in Sleman (30 t ha-1 yr-1) this
would approximate between 135000 Rp and 420000 Rp ha-1 yr-1. Gunatilake and
Vieth [10] estimated slightly higher soil erosion costs depending on the type of
crop and the erosion rate with e.g. ≈ 200000 Rp (21-25 USD) for paddy fields
and ≈ 600000 Rp (68 USD) for market gardens.
On the other hand to minimise soil erosion rates soil conservation measures
are necessary, which are not cost less. Adiningsih and Karama [18] estimated the
additional annual costs for bench terraces and raised bed terraces compared to
conventional farming practices in East Java to be 113 USD (≈ 1 million Rp) and
56 USD (500000 Rp), respectively. Thus, the costs for terracing are not covered
by the benefits from reduced erosion alone. Other potential benefits like the
possibility to intensify/change the agricultural production coming along with
conservation measures have to be considered as well. Quantifying these benefits
is very difficult, but point based studies comparing the net income of farms with
and without conservation measures indicate that adequate soil conservation
measurements can be economically worthwhile (Adiningsih and Karama [18];
Posthumus and De Graaff [19]). Adiningsih and Karama [18] showed that the net
income of the farmers was by a multiple higher after changing to an integrated
farming system with terraces.
0.125 - 0.25
0.25 - 0.5
0.5 - 0.75
0.75 - 1.0
1.0 - 1.5
1.5 - 2.0
over 2.0
Figure 3: Map of the On-farm cost of erosion (at agricultural areas with
significant soil erosion [million Rp ha-1a-1]).
Adventitiously, other decision factors like the availability of credits or the

fact that in many countries soil conservation is not reflected in land prices makes
it difficult or impossible for a farmer to decide if it is worthwhile to invest in a
certain soil conservation measure. Nevertheless, information available on the
costs of soil erosion and on possible benefits is an important economic factor to
help farmers in their decision to invest in soil conservations measure.
Besides, not reflected in farmer’s decision-making are off-site or external
costs of soil erosion, but they play an important part of the economic impact of
soil erosion. In many studies off-site cost are estimated to be higher than on-site
costs. About these costs, which are not reflected by the markets prices, the
society has to be concerned and against the background that decisions on pure
economic basis usually only consider the next maximum 50 years, but
sustainability of soil resources is a matter of the next centuries, soil conservation
has to be a general goal for the society and can’t be shouldered by the farmers
alone. However, it is not easy to design appropriate policies to include the off-
site or external costs into the decision-process of soil conservation on farm level.
Acknowledgements
The project was funded by the Federal Minister for Economic Cooperation and
Development, Germany. The Authors acknowledge the contribution of the
Directorate General of Geology and Mineral Resources of Indonesia to the
project.
References
[1] Uhl, C.J., Carl, C., Clark, H., Herrera, R., Ecosystem recovery in Amazon
Caatinga forest after cutting, cutting and burning, and bulldozer clearing
techniques. Oikos 38(3), pp. 313–320, 1982.
[2] Landa, R., Meave, J., Carabias, J., Environmental deterioration in rural
Mexico: an examination of the concept. Ecological Applications 7(1), pp.
316–329, 1997.
[3] Lubchenco, J., Entering the century of the environment: a new social
contract for science. Science 27(5350), pp. 491–497, 1998.
[4] Pimentel, D., Harvy, C., Resosudarmo, P., Sinclair, K., Kurz, D., McNair,
M., Crist, S., Shpritz, L., Fitton, L., Saffouri, R., Blair, R., Environmental
and Economic Costs of soil Erosion and Conservation Benefits. Science
267(5126), pp. 1117-1123, 1995.
[5] Wolman, M.G., Soil erosion and crop production: A worldwide
perspective. Soil erosion and crop productivity, eds. R.F. Follett & B.A.
Stewart, ASA, CSSA & SSSA: Madison, pp. 10-22, 1985.
[6] Magrath, W. & Arens, P., “The Costs of Soil Erosion on Java: A Natural
Resource Accounting Approach”, Environment Department Working
Paper No. 18, Washington D.C.: The World Bank, 1989.
[7] Barbier, E.B., The Economics of Soil Erosion: Theory, Methodology and
Examples. Special Papers, Fifth Biannual Workshop on Economy and
Environment in Southeast Asia, Singapore, 1995.
[8] Kim, S.H., Dixon, J.A., Economic valuation of environmental quality
aspects of upland agricultural projects in Korea. In: Doxon, J.A.,
Hufschmidt, M.M. (eds.). Validation techniques for the environment: A
case study workbook. Baltimore: Johns Hopkins University, 1986.
[9] Dixon, J.A., Scura, L.F., Carpenter, R.A., Sherman, P.B., Economic
analysis of the environmental impacts. London, Earthscan Publication
Ltd, 1994.
[10] Gunatilake, H.M., Vieth, G.R., Estimation of On-site Cost of Soil Erosion:
A Comparison of Replacement and Productivity Change Methods. Journal
of soil and water conservation 55(2), pp. 197-204, 2000.
[11] Krausse, M., Eastwood, C., Alexander, R.R., Muddied waters Estimating
the national economic cost of soil erosion and sedimentation in New
Zealand. Manaaki Whenua landcare research, Palmerston North, 2001.
[12] Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K., Yoder, D.C.,
Predicting soil erosion by water: a guide to conservation planning with the
Revised Universal Soil Loss Equation RUSLE. Handbook No. 703. US

Department of Agriculture, 404 pp., 1997.
[13] Turnage, K.M., Lee, S.Y., Foss, J.E., Kim, K.H., Larsen, I.L., Comparison
of soil erosion and deposition rates using radiocesium, RUSLE, and
buried soils in dolines in East Tennessee. Environmental Geology, 29, pp.
1-9, 1997.
[14] Millward, A.A., Mersey, J.E., Adapting the RUSLE to model soil erosion
potential in a mountainous tropical watershed. Catena, 38, pp. 109-129,
1999.
[15] Wiriosudarmo, S., Bisri, D.H., Use of environmental geological
information to plan erosion control in the upper part of the Citanduy
catchment, West Java province, Indonesia. Sixth Regional Congress on
Geology, Mineral and Hydrocarbon Resources of Southeast Asia, 1987.
[16] El-Swaify, S.A,. Susceptibilities of certain Tropical Soils to Erosion by
Water. In: Greenland, D.J. and Lal, R. Soil Conservation and Management
in the Humid Tropics. John Wiley & Sons, New York, 1977.
[17] Kusumandri, A., Mitchell, B., Soil erosion and sediment yield in forest
and agroforestry areas in West Java, Indonesia. Journal of Soil and Water
Conservation, 52(4), pp. 376-380, 1997.
[18] Adiningsih, J.S., Karama, A.S., A sustainable upland farming system for
Indonesia. http://www.fftc.agnet.org/library/abstract /eb354b.html, 1992.
[19] Posthumus, H., De Graff, J., Cost-benefit analysis of bench terraces, a
case study in Peru. Land Degradation & Development, 16(1), pp. 1-11,
2004.
Section 2
Environmental modelling
and monitoring
Large scale soil erosion modeling for

a mountainous watershed
P. Zhou1, J. Nieminen2, T. Tokola 2, O. Luukkanen1 & T. Oliver3
1
Viikki Tropical Resources Institute, University of Helsinki, Finland
2
Department of Forest Resource Management,
University of Helsinki, Finland
3
University of Barcelona, Spain
Abstract
Soil erosion control requires a quantitative evaluation of potential soil erosion on
a specific site. The Revised Universal Soil Loss Equation (RUSLE), Remote
Sensing (RS), and Geographic Information System (GIS) were used to model
soil erosion intensity for soil conservation and vegetation rehabilitation in an
Upper Min River (UMR) watershed, which is in the Upper Yangtze River basin.
Data used in this study to generate the soil loss were Landsat Enhanced Thematic
Mapper (ETM) images, Digitized Elevation Model (DEM), soil erodibility,
rainfall erosivity, and inventory data. The non-parametric k-nearest neighbor (k-
NN) method was used to produce the cover management map by integrating the
ETM images and vegetation coverage data measured in the 625 sample plots.
The root mean square errors and significance of biases at pixel level were
evaluated in order to find optimal parameters. Four raster maps have been
produced for the soil erodibility, rainfall erosivity, slope length and steepness,
and cover management factor, and the map with different soil loss risks has been
produced for soil erosion potential. The result can be beneficial to the erosion
control and ecological restoration in the degraded mountainous watershed.
Keywords: soil erosion, RUSLE, DEM, k-NN method, Upper Min River
Watershed.
1 Introduction
Soil erosion is a worldwide environmental problem that degrades soil
productivity and water quality, causes sedimentation and increases the
doi:10.2495/GEO060071
probability of floods. The 1998 flood in the Upper Yangtze raised public
attention to the problems of soil erosion and sedimentation. Soil loss control calls
for a quantitative evaluation of potential soil erosion on a specific site. Scientists
have studied different methods to assess soil erosion loss by water, for instance,
universal soil loss equation (USLE) [1], revised soil loss equation (RUSLE) [2],
the 137Cs technique [3], and WEPP Hillslope model [4]. Among them, the USLE
and RUSLE provided a convenient tool for soil loss evaluation by taking the
climate, geographical terrain, conservation support practice, soil, and vegetation
into considerations. The RUSLE was developed to incorporate the considerable
amount of erosion information and to address specifically the application of the
USLE to land uses other than agriculture [2]. The model can be used to any
geographic region by modifying its factors. The factors are rainfall runoff
erosivity factor, soil erodibility factor, slope length and steepness factor, cover
management factor, and support practice factor [1, 2]. An important reference
manual for applying the USLE to disturbed forestlands is “A guide for predicting
sheet and rill erosion on forest land” [5]. The USLE and RUSLE have been used
widely in evaluating the soil erosion risk in watershed and highland [6, 7, 8].
Slope length and steepness factor, which reflects the terrain on a given site,
can be computed from the digital elevation model (DEM) [9, 10]. Rainfall and
runoff erosivity factor was calculated based on the storm events and rainfall data
in many studies [6, 7, 8, 11]. However, in mountainous watershed, orographic
effects caused by mountainous terrain can result in a significant positive
correlation between precipitation and elevation [12, 13, 14]. In the UMR
watershed, precipitation tends to increase with an increase in elevation because
of the orographic effect of mountainous terrain and the foehn effect [15]. In our
study area, we examined the relationship between elevation and precipitation for
38 stations, and selected cokriging as a method for estimating average annual
precipitation of the whole watershed.
Cover management C-factor in the soil loss equation was defined as the ratio
of soil loss from land cropped under specified conditions from the corresponding
loss from clean tilled, continuous fallow [1]. However, in large scale UMR
watershed, where are not mainly covered by agricultural lands, the cover
management factor is not only affected by the agricultural crops. Ma et al. used
the proposition of vegetation reflectance in pixel end members to calculate C
factor [7]. In our study, we used the non-parametric k-nearest neighbour (k-NN)
multi-source estimation method to estimate coverage data and produce the
coverage map by integrating the satellite images and field data with optimal
parameters. The k-NN method has been widely used in a variety of forest
estimation and biomass mapping applications over the years [16, 17, 18, 19], and
therefore, can be applied in vegetation cover estimation.
Lu et al. [20] explored the relationships between the soil erosion and land use
and land cover distribution, they found that most climax and mature forests are
in low erosion risk areas, while agroforestry and pasture are usually associated
with medium to high risk areas. A good plant cover is generally capable of
preventing surface erosion, and reducing landslides as well. Removal of
vegetation can greatly increase runoff and soil erosion particularly in
mountainous areas [21]. Soil erosion control especially calls for the forest
restoration or rehabilitation to reduce the erosion loss and improve soil stability.
The questions we are going to answer in the paper are: (1) How to model the
soil erosion loss in this mountainous watershed? and (2) How much is the soil
erosion risk in the area?
2 Study area
The Upper Yangtze River Basin is a mountainous region, which has an area of
1.04×106 km2, a mean annual runoff discharge of 4.35×108 m3, a mean sediment
yield of 5.17×108 t and a population of 1.4×108 [22]. The basin is one of the most
severely eroded areas in China. Water erosion results in both on-site soil
degradation and off-site problems related to downstream sedimentation [23].
The Upper Min River, which is one of the most important tributaries of the
Upper Yangtze River, is 341 km long with a drainage area of 23,037 km2. The
watershed is located in Sichuan Province, South West China. The area is
governed by the southeast and southwest monsoons. The complex topography,
with elevations ranging from 900 m to 5 700 m, results in steep gradients of
rainfall. The Upper Min river watershed has been divided into five ecozones: the
Sub-tropical (1300–2200 m), Temperate (2200–2600 m), Sub-alpine (2600–3200
m), Boreal (3200–3600 m) and Arctic zone (3600–5700 m) [24]. At present, the
forest cover is around 21% of the whole watershed area. Our 625 inventory plots
were randomly placed in the middle and upper reaches of the UMR watershed,
between 31º-34º N, 103º-104º E, with an area of about 7 400 km2, see fig. 1. The
vegetation ranges from subtropical evergreen broadleaved forest to the alpine
meadows.
3 Method
3.1 Model structure
The soil loss (A) due to water erosion per unit area per year (Mg ha-1yr-1) was
quantified using RUSLE by the following equation:
A = R × K × LS × C × P (1)
where A is the average soil loss due to water erosion, R the rainfall and runoff
erosivity factor (MJ mm ha-1h-1yr-1), K the soil erodibility factor (Mg h MJ-1
mm-1), L the slope length factor, S the slope steepness factor, C the cover and
management practice factor, and P the support practice.
3.2 Data and processing
3.2.1 Rainfall and runoff erosivity factor (R)

R is the long term annual average of the product of event rainfall kinetic energy
(E) in MJ ha-1 and the maximum rainfall intensity in 30 minutes (I30) in mm h-1.
The R values were correlated with annual precipitation [1, 25]. We used the
following equation to calculate the R factor, which has been adopted for
application in the RUSLE model [2]:
2
R = −0.0334 Pa + 0.006661Pa (2)
where R is rainfall and runoff factor (MJ mm ha-1h-1 yr-1), and Pa the measured
annual precipitation in mm.
The average annual precipitation (APP) and elevation data from 38
meteorological stations in the research area were obtained to check the
correlation between precipitation and elevation. The APP surface was
interpolated with a multivariate geostatistic cokriging model [26]. The R factor
surface was then calculated by eqn. (2) from the APP surface using the raster
calculation in spatial analyst.
Figure 1: Upper Yangtze River watershed and sample plots in the research area.
( • )showed the 625 sample plots, ( ) showed the drainage net
work, ( ) showed the Upper Yangtze River basin, and ( )
showed the detail research area.
3.2.2 Slope length and slope steepness factors (LS)

L factor and S factor, which reflect the topographic erosion susceptibility on a
given site, were computed together from the digital elevation model (DEM). The
DEM used is based on a digital topographic map, with 100-m elevation contour
lines and stream data. In order to achieve a geomorphological realistic surface, it
was interpolated to a 25-m cellsize grid with the Topogrid algorithm [27] which
generates a hydrologically correct grid DEM using contour lines and stream data.
The slope was calculated using the maximum downhill direction method, in
which the slope value for each raster cell is obtained from the angle formed
between the cell itself and the lowest neighboring cell. The flow direction was
calculated with the D∞ (infinite directions) method developed by Tarboton [28],
by which dispersed or rilled flow is estimated for each cell from the slopes to the
lower neighboring cells. In flat areas of the DEM (no lower neighboring cells),
the method approached by [29] was used to calculate flow direction. Flow
accumulation, the number of cells contributing with its flow to each cell was
calculated from the flow direction raster. The DEM sinks filling the slope angle,
the flow direction and the flow accumulation were calculated by Taudem, an
ArcGIS 9.0 extension developed by Tarboton. For this project, an approach
developed by Moore and Burch [9, 10] was used to compute LS factor:
LS = L × S × 1.4 (3)
L = ( fa × cs / 22.13) m
(4)
where LS is computed slope length (L) and slope steepness (S) factor, fa the
flow accumulation (contributing area or upslope area), cs the cellsize, and m the
slope-length exponent, as explained in the equation:
m = β /(1 + β ) (6)
where β is the ratio of rill to the interrill erosion for conditions when the soil is
moderately susceptible to both, and is computed by θ with the following
equation [30]:
β = (sin θ / 0 .0896 ) / [3 .0 × (sin θ ) 0 .8 + 0 .56 ] (7)
Where θ is slope angle in degree. Table 1 shows the values for m computed from
eqn (6) and (7), and applied to eqn (4) to calculate a raster map for L factor. S is
calculated by the following equations:
S = 10 × sin θ + 0.03 If slope < 9 percent (8)

S = 16.8 × sinθ − 0.05 If slope ≥ 9 percent (9)
Table 1: Calculation of m value from angle.
θ angle in degree m value

θ <1° m=0.2
1° ≤ θ <2° m=0.3
2° ≤ θ <3° m=0.4
3° ≤ θ <6° m=0.45
6° ≤ θ <10° m=0.55
10° ≤ θ <26° m=0.65
26° ≤ θ m=0.7
3.2.3 Soil erodibility factor (K)

K factor is soil erodibility factor, which represents both susceptibility of soil to
erosion and the rate of runoff. Specifically, the k factor is a function of particle
size distribution, organic matter content, structure and permeability [1, 2, 11]. K
was calculated using the equation recommended by Wischmeier and Smith [1]:
K = (2.1× (vf 2 + vf × sa)1.14 ×10 −6 × (12 − om) + 0.0325× ( pe − 2) + 0.025× (st − 3)) × 0.1317 (10)
Where K is the soil erodibility factor (Mg h MJ-1mm-1), vf the percentage of very
fine sand plus silt, sa the percentage of sand, om the percentage of organic
matter, pe the permeability class, and st the structure class. In this study area, eqn
(10) was used to calculate K value of each soil type. K values and the map of soil
type were used to produce the raster map of K factor.
3.2.4 Cover management factor (C)

Vegetation cover at three levels (canopy cover, under canopy cover, surface
cover) was recorded from 625 sample plots. The canopy cover was measured by
densiometer. The k nearest neighbor (k-NN) method was used to produce the
canopy cover map and total vegetation cover map by integrating Landsat ETM+
image information and collected vegetation coverage. In this study we used two
consecutive Landsat ETM+ scenes (WRS2 130/037 & 130/038, 10th July 2002
[31]) for the cover factor estimation. A set of parameters was chosen for the k-
NN method in predicting the vegetation coverage map. The parameters were the
image features, the weight for each band, the distance, the number of nearest
neighbors the value of k, and the geographical reference area from which the
nearest field plots are selected. The leave-one-out cross validation method is
applied to calculate the root mean square errors (RMSE) and the average biases
of predictions at the single pixel level for different combination of k-NN
estimation. The RMSE and biases were used as a measure of reliability of the
continuous variables. The cover management factor (C) then was calculated from
vegetation coverage data using the equation recommended by Renard et al. [2].
CC = 1 − Fc × exp( −0.03048 × H ) (11)
where CC is the canopy cover subfactor range from 0 to 1, Fc is fraction of land

surface covered by canopy, and H (m) is distance that raindrops fall after striking
the canopy. The equation was used to calculate both the canopy cover subfactor
and under canopy cover subfactor. For the former, we used the average weighted
tree height 8.341m; and for the latter, we used the estimated shrub and grass
height 0.5 m.
3.2.5 Support practice factor (P)

P values range from 0 to 1, value 0 represents a very good manmade erosion
resistance facility, and value 1 represents no manmade resistance erosion facility.
In the study area, there are some agricultural support practices [7]. However,
most of the farmlands in the study area are small and self-managed lands, and the
spatial resolution of ETM+ imageries is 30 m, so it is impossible to distinguish
the practices in a large scale watershed from the available data. We use P = 1 for
all the lands.
4 Results
The historical precipitation data and station elevations from 38 meteorological
stations were obtained to estimate the average annual precipitation (AAP) over
the entire watershed. The AAP showed a significant (p < 0.01) correlation of r =
0.74 with the station elevation. A multivariate cokriging interpolation method
was used in the analysis since it takes into consideration the elevation which
significantly affects precipitation. Root mean square errors (RMSE) were
calculated to investigate the estimation accuracy. The RMSE by cokriging
estimation was 86.88 mm, which was reduced by 28.2% to kriging estimation
(121.2 mm). The estimated AAP was used for calculation of rainfall and runoff
erosivity R-factor in ArcGIS. The R factor varied from 1288 to 3342 MJ mm ha-
1 -1
h yr-1 (fig. 2).
The watershed occupied a raster grid space of 7700 rows by 2736 columns,
and elevations ranged from 1261 m to 5537 m. Approximately 94.7 percent of
the watershed has slopes steeper than 9 percent. Slope angles ranged from 0 to
77.2 degrees with a mean of 25.9 degrees and standard deviation of 12.0 degrees.
As a result of applying Taudem, flow accumulation ranged from 1 to 15496180
m with 98% less than 1000 m, and slope length factor ranged from 0 to 3398,
with a mean of 85.9 and 99.2% less than 120. The slope steepness S factor varied
between 0.03 to 15.88 with a mean of 6.7 and a standard deviation of 3.1.
The canopy cover map and total vegetation cover map were produced using
k-NN method. The root mean square errors (RMSE) and the average biases of
predictions at the single pixel level were evaluated for each combination of
parameters. The value of k (8), the distance (55 km), the bands (1, 2, 3, 4, 5, 7)
and their optimal weights were chosen when RMSE and bias were minimal. The
C factor was calculated from the produced raster maps by using eqn (5). The
cover management C factor ranged from 0.015 to 0.892 (fig. 2).
Figure 2: Factors for calculation of soil loss potential and the map of classified
erosion risk.
The soil erodibility K factor was between 0.036 and 0.043 Mg h MJ-1mm-1
(fig. 2). The estimated soil loss for the research area varied from 325 to 83240
Mg ha-1 per year. According to the soil loss amount and field inventory result, we
divided them into four ordinal classes: extreme risk (> 10000), high risk (3000 -
10000), moderate risk (1000 - 3000), low risk (< 10000) and No data (fig. 2). No
data values were assigned in two circumstances: firstly, excluded data with flow
accumulation values higher than 1200 m (1.7% of the cells), which are
coincident with the main stream paths; and secondly, excluded data with LS
factor values higher than 1600 (0.0023% of the cells), which happened only in
isolated cells with extremely high slopes and contributing areas. Totally 0.7% of
the cells had No data value. Table 2 showed the area and proportion of each of
the soil erosion potential categories. More than half of the watershed (58.1%)
showed moderate, high or extremely high erosion risks.
Table 2: Derivation of the ordinal categories of soil erosion potential and the
area and proportion of each category.
Numeric range Erosion potential Area (ha) Proportion (%)

(Mg ha-1yr-1)
0 - 1000 low 297007.7 40.2
1000 - 3000 moderate 306909.9 41.6
3000 - 10000 high 120109 16.3
>10000 extreme 9202.3 1.2
no data 4896.8 0.7
5 Discussion
In the UMR watershed, the average annual precipitation was positively
correlated with elevation (r = 0.74, p < 0.01), which supports similar findings in
Algarve (Portugal), southern Nevada and southeastern of California (r = 0.75, p
< 0.05) [12, 13, 14]. The effect of elevation on precipitation can be used to
improve the geostatistical interpolation. The RMSE by cokriging estimation was
reduced by 28.2% to kriging estimation in the UMR watershed, and 54%
reduction has been reported in Nevada and southeastern of California [13].
Slope calculations made with a maximum downhill method conserved the
variability and the maximum slope values. This method produced no
underestimation, since no averaging was used. Flow directions calculated by D∞
(infinite directions) improved significantly the water flow modelling, by
allowing dispersed flow to be modelled over the surface. This method calculated
the flow direction from the lowest continuous neighbouring cells and
fractionated the water flow between them, simulating dispersed water flow and
generating natural looking flow maps. Other studies mostly use the D8 approach
method, by O'Callaghan and Mark [32]. However, D8 method produces
unrealistically rilled water flow with lots of straight lines in flow accumulation
maps, because it can only produce 8 different flow directions, to one of the
neigbouring cells (cardinal or diagonal direction).
The K values of the soils in our study area ranged from 0.036 to 0.043 Mg h
MJ-1mm-1. Compared to the K values of tested soils in USLE (0.03 - 0.69 tons
acre hr/hundreds of acre ft-ton in) [1], which are from 0.004 to 0.09 Mg h
MJ-1mm-1, the soils in the UMR watershed have the moderate erodibility.
To find a suitable C-factor, a canopy cover and total vegetation surface cover
were calculated using the k-NN technique for their estimation. Several methods
were applied for topographic normalization of the imagery [33]. However, the
elevation changes in the study area were so great (1254 m - 5527 m) that not all
of the shadowing effects were removed from the imageries. To maximize
spectral variability bands 1-5 and 7 were included in the analysis. The calibration
of the k-NN parameters was performed as outlined by several articles dealing
with forest estimation using k-NN methods [27, 34, 35]. The RMSE and
significance of biases at sample plot pixel level were evaluated in order to
choose the most optimal parameters, such as numbers of k, distance, and bands
weights. The value of k (8) was chosen when the variation in bias and RMSE
were minimal. The built in cross-validation method of bias and error estimation
was applied in all calculations.
Morgan [36] argues that 10 Mgha-1yr-1 is an appropriate boundary measure of
soil loss over which agriculturists should be concerned. This was identified as
the separation of the low and moderate categories in RUSLE [2]. Soil loss in
highland conditions in Kenya ranged from 30 to 666 Mg ha-1yr-1 [8], while the
calculated soil loss (325 - 83240 Mg ha-1yr-1) in our research area is much
higher. The complex terrain, with elevations from 1261 m to 5537 m, and slopes
from 0 to 77.2 degree could be one reason to compute so high soil loss potential.
Van Remortel et al [37] argue that erosion model can be used to derive patterns
of erosion, but not necessarily the actual loss of erosion, because of the
limitations of the methods used to derive some component factor values.
Millward and Mersey [6] found that relative comparisons of soil loss among land
areas are more critical than assessing the absolute soil loss in a particular cell. A
visual interpretation and validation of the resulting erosion risk map was
performed for all the sample clusters. The sites were given a subjective risk scale
ranging from No risk – Low – Moderate – High – Extreme bases upon the
general site characteristics. The high or extreme high erosion risks mostly
occurred on the downhill gullies with long proceeding slope lengths (red areas in
the map). Some considerations should be given to the vulnerable areas, where
the landslides or mudslides could happen easily according to the soil loss
potential.
A good plant cover is generally capable of preventing surface erosion [21].
The cover management factor with a range from 0.015 to 0.892 indicated that the
loss of soil erosion can be greatly reduced by a higher vegetation cover. For the
large scale soil conservation, little work can be done to reduce rainfall and runoff
erosivity, soil erodibility, slope length and slope steepness, so vegetation
restoration and support practice would be the way to reduce the soil loss risk.
We have estimated the error from ETM+ images to the canopy cover map and
total vegetation coverage map, and calculated the interpolation error of average
annual precipitation surface. However, the model is still subjected to errors due
to the limitations of the methods to estimate some component factor values, and
the lack of possibilities on quantitatively verifying the actual erosivity from our
sites. Problems seemed to be mostly concentrated in areas with thick canopy
coverage and a high measured ground cover percentage. The erosivity
discrepancies in the forests caused by the modelled rainfall influence from the
tree canopies on the ground without taking properly into account the forest floors
ground layer. Problematic areas also include rill and valleys where the calculated
risk values seemed relatively high when compared with field experiences. This
may be explained by the methods used in calculating the LS factor. On the other
hand hill tops within the Arctic zone (3600-5400m) were somewhat
overestimated since the ETM+ cover of these regions was highly cloud covered.
The results could be improved if a cloud free fully topographically normalized
image was available, and the C-factor calculation model would take into
consideration under canopy ground coverage.
References
[1] Wischmeier, W.H. & Smith, D.D. Predicting rainfall erosion losses – a
guide to conservation planning. U.S. Dept. of Agriculture, Agric.
Handbook No. 537, 58p, 1978.
[2] Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K, & Yoder, D.C.
Predicting Soil Erosion by Water: A Guide to Conservation Planning with the
Revised Soil Loss Equation (RUSLE). U.S. Dept. of Agriculture, Agric.
Handbook No. 703, 404p, 1997.
[3] Zhang X., Zhang Y., Wen A., & Feng M., Assessment of soil losses on
cultivated land by using the 137Cs technique in the Upper Yangtze River
Basin of China. Soil and Tillage Research, 69 (1-2), pp. 99-106, 2003.
[4] Grønsten H.A. & Lundekvam H. Prediction of surface runoff and soil loss
in southeastern Norway using the WEPP Hillslope model. Soil and Tillage
Research, 85 (1-2), pp.186-199, 2006.
[5] Dissmeyer, G.E. & Foster, G.R. A guide for predicting sheet and rill
erosion on forest land. Technical Publication SA-TP-11. USDA-Forest
Service-State and Private Forestry-Southeastern Area. 40p. (Technical
report), 1980.
[6] Millward, A.A. & Mersey, J.E. Adapting the RUSLE to model soil
erosion potential in a mountainous tropical watershed. Catena, 38, pp.
109-129, 1999.
[7] Ma, J.W., Xue, Y., Ma, C.F. & Wang, Z.G. A data fusion approach for
soil erosion monitoring in the Upper Yangtze River Basin of China based
on the Universal Soil Loss Equation (USLE) model. International Journal
of Remote Sensing, 24 (23), pp. 4777-4789, 2003.
[8] Angima, S.D., Stott, D.E., O’Neill, M.K., Ong, C.K. & Weesies, G.A.
Soil erosion prediction using RUSLE for central Kenyan highland
conditions. Agriculture, Ecosystems and Environment, 97, pp. 295-308,
2003.
[9] Moore, I.D. & Burch, G. Physical Basis of the Length-Slope Factor in the
Universal Soil Loss Equation. Soil Science Society of America Journal,
50, pp. 1294-1298, 1986.
[10] Moore, I.D. & Burch, G. Modelling erosion and deposition: topographic
effects. Transactions of ASAE, 29 (6), pp. 1624-1630, 1640, 1986.
[11] Renard, K.G., Foster, G.R., Yoder, D.C. & McCool, D.K. RUSLE
revisited: Status, questions, answers, and the future. Journal of Soil and
Water Conservation, 49 (3), pp. 213-220, 1994.
[12] Hevesi, J.A., Istok, J.D. & Flint, A.L. Precipitation estimation in
mountainous terrain using multivariate geostatistics: I. Structural analysis.
Journal of Applied Meteorology, 31, pp.661-676, 1992.
[13] Hevesi, J.A., Flint, A.L. & Istok, J.D. Precipitation estimation in
mountainous terrain using multivariate geostatistics: II. Isohuetal maps.
Journal of Applied Meteorology, 31, pp. 677-688, 1992.
[14] Goovaerts, P. Using elevation to aid the geostatistical mapping of rainfall
erosivity. Catena, 34, pp. 227-242, 1999.
[15] Ma, K.M., Fu, B.J. Liu, S.L. Guan, W.B. Liu, G.H. Lü, Y.H. & Anand, M.
Multiple-scale soil moisture distribution and its implications for
ecosystem restoration in an Arid River valley, China. Land Degradation &
Development, 15, pp. 75-85, 2004.
[16] Tokola, T. The influence of ground truth data location on growing stock
volume estimation in Landsat TM-based forest inventory in Eastern
Finland. Remote Sensing of Environment, 74, pp. 422-431, 2000.
[17] Franco-Lopez, H., Ek, A.R. & Bauer, M.E., Estimation and mapping of
forest stand density, volume and cover type using the k-Nearest Neighbors
method. Remote Sensing of Environment, 77 (3), pp.251-274, 2001.
[18] Katila, M. & Tomppo, E. Selecting estimation parameters for the Finnish
multisource National Inventory. Remote Sensing of Environment, 76, pp.
16-32, 2001.
[19] McCool, D.K., Foster, G.R., Mutchler, C.K., & Meyer, L.D. Revised
slope length factor for the Universal Soil Loss Equation. Transactions of
the ASAE, 32, pp. 1571-1576, 1989.
[20] Lu, D., Li, G., Valladares, G.S. & Batistella, M. Mapping soil erosion risk
in Rondônia, Brazilian Amazonia: using RUSLE, remote sensing and GIS.
Land Degradation & Development, 15 (5), pp. 499-512, 2004.
[21] Gurevitch, J., Scheiner, S.M. & Fox, G.A. The Ecology of Plants. Sinauer
Associates, Inc., Publishers, U.S.A. 523p., 2002.
[22] Zhang, X.B. & Wen, A.B. Current changes of sediment yields in the upper
Yangtze River and its two biggest tributaries, China. Global and Planetary
Change, 41 (3-4), 221-227, 2004.
[23] Zhang, X.B, Zhang, Y.Y, Wen, A.B. & Feng, M.Y. Assessment of soil
losses on cultivated land by using the 137Cs technique in the Upper
Yangtze River Basin of China. Soil and Tillage Research, 69 (12), pp. 99-
106, 2003.
[24] Editorial Board of Sichuan Vegetation, Sichuan Vegetation. People’s
Publishing House of Sichuan, Chengdu, 463p. (in Chinese), 1980.
[25] Renard, K.G. & Freidmund, J.R. Using monthly precipitation data to
estimate the R-factor in the RUSLE. Journal of Hydrology, 157, pp. 287-
306, 1994.
[26] David, M. Geostatistical Ore Reserve Estimation. Elsevier Scientific
Publishing Company, 364p., 1977.
[27] Hutchinson, M.F. A new method for gridding elevation and stream line
data with automatic removal of pits. Journal of Hydrology, 106, pp.211-
232, 1989.
[28] Tarboton, D.G. Anew method for the determination of flow directions and
contributing areas in grid digital elevation models. Water Resources
Research, 33 (2), pp. 309-319, 1997.
[29] Garbrecht, J. & Martz, L. W. The Assignment of Drainage Direction Over
Flat Surfaces in Raster Digital Elevation Models. Journal of Hydrology,
193, pp. 204-213, 1997.
[30] McRoberts, R. E., Nelson, M.D. & Wendt, D.G. Stratified estimation of
forest area using satellite imagery, inventory data, and the k-Nearest
Neighbors technique. Remote Sensing of Environment, 82, pp. 457-468,
2002.
[31] GLCF (Global Land Cover Facility) - www.landcover.org as a source of:
- U.S. Geological Survey. 10th July 2002, Landsat ETM+, Scene, WRS-2
Path 130, Row 037, Orthorectified Geocover, Sioux Falls, South Dakota:
USGS.
- U.S. Geological Survey. 10th July 2002, Landsat ETM+, Scene, WRS-2
Path 130, Row 038, Orthorectified Geocover, Sioux Falls, South Dakota:
USGS
[32] O'Callaghan, J. F. & Mark, D. M. The extraction of drainage networks
from digital elevation data. Computer Vision, Graphics and Image
Processing, 28, pp. 328-344, 1984.
[33] Labrecque S., Fournier R.A., Luther J.E. & Piercey, D. A comparison of
four methods to map biomass from Landsat-TM and inventory data in
western Newfoundland. Forest Ecology and Management, in press, 2006.
[34] Tokola, T., Pitkanen, J., Partinen, S., & Muinonen, E. Point accuracy of a
non-parametric method in estimation of forest characteristics with
different satellite materials. International Journal of Remote Sensing, 17
(12), pp. 2333-2351, 1996.
[35] Tokola, T., Sarkeala, J. & Van der Linden, M. Use of topographic
correction in Landsat TM-based forest interpretation in Nepal.
International Journal of Remote Sensing, 22 (4), pp. 551-563, 2001.
[36] Morgan, R.P.C. Soil erosion and conservation. Addison-Wesley
Longman, Edinburgh, 198p. 1995.
[37] Van Remortel, R., Hamilton, M. & Hickey, R. Estimating the LS factor
for RUSLE through iterative slope length processing of digital elevation
data. Cartography, 30 (1), pp. 27-35, 2001.
Decreased deposition of sulphate and

the responses in soilwater at Estonian
integrated monitoring sites 1995–2004
J. Frey1, K. Pajuste2, K. Treier1, Ü. Mander1, P. Kask3 & T. Frey3
1
Institute of Geography, University of Tartu, Estonia
2
Estonian Environmental Research Centre, Tallinn, Estonia
3
Estonian University of Life Sciences, Tartu, Estonia
Abstract
This paper presents the results of trend analyses of a ten-year study of deposition
(throughfall+stemflow) and soilwater in two pine stands and a spruce stand (ICP
IM areas). The reported decreasing trends of deposited sulphate and cations at
the easterly located Saarejärve integrated monitoring area are in good agreement
with the decline of SO2 and fly ash emissions in Estonia. At Vilsandi pine stand,
which has the westernmost location and is under a marine influence, the decrease
of sulphur deposition was comparable with that of the eastern Saarejärve pine
stand. However, total base cation load in Vilsandi pine stand remained
unchanged resulting in a decrease of throughfall acidity. At Saarejärv coniferous
stands deposition of base cations decreased more than that of acid anions causing
an increase in K leaching from canopies. A good accordance between sulphate
decline in deposition and topsoil water was accompanied by base cation decline
in soilwater under organic horizon at both IM areas. In podzolized soil at
Saarejärve the decline of SO4 and base cations resulted in increased levels of
total soluble Al.
Keywords: deposition, throughfall, stemflow, pine and spruce stand, soilwater,
sulphate, cations, aluminium, trend analyses.
1 Introduction
Predominant sources of SO2 and particle emission in Estonia are four large oil-
shale-based thermal power plants and chemical industries in the Northeastern
region. Oil shale mining and combustion accounted for about 81% of total
doi:10.2495/GEO060081
harmful emissions up till 2005. Estonian oil shale as a fuel is characterized by

high ash share (45-50%) and moderate sulphur content (1.4-1.8%). A peculiarity
of Estonian oil shale is its chlorine content (0.75%) combined with organic
matter. Oil-shale-centred energetics and economy have led to specific deposition
patterns. The high amount of alkaline dust emitted contributed to a considerable
neutralization of acid pollutants already in the atmosphere, and even resulted in
alkalinization and damage of sensitive ombrotrophic raised bogs [1]. A rapid
decrease in emissions took place in 1990s due to reshaping of local economy
after restoration of independent statehood in 1991. During the study period
(1995-2004), the quantities of emitted SO2 and fly ash decreased by 40% and
65%, respectively, mostly due to a decline in annually combusted oil shale
amounts (from about 24 megatons in 1994 to 13 megatons in 2004) i.e. due to a
decrease in production of electric energy in the power plants.
The decrease of emissions led to a decline in concentrations of both anions
(SO4 and Cl) and cations (Ca, Mg, Na and K) in bulk precipitation at all Estonian
monitoring stations during 1994-2001 [2].
The first long-term national intensive ecosystem monitoring programme in
Estonia: International Co-operative Programme on Integrated Monitoring (ICP
IM) under LTRAP Convention was initiated by the Nordic Council of Ministers
in 1994 [3]. In the framework of ICP IM, an intensive monitoring site at Lake
Saare forested catchment area in eastern Estonia and a biomonitoring site in the
westernmost Island Vilsandi were established. The IM monitoring sites represent
areas of boreal coniferous forest receiving background loads of air pollution and
deposition. Although the two sites are not regionally representative, they provide
an opportunity to monitor changes in air pollution and its impacts on the
environment of coniferous stands at the western border of Estonia (Vilsandi),
where higher concentrations of SO2 are measured from southern and south-
western air mass transport direction [4], and in eastern Estonia (Lake Saare area),
which is more affected by local sources of air pollution from the region of oil
shale industry in North-Eastern Estonia.
The current paper presents results of a ten-year study of throughfall,
stemflow and soilwater in two pine stands and one spruce stand under the
conditions of declining SO2 and base cation emissions.
2 Material and methods

2.1 Site characteristics
The study was carried out in the forested subcatchment area (109 ha) of Lake
Saare in eastern Estonia (58º39´ N, 26º45´ E), hereafter referred to as Saarejärve,
and in Vilsandi, Estonia’s westernmost island (58º34´ N, 21º50´ E). At
Saarejärve water and litter samples were collected from two permanent plots of
the Scots pine (120 years old) and the Norway spruce (90 years old) stands. The
permanent plots are situated on nearly flat surfaces, in areas representative of
dominant forest site types in the catchment area. The Rhodococcum type Scots
pine stand is located at an elevation. The Vaccinium type Norway spruce stand is
situated lower than the pine stand, near the bottom of a slope. At both stands the
parent material is glaciofluvial sand, on which moderately eluviated Haplic
Podzols have developed.
The permanent plot in Vilsandi is a 100-year-old Scots pine stand (Fragaria
type) on Calcari-Gleyic Leptosol.
Mean annual precipitation during the study period (1995-2004) was 630 mm
at Saarejärve and 535 mm in Vilsandi.
Table 1: Stand characteristics of sample plots.
Permanent plot Vilsandi Saarejärve Saarejärve

pine stand pine stand spruce stand
Site type Fragario- Rhodococco- Vaccinio-
Pinetum vitis-idaeo- myrtilli-
Pinetum Piceetum
Soil type Calcari- Haplic Podzol Haplic
Gleyic Podzol
Leptosol
Age of dominant trees 100 120 90
Number of trees in ha-1 440 551 672
Mean diameter (m) 0.29 0.37 0.26
Mean height (m) 16.9 28.7 22.5
Annual average throughfall 282 513 430
amount (mm)
Annual average soilwater amount 101 57 113
from depth of 10 cm (mm)
Annual average soilwater amount 98 25 74
from depth of 40 cm (mm)
2.2 Sampling and chemical analyses
Bulk deposition was collected using two NILU-type collectors [5] at both
monitoring areas. Throughfall deposition was collected by polyethylene funnel-
type bulk collectors (20 cm in diameter, at a height of about 150 cm) in snow
free time, and by buckets during winter. Spiral silicone collectors fitted to three
trees per plot were used for collecting stemflow. Water volumes were measured
on the field by graduated cylinder. Sampling frequencies were once a fortnight in
summer and once a month in winter. Sampling areas and stand characteristics are
given in table 1.
Soil water was sampled with zero-tension plate lysimeters of 0.1 m2 [6]. At
Saarejärve pine and spruce stands the lysimeters were inserted into depths of 5 to
10 cm under organic horizon and about 40 cm under eluvial horizon with 6
replications per depth. At Vilsandi the lysimeters were installed under humus
horizon (into depth of 5 to 10) and illuvial horizon (BC(g) into depth of about 35
cm). At both sites percolation water was collected approximately at 1-month
intervals during the snow-free period along with deposition samples.
Water samples from Vilsandi and Saarejärve were analyzed in the Estonian
Environmental Research Centre in Tallinn and in the Environmental Studies
Laboratory in Tartu, respectively. Both laboratories have continuous quality

control programmes, and they participate regularly in international
intercalibration exercises.
Major anions (Cl-, NO3-, SO42-) in precipitation and soilwater were analysed
by ion chromatography. Base cations were determined using atomic absorption
spectroscopy or ion chromatography depending on laboratory, pH was measured
potentiometrically.
The nonparametric Mann-Kendall test was used to estimate significance
(p<0.05) of the trends in annual values. The slope of a linear trend was estimated
with the nonparametric Sen’s method [7].
3 Results
3.1 Decline of sulphur deposition
During 1995-2004 a significant decline was observed in concentration and
deposition of sulphate in bulk precipitation and in pine and spruce stands’
throughfall and stemflow at the integrated monitoring sites in Estonia (Table 2).
Mean annual concentration of SO4-S in bulk precipitation decreased from 2.2 mg
l-1 to 0.37 mg l-1 at Saarejärve and from 1.4 to 0.51 mg l-1 at Vilsandi. Sulphur
load dropped from 9.1 to 3.2 kg ha-1 yr-1 at Saarejärve. At Vilsandi the highest
load was 5.1 kg ha-1 yr-1 in 1995 and the lowest 2.4 kg ha-1 yr-1 in 2002.
The prevailing anion in throughfall of the pine and spruce stand at
Saarejärve was SO42-, which formed 58% and 66% of summed anions,
respectively. Sea salt correction shows that marine fraction of sulphur is less than
8% in bulk- and throughfall precipitation at Saarejärve. At the remote island
Vilsandi, however, the prevailing anion in throughfall was Cl-1 (53%), sulphate
formed only 29% of summed anions. Marine fraction of sulphate was on average
10% in bulk deposition and 25% in throughfall flux.
The throughfall flux of S decreased significantly from 8.2 and 18.8 kg ha-1,
in 1995 to 4.2 and 4.8 kg ha-1 in 2004 in Saarejärve pine and spruce stands,
respectively. At Vilsandi pine stand the decrease in sulphur deposition was about
the same as in Saarejärve pine stand (50%) (Fig. 1). The concentration of the
other two main anions (NO3 and Cl) also seems to have decreased along with the
decline of sulphate at Vilsandi pine stand (Fig. 1). Comparison of TF from the
two monitoring sites shows that NO3-N deposition is higher at Vilsandi (2.3 kg
ha-1) than Saarejärve (1.0-1.4 kg ha-1) pine stand. The decrease of nitrate
deposition on Fig.1 could be explained by variation between the selected
monitoring years. Furthermore, throughfall NO3 content is more influenced by
internal cycling in canopy [8] than the SO4 content.
Actual annual S input flux into the soil comes from S of throughfall,
stemflow and litterfall. Stemflow water was usually negligible in volume
forming 2-5% of throughfall in the spruce and pine stand at Saarejärve and less
than 1% in the pine stand at Vilsandi. However, due to the high concentration of
SO4-S, significant transport of S took place by stemflow, at least at Saarejärve, in
the first years of the monitoring period. Sulphur input by litterfall formed on
average one third of total S input (TF+SF+LF) in the studied stands.
Figure 1: Decline of throughfall+stemflow deposition (keq ha-1 y-1),

comparison of first and last monitoring years in spruce stand and
pine stand at Saarejärve and Vilsandi on the left side. Sulphate
decline in deposition (throughfall+stemflow) and in soilwater under
organic (org.) and eluvial (E) horizons of spruce and pine stands at
Saarejärve and under organic and illuvial (illuv.) horizons in
Vilsandi pine stand on the right side.
During the study period the actual S input decreased about 2 times in the
pine stands and 4 times in the spruce stand. It should also be pointed out that,
judging by annual litterfall amounts (needles only); the crown needle surface of
spruce canopy is about three times bigger than that of pine canopy. The sulphur
content in litter will be a more important part in the S-circle due to decreased
atmospheric loads of S.
3.2 Decline of base cation deposition
The decreasing trend of SO42- concentration in throughfall+stemflow is in good
accordance with decreasing trends of the concentrations of most important base
cations, Ca and Mg at Saarejärve coniferous stands. The share of Ca2+ and Mg2+
was on average 45% and 15% of summed cations, respectively. Decreasing flux
of summed base cations (Ca+Mg+K+Na) during the study period was
statistically significant at Saarejärve (Table 2).
Table 2: Trends in data series (“-“decreasing and “+” increasing) of annual
mean concentration in throughfall (TF+SF) and soil water (SW)
from 1995-2004. Estimated by Mann-Kendall nonparametric test
(significance levels ***p< 0.001; **p<0.01; *p<0.05).
Stand Saarejärve spruce stand Saarejärve pine stand Vilsandi pine stand
TF+ SW SW TF+ SW SW TF+ SW SW
SF 10 cm 40 cm SF 10 40 cm SF 17 35 cm
cm cm
SO4 - -*** -*** -* - -* -* -* -*
***
Ca -** -* + -* - -
Mg -** -* -*** -* - -* -*
K + -* -** + - - -*
Sum of -** -** -* -* - -* -*
cations
Sum of -** - -* -* - -*
anions
AlTot + +** +* +
Al3+ - -* +* +***
H+ - +* + *+ + -* -* +*
There was a clear change in proportion of various cations in throughfall.

Proportions of Ca2+ and Mg2+ in summed cations decreased by 0.1% and 0.4%
per year, accordingly, in pine throughfall, while the proportion of K increased by
1.6% per year (p<0.05) obviously due to leaching (K did not increase in open
area precipitation). In the spruce stand the decrease of the share of Ca2+ and Mg2+
in summed cations was 1.2% and 0.38% per year, accordingly, and the increase
of K+ due to leaching was 2.8% per year (p<0.05).
Due to its location the dominant cation in Vilsandi TF is Na (40% of
summed cations on average), the remaining 60% is Ca+Mg+K in almost equal
parts. There was no significant decreasing trend of base cation concentration or
deposition at the Vilsandi pine stand.
At Vilsandi deposition of base cations neutralised the acidic deposition in

bulk precipitation to a larger extent than in throughfall, while H+ flux also
decreased in the throughfall of the pine stand. At the same time bulk
precipitation was more acidic than total deposition in the pine and spruce stands
at Saarejärve. The annual mean pH decreased in the bulk deposition (from 4.8 to
4.7) as well as in the throughfall (from about 5.4 to 5.2) during the study period,
and the increase of H+ flux in the pine stand was significant (Table 2).
3.3 Responses of soil solution chemistry to decline of deposition
There were statistically significant declines in sulphate concentration, sum of

cations and pH (from 5.5 to 4.9) in soilwater under organic horizon at Vilsandi.
At Saarejärve statistically significant declining trends in sulphate concentration,
sum of cations and sum of strong anions (SO4+Cl+NO3) were estimated in
soilwater under both horizons of the spruce stand, and under eluvial horizon of
the pine stand. Concentrations of SO4 and of most cations also decreased below
the organic horizon of the pine stand but the trends were not significant (Table
2). SO4 concentrations were about 10% higher in soilwater under the eluvial
horizon than under the organic horizon at Saarejärve. Likewise, a statistically
significant decreasing timetrend of sulphate amount in soil water output from
organic horizon was observed in the spruce and pine stands at Saarejärve. The
output of sulphate flux from organic horizon decreased from 0.25 to 0.05 keq ha-
1 -1
y (about 5 times) in the spruce stand and from 0.14 to 0.025 keq ha-1 y-1 (6
times) in the pine stand at Saarejärve area. The decline of sulphate output fluxes
from deeper horizons was not statistically significant, although leaching from
eluvial horizon in the spruce stand decreased from 90 to 30 eq ha-1 y-1 and in the
pine stand from 67 to 18 eq ha-1 y-1. At the remote island Vilsandi sulphate
output fluxes decreased from 0.12 to 0.08 keq ha-1 y-1 (1.5 times) under organic
horizon and from 0.16 to 0.08 keq ha-1 y-1 (about 2 times) under illuvial horizon.
SO4 content under deeper (B) horizon was remarkable higher in Vilsandi pine
stand than at Saarejärve (Fig.1 on the right).
The decline in soilwater sulphate was to a great extent matched by a
decrease in most cations in organic horizons of the spruce and pine stands, and to
a lesser extent also in the eluvial horizon of both stands at Saarejärve. In the
soilwater of carbonate-rich soil at Vilsandi the cation content decreased in good
accordance with SO4 decline under organic horizon but was not influenced by a
decline of sulphate in illuvial horizon.
Total soluble Al contents increased in the soil water from both depths of
both stands at Saarejärve (Table 1), although the trends were statistically
significant only below organic horizon in the spruce stand and below eluvial
horizon in the pine stand.
4 Discussion
Decrease in sulphur deposition is often accompanied by decline of base cations.
At Vilsandi pine stand the total deposition of summed base cations did not
significantly decrease while the cation flux under humus horizon decreased
significantly during the monitoring period. Thereby, it could be suggested that a
substantial share of cations got deposited not along with sulphate ions at
Vilsandi. On the other hand SO4 from deposition plays important role as an
accompanying anion for cations, which readily moves down with percolation
water through organic horizon.
Throughfall composition forming processes (dry deposition, interception,
leaching) in canopy changed in both stands at Saarejärve during the study period.
Deposition of anions via throughfall and stemflow decreased by about 1 keq ha-1
in the spruce stand and 0.8 keq ha-1 in the pine stand. At the same time
deposition of cations at Saarejärve decreased by 1.5 and 1 keq ha-1 in the spruce
and pine stands, respectively. Since the input of dust-associated base cations
decreased more than the acid anion input, acidity of throughfall should increase,
and cause increased leaching of K from canopies.
Good accordance between sulphate decline in deposition and topsoil water is
partly due to the use of zero-tension lysimeters. Percolation water obtained by
zero-tension lysimeters is the soil water fraction that is primarily involved in soil
formation processes, e.g. transport of ions down the soil profile, from one
horizon to another prior to e.g. buffering processes [9].
The decline of sulphate fluxes in deposition and under organic horizon of
soil water at Saarejärve area resulted in a decrease of SO4 retention in organic
horizon. The retention of sulphate via adsorption processes decreased from about
0.44 to 0.24 keq ha-1 in the spruce and pine stands suggesting a decrease in
consumption of protons (potential increase of H+) by about 0.2 keq ha-1 in the
spruce and pine stands’ organic horizon. The decline in the retention of sulphate
in eluvial horizon occurred to a lesser extent, and resulted in the potential
increase of H+ by about 0.15 and 0.07 eq ha-1 in the spruce and pine stand,
respectively. Via proton consumption the sulphate retention had a decisive role
in proton budgets in organic layers of the spruce and pine stands at Saarejärve in
the first five years of the monitoring period [10]. The decline in SO4 retention in
Vilsandi organic horizon was 0.16 keq ha-1, which is about the same as in
Saarejärve. Higher output of SO4 from illuvial horizon than from organic horizon
indicates suitable conditions for adsorption/desorption processes. The increase of
H+ in percolation water of both horizons in the spruce stand and under the
organic horizon of the pine stands could be due to changes in SO4 retention, as
well as due to the decline of summed cations in soilwater, which reflects
intensive accumulation of base cations into biomass or/and a shortage of
exchangeable cations in soil. Although desorption of SO4 anion appeared to be
much slower than adsorption [11], the previously adsorbed sulphate would have
got released and thus delayed the decrease of SO4 concentration in deeper
horizons.
Only additional H+ can increase both total Al and soluble free Al3+
concentration in soil water. The pH of solutions in the studied soil water was
normally well below 5 at Saarejärve and, in fact, the pH of throughfall was
commonly higher than that of the soil solution receiving it. The increase of total
soluble Al in podzolized soil indicates an ongoing process of podzolization due
to dissociated organic acids derived from mineralisation of conifer litter, and is

probably attributable to decreased retention of sulphate in soil horizons.
5 Conclusions
The decreasing trends of deposited sulphate and cations at Saarejärve integrated
monitoring area are in good agreement with the decline of SO2 and fly ash
emissions in Estonia, as well as with SO4 time series from local precipitation
monitoring stations during the study period. At Vilsandi pine stand, which has
the westernmost location and is under marine influence, the decrease of sulphur
deposition was comparable with that of the eastern Saarejärve pine stand, but the
total base cation load in Vilsandi pine stand remained unchanged over the entire
monitoring period resulting in a decrease of throughfall acidity. Deposition of
base cations at Saarejärve coniferous stands decreased more than that of acid
anions causing an increase in K leaching from canopies.
A good accordance between sulphate decline in deposition and in topsoil
water was accompanied by base cation decline in soilwater under organic
horizon at both IM areas. In podzolized soil at Saarejärve the decline of SO4 and
base cations resulted in increased levels of total soluble Al.
Acknowledgements
This study was supported by Estonian Science Foundation grant No. 6083 and
Target Funding Project No. 0182534s03 of the Ministry of Education and
Science of Estonia.
References
[1] Karofield, E., The effects of alkaline fly ash precipitation on the
Sphagnum mosses in Niinesaare bog, NE Estonia, Suo 47, pp. 105-114,
1996.
[2] Treier, K., Pajuste, K. and Frey, J., ‘Recent trends in chemical
composition of bulk precipitation at Estonian monitoring stations 1994-
2001‘, Atmospheric Environment 38, pp. 7009-7019, 2004.
[3] TemaNord, Establishing of the Baltic Integrated Monitoring (BIM) sites
1992-1994, in, Nordic Council of Ministers, Copenhagen, pp. 1-44, 1996.
[4] Pajuste, K., Kimmel, V., Kohv, N. and Truuts, T., Assessment of the
Estonian EMEP data, in, Assessment of the EMEP measurements and
modelling work in Europe from 1977 until today,
http://www.emep.int/assessment/estonia.pdf, 2003.
[5] EMEP: EMEP manual for sampling and chemical analysis, in,
EMEP/CCC-Report 1/95, Chemical Co-ordinating Centre, Norwegian
Institute for Air Research, Kjeller, pp. 1-199, 1996.
[6] Voll, M. and Roots, O., Soil water sample collector, Environmental
Monitoring and Assessment 54, pp. 283-287, 1999.
[7] Salmi, T., Määttä, A., Anttila, P., Ruoho-Airola. T. and Amell, T.,
Detecting trends of annual values of atmospheric pollutants by Mann-
Kendall test and Sen’s slope estimates – the Excel template application
MAKESENS, in, Publication on Air Quality 31, Finnish meteorological
Institute, Helsinki, pp.1-35, 2002.
[8] Pajuste, K., Frey, J. and Asi, E., Interactions of atmospheric deposition
with coniferous canopies in Estonia. Environmental Monitoring and
Assessment 112, pp. 177-196, 2006.
[9] Derome, J., Lindroos, A.J. and Niska, K., Comparison of soil water and
percolation water quality, in, Forest Condition Monitoring in Finland.
National report 1998, 748. The Finnish Forest Research Institute,
Helsinki, pp. 132-137, 1999.
[10] Frey, J., Frey, T. and Pajuste, K., Input-output analysis of macroelements
in ICP-IM catchment area, Estonia, Landscape and Urban Planning 67,
pp. 217-223, 2004.
[11] McBride, M.B., Chemisorption and Precipitation of Inorganic Ions‘, in,
Environmental Chemistry of Soils, Oxford University Press, New York,
pp. 121-164, 1994.
SAKWeb© – Spatial Autocorrelation and

Kriging Web Service
J. Negreiros & M. Painho
Instituto Superior de Estatística e Gestão de Informação – UNL,
Portugal
Abstract
SAKWeb is the result of research integrating developments in spatial

autocorrelation and Kriging as well as in Web services. This paper presents the
first Web application to offer spatial autocorrelation and association measures,
spatial exploratory tools, variography, Ordinary Kriging (OK) and Simple
Kriging with global mean (SK). SAKWeb© version 2.0 exploits IE®, ASP®,
PHP® and IIS® capabilities and was designed in an attractive and straightforward
way for everyone’s use. Major features are presented in summary. As confirmed
by the survey to GIS users, researchers suggest that this Web service and its
enhanced developments are a promising alternative for the near future, including
its use in an E-learning environment.
Keywords: geostatistics, spatial autocorrelation, kriging, web services,
E-learning.
1 Introduction
Spatial analysis technology is now beginning to reach the stage where advanced
users are struggling with new methods of spatial exploratory data analysis
relating to possible explanations for any spatial phenomena. From an explanation
of the unitary to a description of the collective, all views of space are correct if
the where question is answered, what spatial correlations exist and why spatial
distribution is structured in that way. Hence, this article mainly focuses on
methods that address the inherent stochastic nature of patterns, interpolation and
relationships based on measured point properties such as spatial autocorrelation
and Kriging rather than deterministic ones such as buffering areas.
doi:10.2495/GEO060091
Many GIS users do not think in spatial terms and do not ask questions about
spatial pattern and spatial autocorrelation. While conventional users prefer to use
GIS modules because of the trademark that underlies the software or the direct
need to topology access, experts tend to use statistical products and independent
geo-software where interfaces are a non-concern issue. Also, conventional users
continue to manage large volumes of spatial data with very deterministic spatial
operations while experts are more concerned with stochastic issues such as
fuzzier boundaries. If a high quality solution depends on geo-statistical and
geographical knowledge then a W3 service with hypermedia help is an excellent
way to fill this knowledge gap between both communities Negreiros and Painho
[8].
With regard to the implementation of new spatial analysis tools, the future
does not seem very bright on account of the undeclared divorce between the
research community and the software developers. Therefore, the W3
environment emerges as an alternative solution when compared with the
traditional statistical, independent and GIS modules approaches (Negreiros and
Painho [8]). Johnston et al. [5]. From a technological standpoint, cooperative
work, computation distribution and networking contribute to the widespread
dissemination of geographical knowledge with more diverse embedding
technologies. From the commercial perspective, the Internet can be viewed as an
extension of the traditional competitive marketplace Laudon and Laudon [6].
From the science organization point of view, the synergy effect of the online
educational material in which the overall information may be greater than the
sum of its parts is a conviction. Under the GIS view, the greatest W3 impact is to
close the data access gap among users with free and direct retrieval of spatial
analysis tools in an E-learning context. Hence, in this paper we put together four
components that will allow an enhancement and a wider use of spatial analysis,
new and simpler spatial autocorrelation measures, alternatives to handle the
Kriging nugget-effect implementation – everything in an universal access
platform (the Internet).
Thus, this expose is organized as follows: The four sub-sections of the
following section justifies SAKWeb project by highlighting major geo-
statistical innovations, E-learning issues, the poor knowledge concerning spatial
autocorrelation, Kriging and simulation topics by many GIS users and the
challenge to build an alternative framework to the traditional geostatistical
modules (Geostatistical Analyst of ESRI®), independent software (VarioWin®,
GeoMS®, GSLib®, STAC®, Regard®, SAGE®, TRIWaco®) and statistical
packages (SAS-GIS®, SPSS-X®, Glim®, Systat® or Minitab®). The basic
requirements facing technologies are presented in section three while section
four contains a brief description of SAKWeb© functionalities. Section five
analyses the user evaluation of SAKWeb©. The concluding section summarizes
the main inferences of this paper and future prospects.
2 Why develop SAKWeb?
2.1 New features
It is significant to stress that all new features referenced here are being developed
in other articles and documentation such as the inclusion of spatial
autocorrelation measures as a valid exploratory spatial data analysis (ESDA)
tool, the enhancements of the conventional Moran scatter plot, the inclusion of
the four nugget-effect approaches and the implementation of the local error
variance. Nevertheless, a brief overview of some of those will be mention next.
With regard to the weight neighbourhood matrix construction, inverse
distance weighting (IDW) is indicated as a poor process because it is not
sensitive to its direct neighbours Negreiros [7]. Hence, SAKWeb© presents the
covariogram contiguity for which neighbours’ boundaries are setup by the
variogram range. Any other weight can be applied provided it means something
to the researcher and can be justified theoretically Ebdon [3]. In accordance with
various lag distances, Moran I is presented as a good search factor of the
neighbour radius. Based on the range of the highest Moran I, the weight matrix is
defined while the Moran location scatter plot is created (the IDW deterministic
interpolation is also based on the same procedure). This mapping computing
framework allows the study of spatial patterns, outliers, changeover areas and
trends. Furthermore, the Moran I correlogram may lead to a re-estimation
process of the variogram sill, model, nugget-effect and range Negreiros [7].
The discontinuity of the variogram origin can be viewed as a non-sense
situation if spatial reality is considered a continuous surface. Among all
variogram factors, the nugget-effect is also the most unpredictable because of the
lack of close samples Clark and Harper [2]. So, SAKWeb© offers four strategies
to handle the nugget-effect: (A) γ(0) = 0 (model 1); (B) γ(0) = 0 but including C0
for superior lags (model 2); (C) γ(0) with micro-scale, γ1(h), and long-range,
γ2(h), assessment (model 3); (D) If measurement error is given, this attribute will
be incorporated within the Kriging system (model 4). It is assumed that the sill is
a good global population variance estimate, a condition that most of the times is
not true Soares [12]. Just as the arithmetic sample mean is a poor global mean
estimate, preferential sampling with proportional effect leads to a false
variability value. It is crucial to readjust the variogram sill to reflect the true
global variance. Using the nearest neighborhood analysis for the samples
weights, it is possible to find the relationship between the Estimated Global
Variance (EGV) and the original variogram sill. It also avoids the difficulties of
weighting samples at the edge. If this ratio multiplies the variogram model, a
rescaling operation, then the Kriging classical variance and region confidence
interval, both local and global, can be enhanced.
2.2 E-learning
Recently, a lot of interest has been put on E-learning methods. Quite often,
common users request spatial analysis knowledge in a self-learning view because
of global cost reduction, both time and money. In addition, implementing the
technological structure that supports E-learning platform is a scalable solution
(Painho et al. [9]). Spatial analysis wizards, multimedia tools (including
animation and hyperlinks), on-line help, software courses, videos, E-Learning
and M-Learning with WML technology are ingredients of this demand. CRM of
the Universidade Autónoma de Lisboa and GIS&Sc Master of ISEGI, certified
by UNIGIS, are some examples of the Learning Space strategy (Semana
Informática [11]). The International Center for Distance Learning (http://www-
icdl.open.ac.uk), the AT&T Learning Network Virtual Academy
(http://www.att.com/learningnetwork/virtualacademy), La Escuela de Negocios a
Distancia de la Universidade Politecnica de Madrid (http://www.cepade.es), Le
Centre National d’Enseignement à Distance (http://www.cned.fr) and the
Universidade Aberta (http://www.univ-ab.pt) are others. As confirmed by Santos
[10], five million US students were taking distance degrees in 1994 and two
million in the EC in 1997.
In Europe, this learning trend has been successfully tapped into by EC
programs such as Delta (1988-1990), Telematics Applications (1994-1998) and
Multimedia Task-Force (1999). Strong future development is therefore expected,
supported by current brain-ware know-how, and confirmed by the following
examples: (A) The UNAVE project from the Universidade de Aveiro
(http://www.unave.pt/fd) reveals greatly encouraging results with its multimedia
Web interfaces program: 70% of the 350 students passed, 20% were advised to
repeat and 10% failed. (B) With PROF2000 LAN/WAN management classes,
the final grades of nineteen high-school professors ranged between 77% and
90% (Santos [10]). (C) During 1999, the estimate for Portugal Telecom’s cost
savings with e-Learning (http://formare.ptinovacao.pt) reached €270,000 for
ISDN, LAN/WAN and telecommunications courses. (D) Santos [10] presents a
p-Fisher statistical study comparing conventional (276 participants) and distance
learning (42) methodologies on the basis of age (26-35, 36-45 and 46-55) and
school qualifications (grades 6, 7-9, 10-11 and 12), revealing a positive
homogeneity of both approaches with a 95% confidence level. This conclusion
was based on two representative populations of the same number, 30. A positive
Pearson correlation was also found between final grades and school
qualifications (+0.313) and a negative one with age (-0.413).
2.3 Modest expertise in sophisticated spatial analysis
Many GIS users do not possess the skills necessary to make use of sophisticated
tools such as Kriging, spatial autoregressive models, autocorrelation measures
and simulation uncertainty, thus creating a limitation to the understanding of
further explanations to their problem. In a survey conducted to Portuguese users,
ninety-two people were surveyed at six academic and research GIS institutions
reveals a poor situation regarding the knowledge of spatial autocorrelation
software: Statistical software – 5%; GIS software – 1%; Geo-software – 7%; No
knowledge – 89% (Negreiros et al. [8]).
The First Law of Geography is only known by 5% of all respondents, with
20% possessing theoretical knowledge of some kind of spatial autocorrelation
and clustering measures: Variogram cloud (18% of the total population); Moran I
(8%); Gi(d) and Geary C (5%). There are several explanations for this present
situation: (A) The present teaching subjects do not involve these recent
quantitative geography matters. (B) Constant training and updating are an
absolute GIS requirement. (C) The hardware requirements to run these types of
packages can be high. (D) There is a lack of difficulty of access to non-cost GIS
software in high-schools, colleges and universities. (E) Geostatistical
foundations, for instance, are statistical complex for major users.
According to this survey, all respondents believe that the implementation of
a spatial analysis Web service can be a wonderful strategy although there is no
software available. To fill this present spatial analysis software vacuum is a
major aim of this work. After all, E-learning and expertise lack of sophisticated
spatial analysis regarding GIS problems are two different issues of the same
coin. Therefore, the ambition of SAKWeb© is to develop the first spatial
autocorrelation and Ordinary Kriging approach in a Web environment for people
with data observation problems that can be handled by both methodologies.
2.4 The W3 challenge
What the users appreciate and interact with is the GUI (graphical user interface),
because they do not care about the technical structure, as long as the results are
trustworthy, prompt and compatible with their operating system and hardware
(Negreiros and Painho [8]). Users want intuitive and easy-to-use software in
order to give immediate results without having to read pages of documentation.
Luckily, the standard Web browser, irrespective of the background computer
code adopted, fulfills this strategy quite well. It is cost free and already provided
in any operating system. Ultimately, the user will work with all available
software in the same way, regardless of the location of the data or its purpose:
(A) In the beginning, the trend was to find the data with the browser; (B) Today,
download the software and installed it; (C) In the near future, just run the
programs from the Internet.
This represents a significant advance in the user interface because users will
no longer have to worry about the software location and the technical knowledge
necessary to connect to the data. Even today, the keyboards available on the
market hold special keys for the browser such as home, search, back, mail and
refresh. The standard Web browser is the future interface. This solution relies on
an enhanced standard front-end acting as a protected wall against the computer
code in an embeddable, extendable and reusable development. The capability to
undercover technical implementation to the final user via WWW is essential.
After struggling for years for digital information, spatial analysis needs to
concentrate on what the information means, sharing it through the W3 and
distributed architectures (Negreiros and Painho [8]). In effect, the wireless Web
is developing furiously nowadays. Web-enabled wireless devices will enable
millions of people to access the Internet while on the go. Therefore, the
challenge of improving spatial data exploration by applying the available tools of
spatial analysis and back-office technology integration with the Internet (mobile
or not) is also critical.
3 SAKWeb© technology
When common users begin to handle and understand the background concepts of
spatial autocorrelation and Kriging, the difficulty of getting specialized software
becomes all too clear (particularly with spatial autocorrelation and spatial
autoregressive methods). It is important to forget installation procedures and
operating system requirements and to embrace a direct, friendly and known GUI
with mobile access. If the inferences presented in the previous section are added
to this process, the final major SAKWeb© specifications are, thus, setup.
To make this project come to life, several W3 technologies were used
including ASP®, PHP® and Dreamweaver® (cf. figure 1). Data Access
Components®, WebChart® from Component One®, FrontPage® Server
Extensions Service Release, Flash®, HTML, Java Applets®, VBScript® and
JavaScript® are some of the other components required. On the bottom of
everything lies the client-side and server-side structure that supports this
dynamic application.
This Web service is, thus, a collection of ASP® pages, server components
and a Website where the distinction among the applications is made by the root
directory within the site. All content within its directory structure is considered
part of the scope of the same application. Further, each application has its own
set of variables and attributes that define its current state and these are
maintained throughout the application lifetime. Another primary concern with
Web based applications is that HTTP has no memory and retains no information
from one client request to the next. ASP® gets around this issue by using
Application and Session objects to store information during a user’s session.
SAKWeb
Technologies
Internet
Information
Server
From CGI to ASP Java Applets

and PHP
JavaScript DHTML
ComponentOne
WebChart And
Flash
ActiveX Controls
Server and Client-side

Technology Client-side Technology
(SAKWeb Programming) (SAKWeb Help)
Figure 1: SAKWeb© technologies.
The 2D and 3D WebChart® gives also power to any ASP® environment by

generating different graphic charts. Internally, the server-side component shares
a common Application Program Interface (API) with the client-side that
generates an .OC2 or .OC3 control to be passed to the client with the appropriate
HTML tags. If the client does not have olch2x8.cab and olch3x8.cab, the server
cab file generates a runtime-only copy of WebChart® on the client's machine
(these CAB files are compressed versions of Component One® Chart).
Control management of the site is achieved by means of three options: 1)
DOS Commands - The capability to browse the contents of DOS® commands; 2)
PHP Configuration - The ability to display the PHP® configuration of
SAKWeb©; 3) Session Variables – It gives the CGI parameters and the contents
of the session variables created in the current user session, a useful tool for the
programmer to control the application flowchart. It also plots a horizontal 3D
graphic with the distribution of the visitors to this site throughout the day by
dividing it into six groups: 0h-4h, 4h-8h, 8h-12h, 12h-16h, 16h-20h and 20h-24h.
This information should guide the Web users to select the best time to connect to
SAKWeb©.
4 SAKWeb main features
SAKWeb© is not a full geostatistical and simulation software package. Its

functionality is depicted on the basis of four critical procedures that implement
together known spatial autocorrelation measures as well as newly developed
ones:
• Data Input and Exploring View, which focuses on MS-Excel® input, control
management of the user session and exploratory data analysis. New features
include the estimation of the global mean and global variance whose weights
are based on the nearest neighbourhood measures.
• ESDA, Spatial Autocorrelation and Variography, which concerns variogram
setup, the Moran I correlogram, the Moran location and the variance scatter
plot. Recent characteristics include the Moran location and variance scatter
plots.
• SAKWeb© Ordinary Kriging, which concentrates on OK calculus and
surface mapping in accordance with four nugget-effect strategies. Validation
with an extra dataset, 3D-2D surface profiles, cross-validation and region
plumes based on threshold values and confidence levels are also included.
• SAKWeb© Help, presents ten options regarding software and geostatistical
help with the Web driven interface using hypermedia with sound and
movement and an e-Learning structure that includes an IRC, Email and
News service, to close the knowledge gap between experts and common GIS
users.
5 SAKWeb© end-user evaluation

In 2003, a survey was carried out to twenty GIS users. SAKWeb© was setup in
an IIS® server, for ten clients using the intranet of Rainha Dona Leonor college
in Lisboa. The same GS+ Pb dataset was given so that they could test the
SAKWeb© features (cf. Figure 2). Previously, a 45 minutes refresher overview of
spatial autocorrelation, trend surfaces, nearest neighborhood and OK subjects
was presented. However, no tips were given on how to use SAKWeb©.
5 4.5 4.8
4.1
4 3.7 3.6
Figure 2: The average classification of the impact of SAKWeb© major features.
On the basis of the survey presented and with post-statistical processing,

several interpretations can be drawn:
a. The SAKWeb© Moran variance scatter plot was too slow to be computed. In
fact, 80% of the users complained about this issue, stating that this routine
leads to impaired performance, that is, the slowness of the server impacts on
concurrent operations of other users. Thus, the ASP® script language that
underlies this procedure was changed to C. However, the strength of
SAKWeb© in concurrent processes did not reveal any other major flaw.
b. The capability to brush and link different maps was pointed out as seriously
deficient by 60% of the interviewees, a major trend in the ESRI®
Geostatistical Analyst module (particularly for assessing classical statistics
in a moving average context). Accordingly, two new options were setup
within SAKWeb©: local interactive statistics and 3D-2D OK profile.
c. In this survey, the limited number of OK models available, especially when
compared with the ESRI® Geostatistical Analyst module, is another issue to
be taken into consideration. This includes different lag spacing depending on
different directions, smaller separation distances for the first lag and other
variography tools such as the correlogram, the madogram, the general
relative variogram and the pairwise relative variogram. Consequently, these
last four features were included after this survey was carried out.
d. Surprisingly, 95% of the users demand cross-validation.
e. The questionnaire reveals that 45% of the respondents think that it is

important to include the Box-Cox, Normal and logarithmic transformations
in geostatistical software.
f. This questionnaire indicates that 30% of the users think that topographical
and hill shade background maps enhance final Kriging maps. For this
reason, image mapping in a new window was added afterwards. Curiously,
no users mention the absence of the traditional semivariance/covariance
cloud and the QQ-plot in SAKWeb©.
g. All answers reveal that these users support the SAKWeb© Moran location
scatter plot for verifying trends and local pockets of stationary. The bivariate
scatter plot is also a good option. However, the histogram tool in
conjunction with the brushing and linkage location map represents a
wonderful tool for this purpose.
0
Low Marginal Costs Mobile And Easy Access Hardware and Software Standard Web GUI
Management
Figure 3: The final ranking for the main advantages of the Web solution.
After the survey was completed, certain issues were discussed in a brief
meeting. The respondents confirmed previous suggestions: they really appreciate
the capability to generate several Kriging surfaces and compare them without
any extra work. It is expected that future software developments will enhance
this feature. Therefore, the capability to extend this on-the-fly potential based on
different range, model and anisotropy factors could be a remarkable
improvement since the present solution is only based on the four nugget-effect
strategies. In terms of the interface, it was unanimously agreed that this Web
service is extremely easy to use for two main reasons: 1) The hints given and the
validation check for the input parameters; 2) The inability to proceed to the next
step if the previous one is not correct (it works like a process wizard).
Two users reclaim the migration from ASP® 3.0 to ASP.Net® architecture.
However, the introduction of the Moran location scatter plot (instead of the
conventional one) and the Moran variance scatter plot represents a major step in
spatial autocorrelation measures. According to the questionnaire, the Moran
location scatter plot was ranked first in terms of usability and visual impact for
outliers, trends and patterns search. The Moran I and Geary C took second place
while the variogram cloud ranked third. But, this latter measure requires the
brushing and linking feature. Another interesting comment made by two users
was: “Until now, the emphasis on spatial autocorrelation measures by geo-
software has been limited and almost forgotten. It is hoped that SAKWeb©
restores in our minds the need to implement them with Kriging software.”
Another point is that the SAKWeb© Help approach presents alternative help
software to that provided by the conventional Contents-Search-Index pattern.
According to most inquiries, laying out the overall help structure with hyperlinks
among subjects is really an attractive model. As some users confirmed, “it works
like a puzzle where each piece corresponds to each issue, though at the same
time a complete view is given of the structure topics and how they are related to
each other.” Future help applications must be encouraged to follow this concept
too.
6 Conclusions
There is an increased demand for systems that do more than display and data
organization Ebdon [3]. The set of potential applications for spatial analysis is
enormous, for example, accident patterns, victim profiles within a residential
population and spread rates for pollution levels. Thus, spatial statistics must hold
a specific spatial framework to apply quantitative and statistical methods for a
better understanding of spatial relationships. If the finding of spatial structures is
fundamental then autocorrelation, interpolation and autoregressive models are
three major spatial methods that fulfill this constraint.
However, when common users begin to handle and understand the
background concepts of spatial autocorrelation and Kriging, the difficulty of
getting specialized software becomes all too clear. It is important to forget
installation procedures and operating system requirements, to embrace a direct,
friendly and known GUI with mobile access, to incorporate standard exchange
for data input, to validate input parameters including suggestions and hints on-
the-fly, to add new features available and to provide the capability of discussing
these issues on an on-line basis with good hypermedia help.
With the advent of Web technology and modern wireless computing, it has
become necessary to develop a W3 service for interpolation to understand the
often complex spatial autocorrelation that exist among the samples collected in
space. The E-Learning geostatistical solution and, particularly, the close linkage
between spatial autocorrelation and Kriging were never considered. The
inclusion of these two major components provided another inspiration for
SAKWeb© version 2.0.
Although this project is still a work in progress, the future of SAKWeb© can
be bright. Already beyond the scope of this paper, the present SAKWeb©
infrastructure can be applied easily as a WWW interface with GSLib® routines to
avoid the reinvention of the wheel for other geostatistical tools, a situation
required by some respondents in the survey.
The Next Kriging Menu selection has this aim by leading the user to a new
choice of options that include SK, UK, IK and CK. In effect, SK with global
mean has already been implemented with the following options: SK Calculation,
Default SK, SK With No C0, SK With Two Structures, Exact SK Differences, SK
Versus OK, Validation With Extra Dataset and Cross-Validation (as with OK,
this selection includes the location of positive-negative/true-false samples). A
third menu screen has already been sketched for simulation and spatial auto-
regression procedures. Adapting Anselin [1] contemplation, the respectability of
the academic research concerning implementation should be a reality, where the
results are reflected in appealing pictures while the arduous work is taking place
beyond-stage in a W3 environment.
References
[1] Anselin, L., Exploratory Spatial Data Analysis in a Geocomputational
Environment in Geocomputation A Primer. John Wiley and Sons, New
York, 1998, pp. 77-94.
[2] Clark, I., Harper, W., Practical Geostatistics Answers to the Exercises,
Ecosse North America, 2000, pp. 298.
[3] Ebdon, D., Statistics In Geography, Blackwell, 1998.
[4] Goovaerts, P., Geostatistics For Natural Resources Evaluation, Oxford
University Press, 1997.
[5] Johnston, K., Hoef, J., Krivoruchko, K., Lucas, N., Using ArcGIS
Geostatistical Analyst, ESRI, 2001, pp. 287.
[6] Laudon, K, Laudon, J, Management Information Systems. Prentice-hall
International, 2002, pp. 523.
[7] Negreiros, J., SAKWeb - Spatial Autocorrelation and Kriging Web, a W3
Computing Perspective, Unpublished Ph.D. Thesis, ISEGI-UNL, 2004,
pp. 449.
[8] Negreiros, J., Painho, M., The Web Platform for Spatial Statistical
Analysis, Proceedings of the 6th Conference of the Portuguese Association
in Information Systems (CAPSI), Bragança, Portugal (26-28 Nov 05).
[9] Painho, M., Cabral, P., Peixoto, M., Pires, P., E-teaching and GIS: ISEGI-
UNL learning experience, Third European GIS Education Seminar
EUGISES, 2002.
[10] Santos, A., Ensino à Distancia e Tecnologias de Informação: E-Learning,
FCA, 2000.
[11] Semana Informática,
http://www.lotus.com/home.nsf/welcome/learnspace, 2002.
[12] Soares, A., Geoestatistica para as Ciências da Terra e do Ambiente, IST
Press, 2000, pp. 206.
Section 3
Environmental pollution
and remediation
After treatment of landfill leachate in

peat filters
M. Kõiv1, M. Kriipsalu2 & Ü. Mander1
1
2
Department of Technology, University of Kalmar, Sweden
Abstract
The main objective of this study was to determine the treatment capacity of well
mineralised Sphagnum peat in order to reduce BOD and COD values and
nutrient concentration in landfill leachate. The peat filters were suitable for the
reduction (up to 93%) of ammonia nitrogen. Good results were obtained in the
reduction of total phosphorus from both raw and pre-treated leachate (up to 81%
and 70-99% respectively). The purification rate of the landfill leachate depended
on the contamination rate – the outflow results were better with pre-treated
leachate, and the results also improved due to the lowering of the flow rate (on
average by 60 times). Therefore, it is recommended that peat filters be used in
combination with conventional treatment methods, e.g. as soil filters of
subsurface flow constructed wetlands for the secondary or tertiary treatment of
the leachate.
Keywords: ammonia-nitrogen, BOD, COD, peat filter, pre-treated landfill
leachate, raw landfill leachate, total nitrogen, total phosphorus.
1 Introduction
Peat is partially fossilized decomposed plant matter that transforms in wet areas
in the absence of oxygen. Compared to mineral soils, peat has a very high
organic content (60% carbon). Peat has a surface area of >200 m2 g-1 and is
highly porous (80-90%) [1].
In Estonia, peat lands cover 22% of total land area. Estonian peat resources
are estimated at 2.4 billion tonnes, of which 0.2 billion tonnes are less
decomposed, and 1.4 billion tonnes are well decomposed [2].
doi:10.2495/GEO060101
The potentially large availability of peat and its unique combination of

biological, physical and chemical properties make it suitable for a wide variety
of uses, including environmental protection [3]. For example, peat has been used
in the treatment of wastewaters of various origin and quality.
Several laboratory and field experiments have demonstrated that peat as a
filter material for constructed wetlands, and also for bio-filters and other
conventional treatment systems can effectively reduce nitrogen concentration
and remove suspended solids [4, 5, 6], pathogenic bacteria [4, 5], mineralise
organic material, retain phosphorus [7, 8] and other heavy metals [1, 9]. Most
peat filters are designed for the treatment of domestic wastewater [10], and
several systems have also shown very good performance in the treatment of
landfill leachate [11, 12]. Likewise, results of experiments on floodplain fens in
Estonia have shown the high potential of peatlands in the after-treatment of
wastewater [13].
Although peat is an inexpensive and attractive material, such advantages must
be balanced against the importance of peatland conservation and the maintenance
of habitat diversity [14].
The composition of landfill leachate varies greatly, being dependent on waste
quantity and quality (the age of the waste and also its decomposition rate and
landfilling technologies). Leachate is considered difficult to treat due to its
typically high concentration of P (up to 100 mg L-1 [15]), ammonia nitrogen (up
to 300 mg L-1 [11]), high COD value (up to 60,000 mg L-1 [15]), and heavy
metals.
In Estonia there is a lack of long-term experience in the treatment of landfill
leachate. There are only a few landfills where leachate is collected and purified
using various methods.
The main objective of this study was to determine the treatment capacity of
well mineralised Sphagnum peat in order to reduce BOD and COD values, and
nutrient concentration in landfill leachate from different stages of the leachate
treatment system. We also investigated whether peat filters improve the
efficiency of conventional leachate treatment systems. In addition, the effect of
the duration of the experiments on treatment efficiency, as well as changes in the
composition of the peat, was studied.
2 Materials and methods

Two experiments were conducted at Väätsa landfill in Estonia, the first
experiment (E1) in summer 2003 and the second experiment (E2) in 2005.
2.1 Site description
The Väätsa landfill is the first sanitary landfill in Estonia that meets the
requirements of the EU Council Directive [16] and Estonian landfill directives
[17]. The first stage of the landfill (1.0 ha) was in service from 2000-2005, and
the second stage (1.5 ha) is in operation since November 2005. The landfill
serves approximately 40,000 inhabitants. By the present time (February 2006),
60,000 t of mixed waste has been deposited [18]. The landfill has a proper lining,
leachate collection system and two-stage biological leachate treatment system
consisting of an activated sludge treatment plant and aerobic-anoxic pond
(Fig. 1).
Figure 1: Aeration of landfill leachate in aerobic-anoxic pond in Väätsa, Estonia.
2.2 Experimental design
In both experiments, custom-designed peat filters (F) (total four filter bodies)
were used. In experiment 1 (E1) the two metal filter bodies (filters 1 and 2 – F1,
F2) had a volume of 1 m3, were rectangular in shape, and had a permeable floor
(Fig. 2). In the second experiment (E2), two filter bodies (filters 3 and 4 – F3,
F4) with a volume of 0.2 m3, were made of PVC pipe (Ø 372 mm, h=1200 mm)
(Fig. 2).
2.2.1 Peat type

All filters were filled with well mineralised fluffy Sphagnum peat collected from
Lokuta peat bog, which is located near Väätsa landfill. The well mineralised peat
was obtained from the lower deposits of depleted industrial peatlands. The peat
was sieved through a 26 mm sieve to remove stones and roots.
Figure 2: Design of the filter bodies and leachate distribution to the peat filters in
experiment 1 (F1; F2) and in experiment 2 (F3; F4).
2.2.2 Distribution of leachate and loading of filters

For the loading of F1 raw leachate and for the F3 the leachate from the activated
sludge treatment plant was used. For the F2 and F4 we used treated leachate
from the pond (the outflow of the treatment system).
The even distribution of leachate into the F1 and F2 (E1) filters was achieved
using perforated elastic pipes placed above the filter bodies in a spiral pattern
(Fig. 2). The even distribution of leachate into the F3 and F4 (E2) filters was
achieved using perforated pipe inside crushed granite (Ø 6-8 mm) on top of the
filter material (Fig. 2).
All filters were loaded using timer-adjusted pumping. In the first experiment
F1 and F2 were loaded for 36 days, with 2.9 m3 m-2 in a day (in total about 104
m3 per filter). In the second experiment, the filter 4 was loaded for a total of 5
months – from June to October 2005 – and the F3 since April 2005, for a total of
9 months. The loading rate for F3 was 0.083-0.033 m3 m-2 d-1 (total amount
about 13.7 m3), and for F4 0.05 m3 m-2 d-1 (about 7 m3).
2.3 Sampling and statistical analyses
The leachate samples from the inflow and outflow of F1, F2 were taken on days
2, 4, 6, 8, 15 and 22, and once a month from the inflow and outflow of F3, F4:
eight times from F3 and four times from F4.
In both experiments, BOD7, COD, total N, NH4-N, NO3-N, NO2-N, total P
and pH were determined in the certified laboratory using standard methods. In
experiment two we also analysed: PO43-, conductivity, TSS, SO42-, Ca2+, Mg2+
and total hardness. In peat, the content of organic matter (%), N (%), pHKCl and
P, K, Ca, Mg (mg kg-1) was determined.
The normality of variables was checked using the Lilliefors and Shapiro-Wilk
tests; for normally distributed variables (Total P, BOD, COD, NH4-N, pH), the
inflow and outflow values in different peat filters were compared via the
pairwise t-test. When the distributions were skewed, the nonparametric
Wilcoxon pairwise test was used. When the assumptions of ANOVA were
fulfilled, a Fisher LSD test was used for multiple comparisons of mean removal
efficiencies in different filters. For the remaining variables, Kruskal-Wallis
ANOVA and the multiple comparison of mean ranks for all filters was used. The
STATISTICA 7.0 software was used and the level of significance of α=0.05 was
accepted in all cases.

3.1 Landfill leachate treatment at Väätsa
At Väätsa landfill, approximately 10-20 m3 of leachate is produced per day and

treated in a biological treatment system. The treatment efficiency of conventional
activated sludge treatment, followed by aerobic-anoxic treatment in a pond, is
high. However, the average values of BOD7, COD, total N, and total P in the
outflow from the aerobic-anoxic pond (Table 1) exceeded the limits required for
treated wastewater in Estonia [19].
Table 1: Average treatment efficiency in Väätsa landfill leachate treatment

system in 2004 and 2005 [20, 19].
Two-stage treatment system Overall treatment

Para- inflow outflow efficiency (%)
Unit
meter
2004 2005 2004 2005 2004 2005
-1
BOD7 mg L 2193 993 52 93 90 91
COD mg L-1 4133 1880 451 847 90 55
Total N mg L-1 401 282 117 115 60 59
Total P mg L-1 1.7 6.5 3.6 4.6 30
pH 7.4 8.2 8.7 8.8
In Estonia there are special target values for leachate treatment [17]:
contamination rate 25 mg L-1 (purification rate ≥90%) for BOD7; 125 mg L-1
(≥75%) for COD; 2.0 mgL-1 (≥80%) for total P and 75 mg L-1 (≥75%) for total
N. Pre-treatment of leachate in the Väätsa biological treatment system is not
sufficient to fulfil prescribed values for effluent.
3.2 Leachate treatment in experimental peat filters
In the first experiment, biologically pre-treated leachate from the second stage of
treatment purified about 2-5% easier than the raw leachate. When we compare
E1 and E2, however, the treatment efficiency is better in the second. One
explanation of why E2 had better results in the removal of contaminants than E1
may be that the selected loading rate was too high for the filters in E1 [11].
Average values of contaminants in the inflow and outflow of the peat filters,
and differences between inflow and outflow values in all filters (significant
differences when p<0.05, according to pairwise t-tests) are presented in Table 2.
Table 2: Average values and standard deviation of the contaminants in the

inflows and outflows of the peat filters. Significant differences
between the inflow and outflow values of the contaminants in all of
the peat filters studied (pairwise t-test results): * - p<0.05; ** -
p<0.005.
Parameter F1 F2 F3 F4
(mg L-1)
in out in out in out in out
1953 1728 54 27 79 24 19 6
BOD7
±459 ±505* ±19 ±19* ±32 ±15** ±16 ±3
3812 3462 521 466 1158 843 598 592
COD
±446 ±555** ±49 ±60** ±239 ±253* ±37 ±150
2.0
1.3 0.7 1.2 0.6 5.6 3.3 0.2
Total P ±1.7*
±0.2 ±0.3** ±0.2 ±0.1** ±1.9 ±0.7 ±0.2*
*
447 401 112 95 339 302 127 115
Total N
±24 ±63* ±22 ±37 ±121 ±101 ±55 ±23
369 327 52 37 254 97 9.5 7.2
NH4-N
±60 ±44* ±10 ±7* ±115 ±85** ±9 ±2
5.8 5.8 18 17 38 143 50 43
NO3-N
±2.2 ±1.3 ±6.7 ±2.7 ±53 ±93* ±14 ±14
8.2
7.5 7.8 8.8 8.3 8.5 8.8 8.9
pH ±0.2*
±0.1 ±0.2* ±0.1 ±0.2** ±0.2 ±0.1 ±0.6
*
The average value of pH in leachate decreased in F2 and F3, and an increase

took place in F1 and F4. The changes in pH value according to the pairwise t-test
were significant in filters 1, 2 and 3 (Table 2). As stated by Patterson et al. [21],
the slight change in pH could be related to the organic acid components flushed
from the peat.
3.2.1 Reduction of BOD and COD values

A decrease in BOD value was observed after treatment in all filters (Table 2).
The treatment efficiency in filters was 12-70% on average, and maximum
reduction was achieved in F3 (95%). The outflow value that fulfils the prescribed
limit of 15.0 mg O2 L-1 [17] was achieved in F4, where the outflow value was 6.4
mg O2 L-1 on average. The most effective removal of COD was achieved in F3,
where the values were on average almost 30%.
The reasons why BOD and COD removal was some times not successful
enough may be that factors were not favourable (pH range, the existence of
inhibitors, the lack of substrate and phosphorus, temperature, contamination rate
etc). For instance, the BOD and COD values in raw leachate were too high
(Table 2) for it to be treated only with a peat filter.
3.2.2 Removal of nitrogen

The best results for nitrification were achieved in the F3 (Table 2), where
ammonia nitrogen was reduced by 62% on average (Fig. 2). Only in F4 did no
significant nitrification take place. The reduction of total N concentration was
noticeably higher in F2 (35% on average). However, significant differences
between inflow and outflow were only determined in F1. The reduction was not
sufficient to fulfil the Estonian discharge limit [17].
Maximum removal of ammonia nitrogen (62% on average) was achieved by
F3 (Fig. 2). Significant differences between inflows and outflows were
determined in F1, F2 and F3 (Table 2). An approximately 12 and 30% reduction
is obtained in F1 and F2 respectively.
Mæhlum [22] indicated that nitrification may be limited due to lack of
oxygen. The amount of oxygen in peat was not directly measured, but the
intermittent loading of leachate during the experiment was selected in order to
increase the natural inflow of oxygen through the top and the bottom of the filter
body. In experiment 2 the amount of oxygen was measured from the inflow and
outflow leachate of the peat filters. The results show that the peat filters increase
the amount of oxygen in leachate (e.g. 95% on average in F3).
Nitrification occurs at an optimum temperature of about 30ºC [23]. At 5ºC,
the nitrification rate is only 15% of the rate at 20ºC [24]. The temperatures of
raw and pre-treated leachate during the summer period in Väätsa were
14.8±0.3ºC and 18.0± 1.6ºC respectively. Thus there should be favourable
conditions for the creation of nitrifying bacteria in that period.
The nitrification of ammonia also depends on the hydraulic loading [25]. We
can confirm this fact, because E2 had better results in nitrification then E1. Also,
when we lowered the hydraulic loading rate into the F3 filter several times
during experiment 2, the ammonia nitrogen decreased, and NO3-N and NO2-N
increased significantly.
3.2.3 Removal of phosphorus

The removal of total phosphorus from Väätsa leachate (Fig. 2) was very
successful, and average outflow values were below the limits [15]. According to
pairwise t-tests, the inflow and outflow values demonstrated a significant
difference (p≤0.005) in all filters. The total phosphorus reduced in F1 and F2
was on average 49% and the F3 and F4 filters removed 64 and 92% respectively.
Mann [26], Kadlec and Knight [23] and Richardson et al. [27] demonstrated
that the reduction of total P could be caused by the sorption, sedimentation and
combination of complex compounds. A certain proportion of P may be bound to
the biofilm [26, 27].
Phosphorus transforms easily from organic to inorganic forms and constitutes
chemical complexes with organic and inorganic ligands, which can be adsorbed
by or sedimented into the soil. In aerobic wetland conditions the P constitutes in
dissolved complexes with oxidised Ca and Mg in alkaline conditions, and with
Fe or Al in soil with acidic to neutral pH [22]. The content of the Mg and Ca in
peat increased significantly, whereas the pH of the peat was neutral.
According to Mæhlum [22], it can be concluded that the fixation of P resulted

in adsorption with Ca and Mg compounds, and these were settled in the peat
filter. P content increased in peat, supporting this statement.
Figure 2: Average inflow and outflow concentrations of NH4-N and total P in all
filters. * - p<0.05, the inflow value is significantly different from
outflow values, according to the pairwise t-test.
3.3 Composition of peat
The initial concentrations of Ca, Mg and K in the well mineralised peat are
relatively high (Table 3). Before the experiment, the content of organic matter in
peat was almost twice lower (37%) than on average, which could be due to the
small stones that were found in the peat. During the experiment there was a
slight increase in the content of organic matter. After the experiment, Ca, Mg
and especially K concentrations increased significantly (96%).
Table 3: Composition of peat in filters 1 and 2 before and after the first
experiment. Units are mg kg-1 (unless otherwise noted).
pHKCl N (%) P K Ca Mg Organic matter (%)

Before 6.7 0.5 23 93 8294 1416 37.2
After 7.1 0.6 37 2177 9178 2149 36.9
The removal of contaminants from leachate can be linked to chemical

changes in the peat. The results are affected by interactions between peat type,
water quality, loading rates, duration of treatment etc. The removal of total P
with the filters may be caused by the high content of Mg- and Ca-compounds in
peat.
4 Conclusions
In Estonia, the use of peat as a filter material has good potential. It has shown
sufficient purification efficiency and can be considered to be an ecologically
sound and economically beneficial material in leachate treatment.
The pre-treatment of landfill leachate considerably reduces the pollutant load
on the peat filter, which increases its performance. A smaller flow rate and
respectively longer retention time will be required for the optimal performance
of peat filters.
We can conclude that peat filters are well suited to the reduction of Total P
from raw leachate (up to 81%), as well as from pre-treated leachate (up to 70-
99%).
The remarkable efficiency of well-mineralised peat in the reduction of BOD
values (up to 95%) and NH4-N concentrations (up to 93%) encourages us to use
peat filters in combination with conventional treatment methods, e.g. as soil
filters of subsurface flow constructed wetlands for secondary or tertiary
treatment of landfill leachate.
Acknowledgements
Science of Estonia. Pille Kängsepp, Tõnu Salu, the staff from Väätsa landfill and
colleagues from the Institute of Forestry and Rural Engineering at the Estonian
University of Life Sciences are acknowledged.
References
[1] Brown, P.A., Gill, S.A., Allen, S.J., Metal removal from wastewater using
peat. Water Res, 34 (16), pp. 3907-3916, 2000.
[2] Statistical Office of Estonia Web site, Extraction of mineral resources.
http://www.stat.ee
[3] McLellan, J.K., Rock, C.A., Pretreating landfill leachate with peat to
remove metals. Water Air Soil Poll, 37, pp. 203-215, 1987.
[4] Lens, P.N., Vochten, P.M., Speelers, L., Verstraete, W.H., Direct
treatment of domestic waste-water by percolation over peat, bark and
woodchips. Water Res, 28 (1), pp. 17-26, 1994.
[5] White, K.D., Byrd, L.A., Robertson, S.C., O’Driscoll, J.P., King, T.,
Evaluation of peat biofilters for onsite sewage management. J Environ
Health, 58(4), pp. 11-17. 1995.
[6] Gunes, K., Ayaz, S.C., Wastewater treatment by peat filtration. Fresen
Environ Bull, 7 (9A-10A). Sp.iss. SI: 777-782, 1998.
[7] James, B.R., Rabenhorst, M.C., Frigon, G.A., Phosphorus sorption by peat
and sand amended with iron-oxides or steel wool. Water Environ Res; 64
(5), pp. 699-705, 1992.
[8] Roberge, G., Blais, J.F., Mercier, G., Phosphorus removal from
wastewater treated with red mud-doped peat. Can J Chem Eng; 77, pp.
1185-1194, 1999.
[9] Davila, J.S., Matos, C.M., Calcavanti, M.R., Heavy-metals removal from
wastewater using activated peat. Water Sci Technol; 26 (9-11), pp. 2309-
2312, 1992.
[10] Geerts, S.M., McCarthy, B., Axler, R., Henneck, J., Heger
Christopherson, S., Crosby, J., Guite, M., Performance of peat filters in
the treatment of domestic wastewater in Minnesota. 9th National
Symposium on Individual and Small Community Sewage Systems, Fort
Worth, TX, March 11-14, 2001, Amer Society of Agricultural Engineers,
Michigan, USA, 2001.
[11] Kadlec, R.H., Integrated natural systems for landfill leachate treatment.
In: Vymazal, J. editor. Wetlands – Nutrients, Metals and Mass Cycling.
Backhuys Publishers, Leiden, Netherlands, pp. 1-34, 2003.
[12] Kinsley, C.B., Crolla, A.M., Kuyucak, N., Zimmer, M., Laflèche, A., A
pilot peat filter and constructed wetland system treating landfill leachate.
9th International Conference on Wetland Systems for Water Pollution
Control, Avignon (France), 26-30th of Sept. 2004, Cemagref. Antony, pp.
635-642 2004.
[13] Öövel, M., Tarajev, R., Kull, A., Mander, Ü., Tertiary treatment of
municipal wastewater in a floodplain peatland. In: de Conçeicao Cunha,
M., Brebbia, C.A. editors. Water Resources Management III. WIT
Transactions on Ecology and the Environment, Vol. 80 WIT Press, pp.
433-444, 2005.
[14] Ma, W., Tobin, J.M., Determination and modelling of effects of pH on
peat biosorption of chromium, copper and cadmium. Biochem Eng J, 18,
pp. 33-40, 2004.
[15] Tchobanoglous, G., Theisen, H., Vigil, S., Integrated solid waste
management: Engineering principles and management issues. McGraw-
Hill, Inc. 2nd Edition, New York, 1993.
[16] European Union Council Directive, 1999/31/EC of 26 April 1999 on the
landfilling of waste. Official Journal of the European Communities,
L182/1-19, 0001-0019, 16 July. 1999.
[17] RT I 2003.83.565, Requirements for wastewater discharged into water
bodies or into soil. Regulation No. 327 of 19 December 2003 of the
Government of the Republic of Estonia, Tallinn, Estonia, In: State
Gazette, 2003.
[18] Lõhmus, A. Personal communication, 20 January 2006, Chairman of

Väätsa landfill Ltd., Estonia.
[19] Sooäär, M., Activated sludge treatment of leachate from Väätsa landfill.
Estonia, Eco-tech 03, Kalmar, Sweden, 47 pp, 2003.
[20] Maves Ltd., Väätsa landfill leachate, surface water and ground water
monitoring data. Tallinn, 2005.
[21] Patterson, A.A., Davey, K., Farnan, N., Peat bed filters for on-site
treatment of septic tank effluent. In: Patterson, R.A. and Jones, M.J.
editors. Advancing On-site Wastewater Systems 25-27th September 2001,
Armidale, Lanfax Labs Armidale, pp. 315-322, 2001.
[22] Mæhlum, T., Cold-climate constructed wetlands: Aerobic pre-treatment
and horizontal subsurface flow systems for domestic sewage and landfill
leachate purification. Agricultural University of Norway, PhD thesis,
1998.
[23] Kadlec, R.H., Knight, R.C., Treatment Wetlands. Lewis Publishers, New
York, pp. 3-893, 1996.
[24] Ødegaard, H., Treatment of wastewater. (Rensing av avløpsvann. Tapir
forlag. In Norwegian), Trondheim, pp. 133-284, 1992.
[25] Heavey, M., Low-cost treatment of landfill leachate using peat. Waste
Manage, 23, pp. 447-454, 2003.
[26] Mann, R.A., Phosphorus removal by constructed wetlands: substratum
adsorption. In: Cooper, P.F., Findlater, B.C. editors. Use of Constructed
Wetlands in Water Pollution Control, Pergamon Press, Oxford, pp. 97-
105, 1990.
[27] Richardson, C.J., Qian, S.S., Craft, C.B., Predictive models for
phosphorus retention in wetlands. In: Vymazal, J. editor. Proc. Conf.
Nutrient Cycling and Retention in Wetlands and Their Use for
Wastewater Treatment, Prague, Inst. Of Botany, Trebon, Czech Republic,
pp. 125-150, 1996.
Studies on the distribution of heavy metal Cd

in contaminated soils of various particle sizes
and removal efficiencies of heavy metal
using acid washing
S. F. Cheng1, C. Y. Huang2 & L. S. Hsiao1

1
Department of Environmental Engineering and Management,
Chaoyang University of Technology, Taiwan
2
Ordnance Readiness Development Center, Taiwan
Abstract
Improper treatment and disposal of industrial wastewater and solid wastes result
in serious heavy metal contamination of soil. A commonly used method for
remediation of the soil contaminated by heavy metals is acid washing. This
method is simple in principle, easy to operate and efficient in achieving the
removal of heavy metals.
The experience gained on soil remediation in Taiwan reveals that the soil
particle distribution influences the heavy metal removal efficiency to a large
extent. Hence, in this research, soil samples were collected from the
contaminated sites and used in the investigation on the distribution of heavy
metal Cd in soils of various particle sizes and the efficiencies of acid washing to
remove the Cd from the soil of various particle distributions. The results will be
used for future engineering implementation of the acid washing technique.
The research results indicate that soils containing particles with sizes below
0.150 mm will hold increasing quantities of Cd and organic matter at decreasing
particles sizes. As far as the acid washing efficiency is concerned, smaller
particle diameters and higher organic matter contents result in decreasing
efficiencies.
Keywords: soil, Cd, particle size, CEC, organic matter, acid washing.
doi:10.2495/GEO060111
1 Introduction
The rapid industrial development in Taiwan during last several decades have not
gone together with adequate planning and management of industrial wastewater
treatment and disposal. Irrigation channels have often been used to discharge
industrial waste effluent thus causing serious pollution to large tracts of
farmland. Among all problems, the contamination caused by heavy metals,
which are accumulative and non-degradable, is the most serious.
The acid-washing method is commonly used for remediation of heavy metal
contaminated soils. This method is simple in principle and easy to implement to
come up with good results in a short period of time [1]. It is considered as a
preferred method for remediation metal contaminated farmland in Taiwan.
However, the experience with several recent cases of soil remediation using the
acid-washing method shows that the efficiency depends on the type of soil to a
great extent. For sandy loam soil, this method can achieve metal removal
efficiency of as high as 90% while for soils containing fine clay particles, a poor
separation of the soil particle and the washing liquid causes a great deal of
operational problems. Additionally, the soil that contains fine soil particles
exhibits poor permeability thus making it impossible to have effective contacts
between the washing liquid and soil particles. Hence, the on-site treatment has
removal efficiency lower than 30%; it may sometimes not show any removal
efficiency at all.
The literature information [1] shows that most pollutants exist in the portion
of soil particles having less than 63 µm with relatively larger specific surface
area of the contaminated soil. Hence, sieving the soil into two portions prior to
the acid treatment is recommended. The larger particles have a lower metal
contamination and are considered as clean soil; only the smaller particles need to
be washed in the subsequent acid-washing treatment [2]. Anderson et al. [3]
pointed out that most of the contaminants concentrate in the portion consisting of
silts and clays with less than 50 µm diameters. The volume of this portion is a
small percentage of the overall soil volume. Thus, the prior sieving procedure
may reduce 60~80% volume of the soil to be treated thus reducing the treatment
cost. Results obtained by Sheets and Bergquist [4] on PCB and lead contents of
soil samples collected at superfund sites indicate that soil particles with
diameters less than 74 µm constitute about 8~15% of the total soil mass.
Nevertheless, most PCBs exist in the soil particles with less than 74 µm
diameters. Smaller particles, which have higher specific surface area and contain
more organic substance to enhance the adsorption of contaminants, show higher
PCB concentrations. The distribution of heavy metals generally follows the
similar trend that smaller particles have higher metal concentrations. Particles
with diameters between 0.425 to 2 mm have much higher metal concentrations.
Thus, the presence of small lead particles is regarded as the cause of the
observed high metal concentration in this soil particle diameter range. Di Palma
and Medici [5] combined sulfuric acid and EDTA as the washing agent to wash
copper contaminated soil and reported that soils with higher content of organic
substances exhibit lower copper removal efficiencies mainly because the

complex reactions between copper and the organic substances.
In this research, soil samples collected from the heavy metal contaminated
field in Taiwan are used to study particle distribution as well as concentrations of
heavy metal Cd and organic substances in soil particles of various sizes. The
objective is to understand the correlation between distribution of heavy metals
and soil particle size as well as organic substance contents. Additionally, the
efficiency of extracting the metals from the contaminated soils using various acid
solutions will be studied and the results will be used as references for future
implementation of the acid-washing method.
2 Materials and methodology
2.1 Soil preparation
Samples used in this research were collected from the top 15 cm layer of a
cadmium-contaminated farmland located at Huwei Township, Yunlin Hsieh in
central Taiwan. The collected sample was thoroughly mixed; a portion of it was
subject to basic characteristic analyses on pH, CEC, organic substances, water
content and heavy metals. The remaining portion was dried, crushed and
screened through a 10-mesh (2 mm) sieve to remove large pebbles and
impurities.
The screened samples were then separated into 4 groups of different particle
ranges using the wet screening and the gravitational sedimentation methods.
First, aqueous (NaPO3)6 solution was added to the sample to disperse soil
particle. After thorough mixing, the soil solution was passed thorough 100 mesh
(0.150 mm) and then 270 mesh (0.053 mm) wire screens for separating the
particles into “sand 1” with particle diameters between 2 and 0.150 mm and
“sand 2” with particle diameters range of 0.150 to 0.053 mm. The solution that
had passed through the wire screens was poured into a 1000 ml sedimentation
container. Based on Stock’s law, the particles settled at the bottom after 7 hours
and 36 minutes are silt with diameters between 0.053~0.002 mm. The remaining
particles suspended in the solution are clay with diameters below 0.002 mm; they
were filtered out through 0.20 µm and then dried.
2.2 Soil characteristic analyses
Both the original and the processed soil samples were analyzed for organic
substance and CEC. The organic substance analyses were done using the
combustion method by heating 20 grams of the dried soil sample at 500~600oC
for 4 hours. The weight loss after heating is taken as the soil organic substance
content. For measuring the soil CEC, 4 grams of the dried sample was mixed
with sodium acetate solution such allowing Na+ ion to replace all exchangeable
cations contained in the soil sample. The mixture was then added with
ammonium acetate to replace all Na+ ions by ammonium ions; the concentration
of replaced Na+ ions is determined using the induced couple plasma (ICP)
method as the soil CEC.
2.3 Soil heavy metal analyses
The dried soil samples were digested with aqua regia by adding 21 ml
concentrate hydrochloric acid and 7 ml nitric acid with 3 g of soil slowly. After
mixing, the sample was undisturbed for 16 hours. It was then diluted to 100 ml
and filtered through 0.20 µm filter paper. The filtrate was analyzed for cadmium
using the ICP method.
2.4 Acid-washing of soil
Soil samples containing various particle sizes were washed with single acids
including two inorganic acids (hydrochloric acid and nitric acid) and one organic
acid (citric acid) as well as the mixture of hydrochloric and citric acids to study
the efficiencies of heavy metal removal from the acid-washed soil. The acid
concentrations were 0.01 M, 0.05 M and 0.1 M; the mixed acid solution was
prepared by mixing 1 volume of hydrochloric acid and 1 volume of citric acid to
final concentrations of 0.01 M, 0.05 M and 0.1 M.
An aliquot of 10 ml washing solution was added to 1 g soil sample. After
mixed on a 150 rpms vibrating shaker for 10 minutes, the mixture was filtered
through 0.20 µm filter paper. The washed soil sample was then analyzed for
heavy metals to calculate the metal removal efficiency.
All analyses were carried out with triplicate samples and the results were used
to calculated average and standard deviations.

3.1 Soil characteristics
Table 1 lists the results of basic soil characteristics. According to the USDA soil
classification standards, the contaminated soil is sandy loam containing 125.2 ~
128.2 mg/kg cadmium. It has average 22.7 meq/100 g CEC and contains 5.2%
organic matter.
Table 1: Properties of the Cd-contaminated soils.
pH 6.83 (0.09)
Organic content (by weight)% 5.15 (0.05)
Water content (by weight)% 2.81 (0.06)
CEC (meq per 100 g) 22.7 (1.30)
Sand (by weight)% 68.5 (1.63)
Silt (by weight)% 27.0 (0.21)
Clay (weight)% 5.5 (0.46)
Values in parentheses denote standard deviations (n=3).
3.2 Soil particle and Cd distributions
Figure 1 shows variations of the particle and cadmium distributions. The results
indicate that “sand 2” particles; i.e. particle with diameters between 0.150 mm to
0.053 mm, constitute a major portion of the soil of nearly 48%. Additionally, the
cumulative mass of all soil particles with diameter greater than 0.053 mm is 70%
of the total soil mass. Fig. 1 also indicates that the cadmium content in these
70% soil particles is much lower than those in the clay and silt portions. The
“sand 2” particles have the highest mass percentage but the lowest cadmium
content of 48.74 mg/kg. On the contrary, the clay portion has the lowest mass
percentage but the highest cadmium concentration of 557.37 kg/mg that is 11
time the cadmium concentration contained in “sand 2”.
During the acid-washing process, clay and silt cannot be effectively separated
from the liquid due to their small particle sizes thus lowering the washing
efficiency and raising the treatment costs. Adopting the screening procedure to
separate smaller particles that contain higher cadmium concentrations to be
treated separated will greatly enhance the remediation efficiency for heavy metal
removal.
weight distribution %
800 48.04 50
Weight % Cd conc.
40
Cd conc.(mg/kg)
600
557.37
26.98 30
400
20.46 358.20
20
200
10
5.52
77.22
48.74
0 0
Sand1 Sand2 Silt Clay
Figure 1: Distribution of particle sizes and Cd content of various particle

sizes for contaminated soil. Error bars are ± standard deviations for
n=3.
3.3 Correlation between soil characteristics and heavy metal contents
The correlation between organic matter content and heavy metal concentration
for soil samples containing particles of various sizes is shown in Figure 2. For
sand 2 to clay, smaller particles contain higher concentrations of organic
substances. Since the average soil has higher humic substances associated with
particles of smaller diameters, the organic content is expected to be higher in
soils containing smaller particles. This is contrary to the observation made in
this study that “sand 1” soil has the organic content more than “sand 2”.
However, most of these organic substances are non-decomposed animal and
vegetation impurities.
Results shown in Figure 2 indicate similar trends of variations for heavy
metal concentration and organic content in soils of various particles. For “sand
2” to clay, smaller particles contain more organic substances and heavy metals.
While “sand 1” shows the increasing organic content is not seen by a
proportional increase of cadmium although the cadmium is seen to increase.
This observation may be caused by the fact that most organic substances
contained in “sand 1” are large non-humic thus they do not have high affinity for
heavy metals as those contained in “sand 2”.
800 6.00
700
Cd conc. organic matter 5.00
600
Cd conc. ( mg/kg)
4.00
500
organic content %
400 3.00
300
2.00
200
1.00
100
0 0.00
Figure 2: The correlation between organic matter content and heavy metal
concentration for soil samples containing particles of various sizes.
The correlation between CEC and heavy metal content for soils containing
particles of various sizes as shown in Figure 3 indicates that “sand 1” soil has the
lowest CEC. Further, variations of CEC and organic content show similar
tendency. In “sand 2” and Clay, higher CECs are associated with small soil
particles while “sand 1” has higher CEC than “sand 2”. The small increase of
CEC in “sand 1” proves that portions of the organic substances contained in
“sand 1” are non-decomposed large particles and they do not have much
adsorption capacity for heavy metals.
800 160
700 140
Cd CEC
600 120
Cd conc. (mg/kg)
CEC (meq/100 g )
500 100
400 80
300 60
200 40
100 20
0 0
Figure 3: The correlation between CEC and heavy metal content for soils
containing particles of various sizes.
100 96.57
91.42
90 86.73 84.85 0.01 M 0.05 M 0.1 M
80
Cd removal efficiency %
69.57 70.95
70
60 52.51 51.76 54.03
50
40 33.78
30.05
30 25.51
20
10
0
Figure 4: The metal Cd removal efficiencies from soil containing particles of

various sizes using hydrochloric acid.
3.4 Efficiencies for heavy metal removal by acid-washing for particles of

various sizes
The metal Cd removal efficiencies from soil containing particles of various sizes
using hydrochloric acid, nitric acid, citric acid and hydrochloric acid/citric acid
mixture are shown in Figure 4 to 7, respectively. Results shown in all figures
illustrate that the “sand 1” and “sand 2” particles have much higher removal
efficiencies than clay and silt with as high as 65% difference. Using the single
acid for washing, inorganic acids are more efficient than organic acids;
hydrochloric acid is more efficient than nitric acid. For “sand 1” and “sand 2”
particles, the acid-washing method (by 0.1 M HCl or HNO3) can remove more
than 90% cadmium while the efficiency is around only 50% for silt and 32% for
clay. Figure 7 shows the cadmium removal efficiency using the
hydrochloric/citric mixture. The results show that the mixed acid solution is
2.5% more effective than single acids in removing cadmium from soils of large
particles.
100
90.47 90.35
90 86.50
82.67
0.01 M 0.05 M 0.1 M
80
69.58
70 65.89
60
50 44.51
41.25
40 35.42
31.31
30 25.99
22.44
20
10
0

various sizes using nitric acid.
100
90 86.22
80.77 0.01 M 0.05 M 0.1 M
80 75.69
72.43
70
60 56.40 55.43 53.69

50 44.88
40
30 25.22
20 15.98
12.08
8.57
10
0

various sizes using citric acid.
97.48
100 93.84
89.58 0.05 M 0.1 M
90 0.01 M
82.47 81.84
80 74.57
70
56.99
60
48.57
50
37.47
40 32.57
30 25.41
20.47
20
10
0

various sizes using the hydrochloric/citric (1:1) mixture.
4 Conclusions and recommendations

Results obtained in this research show that the organic content, CEC and heavy
metal content in the soil are somewhat consistent. Particles with diameters
greater than 0.053 mm constitute more than 70% of the total soil mass but
contain the lowest metal contents. They have the high metal removal efficiency
and are easily separated from liquid. On the contrary, sol particles with
diameters less than 0.053 mm contribute to 30% of the total soil mass especially
the clay portion that is only 5.5% of the total mass. This portion of soil has the
highest metal content and the lowest removal efficiency in addition to being
more difficult to be separated by sedimentation from the liquid. The results
obtained in this research indicate the use of a prior screening procedure to
removal particles of small diameters that have the lowest acid-washing
efficiencies. The acid-washing carried out on large particles will be efficient in
removing heavy metal while the separated smaller particles can be treated using
other treatment methods such that the overall treatment efficiency can be
optimized.
References
[1] Evanko, C.R. & Dzombak, D.A., Remediation of metals-contaminated
soils and groundwater. Technology evaluation report, TE-97-01, Ground-
water Remediation Technologies Analysis Center, 1997.
[2] Rosetti, P.K., Possible methods of washing fine soil particles
contaminated with heavy metals and radionuclides. M.S. Thesis, Carnegie
Mellon University, Pittsburgh, PA, 1993.
[3] Anderson, R., Rasor, E. & Ryn, F.V., Particle size separation via soil
washing to obtain volume reduction. Journal of Hazardous Materials, 66,
pp. 89-98, 1999.
[4] Sheets, R.G. & Bergquist, B.A., Laboratory treatability testing of soils
contaminated with lead and PCBs using particle-size separation and soil
washing. Journal of Hazardous Materials, 66, pp. 137-150, 1999.
[5] Di Palma, L. & Medici, F., Recovery of copper from contaminated soil by
flusing. Waste Management, 22, pp. 883-886, 2002.
Mitigation of alkaline mine drainage in a

natural wetland system
J. Kelly1, P. Champagne2 & F. Michel1
1
Department of Earth Sciences, Carleton University, Canada
2
Department of Civil Engineering, Queen’s University, Canada
Abstract
Many studies have focused on the generation and mitigation of acidic drainage
generated when sulfide bearing material is exposed to the atmosphere and
undergoes oxidation. Neutral or alkaline mine drainage can be produced from
mining waste containing little or no sulfides, and such drainage can also contain
elevated metal concentrations, potentially impacting receiving environments.
The goal of this study was to characterize the biogeochemical interactions
occurring throughout a natural wetland located in the Farr Creek drainage area in
Cobalt Canada and to evaluate the ability of the system to effectively attenuate
alkaline mine drainage. The biological characterization of the sediment samples
demonstrated the presence of acid producing bacteria in consistent numbers with
sulphate reducing bacteria and iron reducing bacteria. The data suggested that the
acid produced by these active bacterial populations was immediately neutralized
by the dissolution of carbonate minerals within the tailings, yielding a neutral to
alkaline drainage. The distribution of metals including As, Co, Cu, and Zn
throughout the sediments, pore water, and vegetation samples collected at
various core locations the metal mass was primarily bound in the sediments or
adsorbed onto organic matter or oxide fractions of the sediments.
Phytoremediation processes involving Typha latifolia were shown to attenuate
metals, particularly Cu and Zn. Adsorption onto organic matter and oxides was
another attenuation pathway that significantly improved metal retention. Iron and
sulfate reduction were also found to lead to the formation of metal sulfide
precipitates, thereby immobilizing the metals.
Keywords: wetland, alkaline mine drainage, heavy metals, sequential extraction,
phytoremediation, sulphate reducing bacteria, iron reducing bacteria, acid
producing bacteria, Typha latifolia.
doi:10.2495/GEO060121
1 Introduction
Anthropogenic sources of metals, including mining wastes, runoff waste streams,
and air fall deposition from industrial operations, are becoming increasingly an
issue for contamination of downstream environments and health impacts to
aquatic species and humans. Acid mine drainage (AMD), which results from the
oxidation of sulfide bearing material in waste rock or tailings, has received
considerable attention. The oxidation of sulfides, which may be catalyzed by
microbial populations, produces sulfuric acid, which can result in extremely low
pH waters and enhance heavy metal dissolution. Metals are typically more
soluble under low pH conditions due to the increased competition between
protons and metal cations for organic ligands (Schnoor [1]). The dissolved
metals have the potential to migrate offsite resulting in potential negative
impacts to both humans and aquatic organisms. The acidity produced can
potentially be neutralized if there is adequate carbonate minerals present in the
tailings and mixing waters, resulting in a net neutral or alkaline mine drainage
(NAMD). NAMD can also be produced from mining wastes containing little to
no sulfides. Such drainage, although non-acidic, can also contain elevated
dissolved metal concentrations, which can potentially impact receiving
environments.
Historically many mining companies have used active treatment options for
mitigating impacts to the environment as a result of acid mine drainage. These
options typically involved chemical additions of neutralizing agents such as lime,
which was quite expensive in the long-term. This sparked the movement towards
passive treatment options, one of which is the use of constructed or natural
wetlands to attenuate the metals. As a result, it is important to characterize metal
distributions and transformations in aquatic environments to better understand
the geochemical and biological processes regulating transformations in these
environments. The main objective of this research was to characterize the
biogeochemical interactions occurring throughout a natural wetland receiving
NAMD and to evaluate the ability of the wetland in the attenuation of metals.
The study area is located in the Farr Creek drainage area, in Cobalt, Canada.
The wetland area investigated is confined to a relatively narrow northeast
oriented valley that is bounded to the south by Crosswise Lake and the remnants
of a gravel dam, and bounded to the north by a water level control dam.
Alkaline tailings underlie the entire study area as well as upgradient of the study
area. Mill Creek, which transports metal loadings from several upstream tailings
deposits and organic loadings from the municipal wastewater lagoon flows into
Farr Creek as shown on Figure 1. The northern portion of the area is maintained
under a water cover for much of the open water season, whereas the southern
portion of the area is relatively dry throughout the summer and fall. This is
further evidenced by the establishment of grasses and sedges in the drier areas,
while waterlogged areas are primarily populated with cattails (Typha latifolia).
Figure 1: Site plan indicating core sample locations.
2 Experimental approach
2.1 Field sampling
Five 5 cm diameter x up to 1.4 m long sediment cores were collected from the
study area in September 2004 (Figure 1). The cores were capped and sealed from
the atmosphere throughout the extraction process. They were kept in a
refrigerator and transported in coolers with icepacks to ensure the samples
remained below 4°C until they were prepared for analysis.
2.2 Laboratory methods
Sediment samples were collected from the 5 cores at approximately 25 cm

intervals and analyzed for acid producing bacteria (APB), iron reducing bacteria
(IRB) and sulfate reducing bacteria (SRB), sequentially extracted metals (SEM),
water content (WC), and organic matter (OM) content. Pore water was extracted
from the remaining sediment samples by ultracentrifugation and analyzed for
alkalinity, pH, dissolved oxygen (DO), Fe(II), sulfate, sulfide, and dissolved
metals. Cattail shoot samples, collected near each of the monitoring wells, were
analyzed for metals.
The most probable number (MPN) technique was used to estimate the number
of APB, IRB, and SRB in each sample. MPN values were calculated for each set
of samples from statistical tables (Cochran [2]) and the results were expressed as
colony forming units per gram of sediment dry weight (CFU/gdw). Ten different
dilutions, at 1 mL in 10 mL were completed with 5 replicates of each dilution.
Sixteen sediment section samples taken at depths corresponding to the

microbiology samples, were analyzed in duplicate for metals using a SEM
procedure outlined by Tessier et al. [3] where the extracted metals are separated
into five fractions: exchangeable, bound to carbonates, bound to Fe and Mn
oxides, bound to OM, and residual fraction. Seven cattail (T. latifolia) shoot
samples were washed with deionized water, broken into small pieces and
crushed with a mortar and pestle. Thirty mL of 4:1 HNO3:HCl mixture was
added to 200 mg of ashed cattail shoots. The mixture was digested at 100-130°C
until the solution became clear. Prepared sequential extraction and vegetation
samples were stored below 4°C until further analysis and the metals in the
samples were analysed by ICPMS.
3 Results
3.1 Sequential Extraction of Metals (SEM)
Figures 2(a)-(d) present the results of the SEM for selected metals from the 5
cores. Arsenic was predominantly associated with the residual fraction, ranging
from 35 to 75%, while the oxide fraction ranged from less than 5% to 31% and
generally decreased with depth. The fraction of As associated with the OM
fraction of the sediment ranged from 10 to 35%. Cobalt was strongly associated
with the OM fraction of the sediment, ranging from 22 to 80%. Cobalt was also
associated with the residual fraction, ranging from 5 to 40%. Cobalt associated
with the oxide fraction ranged from less than 5% to 32% and generally decreased
with depth. Cobalt associated with the carbonate fraction of the sediment ranged
from 8 to 30%. Cobalt has been shown to have an affinity for carbonates, thus, it
is likely that there was competition for sorption sites between the carbonate and
other fractions of the sediment. Copper was primarily associated with the
residual fraction of the sediment, ranging from 10 to 50%, as well as the OM
fraction of the sediment, 40 to 85%. Zinc was predominantly associated with the
residual fraction, ranging from 15 to 65%. Zinc associated with the oxides
fraction ranged from less than 10 to 25% and generally decreased with depth.
Zinc associated with the OM fraction of the sediment ranged from 35 to 45%.
3.2 Pore water chemistry
Figures 3(a)-(h) display porewater concentration profiles with depth. In core S2,
both As and Co concentrations increased with depth, while in cores S3 and S5,
As and Co concentrations increased to a depth of 50 cm and then decreased. In
cores S1 and S4, no observable trends were noted. For cores S3, S4, and S5, Cu
concentrations increased to a depth of 50 cm and then decreased. There were no
observable Cu trends with depth in cores S1 and S2. Zinc concentrations
increased with depth in cores S1 and S3. In core S5, Zn showed increasing
concentrations to a depth of 50 cm, followed by decreases, while in core S4 Zn
decreased with depth. In core S2, Zn showed no observable trend. Calcium
values generally remained fairly constant with depth at all core locations.
Percent Distribution (%) Percent Distribution (%)
0
10
20
30
40
50
60
70
80
0
10
20
30
40
50
60
70
80
90
S1- 25 cm S1- 25 cm
S1- 50 cm S1- 50 cm
Exchangeable
Exchangeable
S1- 75 cm S1- 75 cm
S2- 25 cm S2- 25 cm
S2- 50 cm S2- 50 cm
S2- 75 cm S2- 75 cm
Carbonates
Carbonates
S3- 25 cm S3- 25 cm
S3- 50 cm S3- 50 cm
S3- 75 cm S3- 75 cm
Location
Location
Fe, Mn-oxides
Fe, Mn-oxides
S4- 25 cm S4- 25 cm
S4- 50 cm S4- 50 cm
(b) Cobalt
(a) Arsenic
S4- 75 cm S4- 75 cm
Organic
Organic
S4- 100 cm S4- 100 cm
S5- 25 cm S5- 25 cm
S5- 50 cm S5- 50 cm
Residual
Residual
S5- 75 cm S5- 75 cm
Figure 2: Percent Distribution (%) Percent Distribution (%)

0
10
20
30
40
50
60
70
0
10
20
30
40
50
60
70
80
90
100
S1- 25 cm
S1- 25 cm
S1- 50 cm
S1- 50 cm
Exchangeable
Exchangeable
S1- 75 cm S1- 75 cm
Sequential extraction results.
S2- 25 cm
S2- 25 cm
S2- 50 cm S2- 50 cm
S2- 75 cm
Carbonates
Carbonates
S2- 75 cm
S3- 25 cm S3- 25 cm
S3- 50 cm S3- 50 cm
S3- 75 cm S3- 75 cm
Location
Location
S4- 25 cm
Fe, Mn-oxides
Fe, Mn-oxides
S4- 25 cm
(d) Zinc
S4- 50 cm S4- 50 cm
(c) Copper
S4- 75 cm S4- 75 cm
Organic
Organic
S4- 100 cm S4- 100 cm

S5- 25 cm S5- 25 cm
S5- 50 cm S5- 50 cm
Residual
Residual
S5- 75 cm S5- 75 cm

C o n c e n t r a t i o n ( m g / L) C on c e nt r a t i o n ( m g / L)
0 100 200 300 400 500 600 700 0 0.2 0.4 0.6 0.8
0 0
20 20
40 40
60 60
80 80
100 100
120 120
140 140
Core S1 Core S2 Core S3 Core S4 Core S5 Core S1 Core S2 Core S3 Core S4 Core S5
(a) Sulphate (b) Sulphide

C oc ne nt r a t i on ( mg/ L)
C on c e nt r a t i on ( m g / L )
0 100 200 300 400
0 0.1 0.2 0.3 0.4 0.5 0.6
0
0
20
20
40
40
60
60
80
80
100
100
120
120 140
140 160
Cor e S1 Cor e S2 Core S3 Cor e S4 Cor e S5 Core S1 Core S2 Core S3 Core S4 Core S5
(c) Ferrous iron (d) Calcium

Conc e nt r a t i ons ( mg/ L) C o n c e nt r a t i on ( m g / L )
0 2 4 6 8 10 0 0.2 0.4 0.6 0.8 1
0 0
20 20
40 40
60 60
80 80
100 100
120 120
140 140
160 160
Cor e S1 Core S2 Core S3 Core S4 Core S5 Core S1 Cor e S2 Core S3 Core S4 Core S5
(e) Arsenic (f) Cobalt

Conc e nt r a t i on ( mg/ L) C onc e nt r a t i on ( mg/ L)
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 0.1 0.2 0.3 0.4 0.5 0.6
0 0
20 20
40 40
60 60
80 80
100 100
120 120
140 140
160 160
Core S1 Core S2 Core S3 Core S4 Core S5 Cor e S1 Core S2 Cor e S3 Cor e S4 Core S5
(g) Copper (h) Zinc

Figure 3: Porewater concentrations.
A reduction in sulfate concentrations with depth was observed in all of the

cores. Increasing or measurable Fe(II) and HS- concentrations with depth were
only noted in cores S3, S4, and S5, a further indication of the development of
reducing conditions with depth. Porewater DO concentrations remained fairly
consistent with depth, ranging from 1.50 to 2.15 mg/L. The pH values were also
relatively consistent throughout the porewater profile, averaging 7.0.
3.3 Vegetation
Table 1 presents the results of selected metal extractions from the T. latifolia leaf
samples collected. Significant concentrations of selected metals were taken up
into the leaf samples. Higher metal concentrations were extracted from leaf
samples collected near the wetland inlet, which was expected due to higher metal
loading in this area, and is supported by other studies (O’Sullivan et al. [4];
Miller et al. [5]).
Table 1: Extracted metal concentrations from T. latifolia.
Table 2: Microbiological enumeration results.
3.4 Microbiology
Table 2 presents the results from the bacterial enumerations. There appear to be
equally strong numbers of acid producing (oxidizing) and reducing bacteria, with
populations ranging from 102 to 106 MPN/gram dry weight. The presence of all
three types of microorganisms throughout the wetland and at relatively
consistent concentrations with depth at each core location may suggest that both
oxic and anoxic conditions exist throughout the wetland.
4 Discussion
This natural wetland appears to be an overall sink for metals, with the majority
of the metals being tied up in the sediments or being taken up into the leaves of
Typha latifolia. A detailed mass balance would be required to confirm this. In
this study, results from the SEM indicated that most of the metals were
associated with the residual and OM fractions of the sediment. The residual
fraction accounts for silicate minerals and sulfides. This is important since one of
the sinks for select metals is the formation of metal sulfide precipitates under
reducing conditions. Metals associated with the exchangeable fraction are
typically highly mobile and are often correlated with the cation exchange
capacity of the sediment (Gambrell et al. [6]). The surfaces of oxides and OM are
also prime adsorption sites for metals. Previous studies (Fuller [7]; Lagmuir et al.
[8]; Soprovich [9]) have shown that As readily adsorbs to or coprecipitates with
Fe and Mn oxides. The results from the SEM indicated that much of the Fe was
associated with the residual fraction of the sediment and not the oxide fraction,
while much of the As was associated with the residual and OM fractions of the
sediment. Thus, it would appear that given the presence of adequate Fe and Mn
oxides under oxidizing conditions and adequate sulfides or OM under reducing
conditions, As is likely to be well attenuated in this wetland. Zinc and Cu were
predominantly associated with the OM and residual fractions of the sediments.
Cobalt was primarily associated with the residual fraction, however, it was also
significantly associated with carbonates. Cobalt has known affinities for
carbonates (Brookings [10]) and therefore competition for sorption sites between
OM and carbonates are likely.
The porewater sulfate and sulfide concentration profiles indicated that sulfate
reduction was occurring throughout the wetland. The microbiology results
identified the presence of APB populations in consistent numbers with the SRB
and IRB populations. This may indicate that all these bacteria are active and the
acid produced by these populations is immediately neutralized by the dissolution
of carbonate minerals within the tailings. Above pH 6.3, which was the case
throughout much of the study area, the dissolution of one mole of calcite
consumes one mole of H+ (Blowes et al. [11]). This would also account for the
elevated dissolved concentrations of Ca2+ in the pore water. It is also possible
that APB populations are supported at depth throughout the wetland in the
vicinity of vegetation root zones. Oxygen is transported by wetland plants to the
root zones generating localized zones of oxidation. This could allow for
sustainable populations of APBs throughout the wetland. In regions located away
from the root zones, conditions are likely anoxic, which are favourable for
biogenic iron and sulfate reduction. It is also likely from the microbiology results
that there is considerable competition between bacterial species for organic
substrates. This was supported by changes in population trends at different
locations in the study area, where small changes in sulfate and Fe(II)
concentrations occurred. Prior studies have shown competitiveness between SRB

and IRB, which appears to be dependent on electron donor availability and pore
water chemistry (Fortin, pers. comm..). It should also be noted that Fe reduction
can occur both biotically and abiotically, while sulfate reduction can only be
mediated by SRB.
The results from the extraction of metals from T. latifolia leaf samples
indicated that up to 20% of the total mass of metals was taken up by the leaves.
Previous studies have reported metal uptake in both the leaves and root zones of
T. latifolia and found much higher percentages of metals associated with the
roots than the leaves (Jackson et al. [12]). This suggests that phytoremediation
may be an even more powerful source of metal immobilization in this wetland.
There has been conflicting research on the ability of wetland plants to
immobilize metals in the long term. This is in part due to the fact that plants
transport oxygen to their root zones generating a zone of radial oxygen loss,
which results in the generation of localized oxic conditions in the immediate
vicinity of the roots, resulting in the oxidation of reduced sulfur and iron species.
Under these conditions metals such as As and Co may adsorb or coprecipitate
with the Fe oxides, however, metals such as Zn would likely remain mobilized in
the dissolved state (Jacob and Otte [13, 14]). Background conditions were not
known for the T. latifolia in the area, which would have been useful in
determining the degree of metal uptake by T. latifolia as a direct result of the
mine tailings deposits. However, the vegetation was flourishing in the area and
thus it was not expected that the presence of metals at these concentrations were
having a negative effect on T. latifolia. It is possible, however, that the growth of
other wetland plants was inhibited due to elevated metal loading to the wetland
system. There was also considerable variability between sample locations. This
could most likely be attributed to variations in metal loadings at each location.
5 Conclusions
The data collected suggest that this is a relatively stable system. Should the
current state of the system change, such as the input of a waste stream elevated in
organic substrate, it is likely that a considerable amount of the metals retained
within this system would become mobilized. Over time, the geochemical
processes regulating metal mobilizations throughout the system would change,
resulting in different biogeochemical controls on the metals throughout this
system. Alkaline drainage systems are geochemically different than acidic
drainage systems. Alkaline systems can immobilize metals under oxic
conditions, such as adsorption onto OM or oxide precipitates or uptake by
wetland vegetation, and anoxic conditions via microbial transformations, such as
biogenic sulfate reduction, generating reduced metal sulfide precipitates. Acidic
drainage systems typically require the formation of strongly reducing conditions
in order to immobilize metals. These results have demonstrated the importance
of considering the geochemical conditions of the wetland system treating the
mine drainage, as well as having a detailed understanding of the metals of
concern within the mining waste, as these will geochemically interact based on
redox conditions, presence of sulfides, Fe and Mn oxides, and organic matter.
References
[1] Schnoor, J., Modeling trace metals (Chapter 8). Environmental Modeling:
Fate & Transport of Pollutants in Water, Air, Soil, eds. J. Schnoor & A.
Zehnder, Wiley, New York, 381-451, 1996.
[2] Cochran, W.G., Estimation of bacterial densities by means of the most
probable number. Biometrics, 6, pp. 105-116, 1950.
[3] Tessier, A., Fortin, D., Belzile, N., DeVitre, R, & Leppard, G.G., Metal
sorption to diagenic iron & manganese oxyhydroxides & associated
organic matter: narrowing the gap between field & laboratory
measurements. Geochimica et Cosmochimica Acta, 60, pp. 387-404, 1996.
[4] O’Sullivan, A.D., Moran, B.M., & Otte, M.L., Accumulation & fate of
contaminants in substrates of wetlands constructed for treating mine
wastewater. Water, Air, & Soil Pollution, 157, pp. 345-364, 2004.
[5] Miller, W., McFee, W., & Kelly, J., Mobility and retention of heavy
metals in sandy soils. J.Environmental Quality, 12(4), pp. 579-584, 1983.
[6] Gambrell, R.P., Wiesepape, J., Patrick Jr., W., & Duff, M., The effects of
pH, redox, and salinity on metal release from a contaminated sediment.
Water, Air, & Soil Pollution, 57-58, pp. 359-367, 1991.
[7] Fuller, J., Surface chemistry of ferrihydrite: Part 2. Kinetics of arsenate
adsorption and coprecipitation. Geochimica et Cosmochimica Acta, 57,
pp. 2271-2282, 1993.
[8] Lagmuir, D., Mahoney, J., MacDonald, A., & Rowson, J., Predicting
arsenic concentrations in the porewaters of buried uranium mill tailings.
Geochimica et Cosmochimica Acta, 63(19/20), pp. 3379-3394, 1999.
[9] Soprovich, E., Arsenic Release from Oxide Tailings Containing Scordite,
Fe-Ca Arsenates, & As-Containing Geothites. Environmental Protection
Branch-Yukon Division, Environment Canada, Whitehorse, Yukon, 1995.
[10] Brookings, D.G., Eh-pH Diagrams for Geochemistry. Springer-Verlag,
USA, 1988.
[11] Blowes, D.W., Jambor, J.L., & Hanton-Fong, C.L., Geochemical,
mineralogical and microbiological characterization of a sulfide-bearing
carbonate-rich gold mine tailings impoundment, Joutel, Quebec. Applied
Geochemistry, 13(6), pp. 687-705, 1998.
[12] Jackson, L.J., Kalff, J., Rasmussen, J.B., Sediment pH & redox potential
affect the bioavailability of Al, Cu, Fe, Mn & Zn to rooted aquatic
macrophytes. Can. J. Fish & Aquatic Science, 50, pp. 143-148, 1993.
[13] Jacob, D.L., Otte, M.L., Long-term effects of submergence and wetland
vegetation on metals in a 90 year old abandoned Pb-Zn mine tailings
pond. Environmental Pollution, 130, pp. 337-365, 2004.
[14] Jacob, D.L., Otte, M.L., Influence of Typha latifolia & fertilization on
metal mobility in two different Pb-Zn mine tailings types. Science of the
total environment, 333, pp. 9-24, 2004.
A sequential aerated peat biofilter system for

the treatment of landfill leachate
P. Champagne1 & M. Khalekuzzaman2
1
Department of Civil Engineering, Queen’s University, Canada
2
Department of Civil and Environmental Engineering, Canada
Abstract
In recent years, researchers have identified peat as an alternative low-cost filter
medium for on-site wastewater treatment, including landfill leachate. Peat
possesses several physical, chemical and biological characteristics that make it a
favorable filter medium for the mitigation of contaminants. The effectiveness
and the impact of clogging of peat biofilter in terms of organic (COD, CBOD5),
ammonia (NH3-N) and total suspended solid (TSS) loading are crucial in the
operation of such systems. The main purpose of this research was to evaluate the
performance of a bench-scale sequential aerated peat biofilter system treating
landfill leachate at different hydraulic loading rates (HLRs) under continuous
flow condition. The system consists of two major components: an aeration
chamber with an attached growth media, followed by a peat biofilter. The
leachate was aerated at a constant air flow rate of 3.40 m3/day for a hydraulic
retention times (HRTs) of 2 or 5 days. The aerated leachate was then fed to two
sets of triplicate peat columns, which were operated at average HLRs of 8.28
cm3/cm2/day and 10.82 cm3/cm2/day. The result of the study showed that similar
CBOD5, COD, NH3-N and TSS removal efficiencies and column life
expectancies could be obtained from the two different hydraulic loading rates to
the peat biofilter. However, the HRT in the aeration basin was found to
significantly increase the life expectancy of the peat biofilter by reducing the
overall contaminant loading to the biofilter. For a HRT of 5 days and constant air
flow rate of 3.4 m3/day 99% NH3-N was removed in the aeration tank after 3
weeks. Removal efficiencies above 80%, 90% and 86 % were noted for COD,
CBOD5 and NH3-N, respectively, in the peat columns after 6 weeks of operation.
Keywords: peat, landfill leachate, aeration, biofilm, hydraulic loading, leachate
treatment.
doi:10.2495/GEO060131
1 Introduction
The Trail Road landfill in the City of Ottawa, commissioned in 1980, generates
an average rate of 190 m3 of leachate per day. Currently, leachate from Trail
Road landfill is hauled by tanker truck for treatment and discharge at the Robert
O. Pickard Environmental Center (ROPEC), the City’s wastewater treatment
facility. However, the concentrations of several contaminants of the leachate
exceed or closely approach the City’s Sewer Use By-law limit, particularly TKN,
TSS, BOD5, H2S, boron, chloride, xylene, toluence, and barium. As such, the
solid waste disposal facility must pay a surcharge for those contaminants that
exceed the City Sewer Use By-Law limits. An on-site treatment system pre-
treating the landfill leachate could reduce operational costs to the landfill, by
bringing the landfill leachate to compliance with the Sewer Use By-law limits.
In recent years, studies (Heavey, [1]; Kinsley et al., [2]; Kennedy and Van
Geel, [3]; Lyons and Reidy, [4]; Talbot et al., [5]; Viraraghavan and Ayyaswami,
[6]; Rock et al., [7]) have identified peat as an alternative low-cost filter medium
for on-site wastewater treatment, including landfill leachate. Peat possesses
several characteristics that make it a favorable filter medium for contaminants
removal, such as high water holding capacity (Bergeron, [8]), low density
(Buttler et al., [9]), large surface area (>200 m2/g) (McLellan and Rock, [10]),
high porosity (Mclellan and Rock, [10]; Buttler et al., [9]; Mitsch and Gosselink,
[11]), and excellent ion exchange properties (Sharma and Forster, [12]; Mckay,
[13]). The properties of peat depend on several factors, including the ambient
conditions during its formation, the extent of its decomposition and the method
of harvesting (Couillard, [14]). To date, limited information regarding the
behavior of peat filter systems under varying contaminant, as well as hydraulic
loading rates when operated in a biofilter configuration is available. In addition,
the treatment efficiency and the total operational life of the peat filter systems,
are vulnerable to varying contaminant loads, particularly organic (COD, BOD5),
NH3-N, and TSS concentrations, as well as hydraulic loading rates.
2 Methodology
Contaminant removal efficiency and the total operational life of the peat biofilter
system are dependent on organic (COD, BOD5), NH3-N, and TSS constituent, as
well as hydraulic loading rates (HLR). The removal efficiency and operational
life of a peat biofilter preceded by an aeration chamber with a support media to
promote the growth of an attached biofilm were investigated, under different
hydraulic and contaminant loading rates and continuous flow condition. The
attached growth medium provides a large active surface area and texture, which
can promote the rapid growth of a biofilm, thereby, reducing contaminant loads,
particularly NH3-N and BOD5, on the peat filter leading to an increase in the
operational life of the peat biofilter system. Laboratory investigations were
conducted using the bench-scale experimental set-up illustrated in Figure 1.
Peristalti
c Pump
Distilled
Water
Peristaltic
Pump
Triplicate Peat Column Triplicate Peat Column
Avg. 8.28 cm3/cm2/day Avg. 10.82 cm3/cm2/day
Control Column
Avg.10.82 cm3/cm2/day
Attached Growth Media
Diffuser
Raw
Leachate
Air Pump
Aeration Basin
Figure 1: Sequential aerated peat biofilter experimental set-up.
Raw leachate was collected in 20 L (28 cm X 23 cm X 40 cm) containers

from the City of Ottawa Trail Road landfill and stored in a refrigerators at 4ºC.
The raw leachate was passed through a cylindrical aeration tank (64 cm X 44 cm
ID) using a peristaltic pump at a flow rate 4.5 L/day, which was equal to the sum
of the influent rates of the peat filter columns. An air pump, MAP2X Maxair
2XL, was utilized to inject air into the leachate at a constant air flow rate of
3.40 m3/day. To aerate effectively, a 28 cm long perforated hose with a 1 cm OD
was placed at the base of the aeration tank. In addition, a spun plastic medium
was used in the aeration basin to get a better performance of the aeration basin
by providing a support medium for biofilm growth.
The aerated leachate was fed from the aeration basin to two sets of triplicate
peat columns by two sets of peristaltic pumps. One set of triplicate columns was
fed at a rate of 8.28 cm3/cm2/day, while the other set was fed at a rate of
10.82 cm3/cm2/day. Masterflex® TYGON tubing was used to connect each of the
stages: from the raw leachate container to the aeration basin, from the aeration
basin to the peat column inlet, and from the distilled water container to the
control peat column inlet. There were also two calibrated Masterflex® peristaltic
pumps, maintaining constant flows from the raw leachate to the aeration basin,
and from the distilled water container to the control peat column. Two sets of
pumps were engaged to feed the two sets of triplicate peat columns with aerated
leachate from the aeration basin. Each pump set was assembled with three
Masterflex® Easy-Load® pump heads attached to a Masterflex® peristaltic
pump to ensure a constant flow rate. The pumps were attached to a GRASSLIN
(model CP-924) timer which turned all pumps on intermittently five times a day,
for a total of ten minutes per day.
Samples of the raw leachate, aerated leachate and column effluents were
collected and analyzed for COD, CBOD5, NH3-N, NO3-N, and TSS removal in
order to assess the performance of the aeration basin with biofilm growth for
contaminants removal at 5-day and 2-day HRTs, as well as the removal
efficiencies and life expectancies of the peat biofilters. A blank column was
operated with distilled water in the same manner as the higher HLR an average
10.82 cm3/cm2/day to observe the potential leaching of constituents from the peat
and the behavior of the peat filter under control condition.
The contaminant load in the Trail Road landfill leachate was much higher
than is typically reported for untreated domestic wastewater especially in terms
of the higher ammonia-N, TSS, COD, and CBOD5 concentration. The average
influent COD, CBOD5, NH3-N, NO3-N, and TSS concentration were 899 mg/L,
340 mg/L, 511 mg/L, 2 mg/L, and 51 mg/L, respectively for the 5-day HRT. The
average influent COD, CBOD5, NH3-N, NO3-N, and TSS concentration were
1052 mg/L, 534 mg/L, 392 mg/L, 2 mg/L, and 135 mg/L, respectively, for the 2-
day HRT. Therefore, these high contaminant concentrations indicate that the
leachate is a high-strength wastewater in comparison to municipal wastewater.

3.1 Aeration basin
The results of this study showed that the aeration basin did not significantly
remove COD from the raw leachate for both the 5-day and 2-day HRTs, while
CBOD5 concentrations in the aeration basin were observed to decrease from an
average 340 mg/L and 534 mg/L to 98 mg/L and 139 mg/L for the 5-day and
2-day HRTs, respectively as shown in Figure 2(a). The TSS concentrations of
aerated leachate were observed to decrease prior to days 70 and 78 for the 5-day
and 2-day HRTs, respectively. Then the TSS concentration of aerated leachate
was found to exceeded the raw leachate TSS concentration, which is likely due
to the fact that sludge in the aeration basin was not collected and disposed of
throughout the course of each experimental HRT (Fig. 2(b)). From Figure 3(a), it
can be noted that steady-state removal of NH3-N was observed for the 5-day
HRT after approximately 2 weeks of operation, while similar NH3-N removal
was not observed for the 2-day HRT even after 3 weeks of operation. The 5-day
HRT also exhibited better nitrification than the 2-day HRT. In addition, an
average NO3-N generation of 108 mg/L was found for the 5-day HRT compared
to 21 mg/L for the 2-day HRT (Fig. 3(b)). Denitrification was also noted in the
aeration basin after 44 and 42 days of operation for the 5-day and 2-day HRTs,
600 400
5-day HRT: 350 5-day HRT:

500
300
400
CBOD5 (mg/L)
250
TSS (mg/L)
300 200
200
150
100
100
50
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Day Day
600
2-day HRT: 300
Geo-Environment and Landscape Evolution II

500 250 2-day HRT:
400 200
CBOD5 (mg/L)
TSS (mg/L)
300 150
200 100
100 50
0 0
0 20 40 60 80 100 120
0 20 40 60 80 100 120
Day Day
Raw AB Control Column Raw AB Control Column
C-1: Avg. 8.28 cm3/cm2/day C-2: Avg. 8.28 cm3/cm2/day C-3: Avg. 8.28 cm3/cm2/day C-1: Avg. 8.28 cm3/cm2/day C-2: Avg. 8.28 cm3/cm2/day C-3: Avg. 8.28 cm3/cm2/day
(a) (b)
129
Figure 2: CBOD5 and TSS of raw leachate, aerated leachate and column effluents for the 5-day and 2-day HRTs.
1200 400 5-day HRT:
5-day HRT: 350

1000
Ammonia-N (mg/L) 300
800
Nitrate-N(mg/L)
250
600 200
150
400
100
200
50
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120

Day
Day
600 200
180 2-day HRT:
500 2-day HRT: 160
Ammonia-N (mg/L)
140
400
Nitrate-N(mg/L)
120
300 100
80
200 60
40
100
20
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120

Day Day
Raw AB Control Column Raw AB Control Column
(a) (b)
Figure 3: NH3-N and NO3-N of raw leachate, aerated leachate and column effluents for the 5-day and 2-day HRTs.
respectively. The aeration basin allowed for NO3-N removal as a result of the
rapid formation of a biofilm onto the attached growth media in the aeration
basin. As the microorganisms grow, the thickness of the biofilm layer increase,
and the diffused oxygen is consumed before it can penetrate the full depth of the
biofilm layer. Thus, an anaerobic environment is established near the surface of
the media, which is likely the main mechanism for NO3-N removal in the
aeration basin after an extended period of operation. The concentration of NO3-N
was observed to decrease from 319 mg/L (day 44) and 96 mg/L (day 42) to
90 mg/L (end) and 1 mg/L (end) for the 5-day and 2-day HRTs, respectively.
3.2 Peat columns
In this study, two sets of triplicate columns were operated at HLRs of

8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively, for both the 5-day and
2-day HRTs. The averages of the triplicate column effluents are used in the
discussion of the results. The average effluent concentrations from the peat
columns were 356 mg/L and 383 mg/L at HLRs of 8.28 cm3/cm2/day and
10.82 cm3/cm2/day, respectively, for the 5-day HRT; while average effluent
COD concentrations of 413 mg/L and 415 mg/L were observed for the 2-day
HRT at the same HLRs. From the control column effluent COD concentrations,
it was found that the peat contributed COD to the effluents, with average COD
concentrations of 39 mg/L for both the 5-day and 2-day HRTs. As a
consequence, the overall COD removal was limited. Biodegradation of organic
matter was observed in the peat columns through CBOD5 removal (Fig. 2(a)).
Average effluent CBOD5 concentrations of 22 mg/L and 24 mg/L for
8.28 cm3/cm2/day and 10.82 cm3/cm2/day HLRs, respectively, were noted for the
5-day HRT. Concentrations of 18 mg/L and 29 mg/L were obtained for the 2-day
HRT, for the 8.28 cm3/cm2/day and 10.82 cm3/cm2/day HLRs, respectively.
Average effluent TSS concentration of 9 mg/L and 6mg/L in 5-day HRT, and
34 mg/L and 42 mg/L in 2-day HRT, were found for the HLRs of
8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively (Fig. 2(b)). The TSS
removal was achieved through adsorption and physical filtration via its porous
structure. Effluent NH3-N concentrations were less than 2.18 mg/L and
2.15 mg/L after 30 days of operation at 8.28 cm3/cm2/day and
10.82 cm3/cm2/day, respectively, for the 5-day HRT (Fig. 3(a)). Comparatively,
NH3-N effluent concentrations were below 4.29 mg/L and 5.30 mg/L after
36 days of operation at 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively,
for the 2-day HRT. The concentrations steadily increased at the end of the
experimental run before clogging of the columns were observed. The likely
mechanisms of NH3-N removal were adsorption of NH4+ onto the peat up to the
CEC saturation for NH4+, followed by leaching of NH3-N, nitrification and
denitrification. An average, NO3--N concentration of 121 mg/L and 119 mg/L in
5-day HRT, and 38 mg/L and 48 mg/L in 2-day HRT were observed for the
HLRs of 8.28 cm3/cm2/day and 10.82 cm3/cm2/day, respectively (Fig. 3(b)).
Denitrification in the peat columns was generally observed after 49 and 36 days
of operation; this might be due to the formation of anoxic zones at the bottom of
the columns.
One of the main objectives of this research was to investigate the lifetime of
the peat biofilter system under different contaminant loadings, HRT in aeration
basin, as well as hydraulic loading rate. The operational life of each of the peat
filters was defined as the number of days of operation between when the peat
columns were initially fed with leachate to the time clogging was observed as
exhibited by surface ponding. The total cumulative COD, BOD5, and TSS
removal of peat columns at the time of clogging for the two sets of triplicate
columns were computed under the different operational conditions. The results
are summarized in Figure 4 and Table 1.
A single factor ANOVA was conducted with an alpha value of 0.05 to
statistically compare the performances of peat columns operated under different
conditions. The results of this study indicated that statistically similar total
cumulative organic (COD, CBOD5) removals were observed in the peat columns
under different HLRs and HRTs since the F values were always less then Fcritical
values in the ANOVA test. However, the higher 5-day HRT of the aeration basin
increased the operational life of the peat biofilters when compared to the 2-day
HRT through the lowering of the contaminant loading onto peat biofilters.
12
5-day HRT:
11
Hydraulic Loading Rate (ml/day)
10
9
8
7
6
5
4
3
0 20 40 60 80 100 120
Day
12
2-day HRT:
Hydraulic Loading Rate (ml/day)
11
10
9
8
7
6
5
4
0 20 40 60 80 100 120
Day
Control Column
Avg. 8.28 cm3/cm2/day: Col. Avg.
Avg. 10.82 cm3/cm2/day: Col. Avg.
Figure 4: Operational life and cumulative contaminant removal of peat

columns.
Table 1: Operational life and cumulative contaminants removal of peat filters.
Total Cumulative Removal

Column Operational (mg/ g of Peat)
Phase
ID Life
(day) COD BOD TSS
Controlled Column(DW) No Clogging ― ― ―

Column 1 104 34.68 6.42 10.92
Avg. 8.28
Column 2 108 46.88 9.42 15.28
5-day HRT
cm3/cm2/day
Column 3 115 48.12 8.86 15.59
Column 1 108 41.31 7.54 14.96
Avg. 10.82
Column 2 101 48.74 10.42 16.71
cm3/cm2/day
Column 3 101 42.06 8.17 14.37
Controlled Column(DW) No Clogging ― ― ―
Column 1 82 30.04 7.65 2.91

Avg. 8.28
Column 2 64 20.90 5.51 1.40
2-day HRT
cm3/cm2/day
Column 3 93 37.79 9.57 4.23
Column 1 93 51.68 13.50 5.20

Avg. 10.82
cm3/cm2/day Column 2 64 31.10 5.80 1.32
Column 3 82 46.77 10.60 3.26
4 Conclusions
One of the main objectives of this research was to investigate the total
operational life of the peat biofilter under varied contaminant loading and
hydraulic loading rates. The contaminant loadings to the peat columns were
considered to be a function of the HRT in the aeration basin. The results of this
research showed that the impact of the hydraulic loading rate was less significant
than the effect of contaminant loading rate leading to a longer life of the peat
filters. Statistically similar organic (COD, CBOD5) removal performances and
life expectancies could be obtained at hydraulic loading rates of
8.28 cm3/cm2/day and 10.82 cm3/cm2/day in both 5-day and 2-day HRTs.
However, the higher HRT, 5 days, increased the life expectancy of the peat
biofilter by approximately one month, due to the considerable decrease in the
organic, NH3-N, and TSS loading through the aeration basin. The results also
suggested that the contaminant removal efficiencies of the peat biofilter columns
were similar for the 8.28 cm3/cm2/day and 10.82 cm3/cm2/day HLRs.
The results indicated that the peat columns were unstable during the first
month of operation, since leaching of COD to effluents by peat itself and
saturating of CEC for ammonia-N followed by leaching of ammonia-N was
observed during the first month of operation. The aeration basin with support
media for biofilm growth was primarily effective for the removal of NH3-N and
NO3-N through nitrification and denitrification. Steady-state nitrification was
initially observed in the aeration basin after approximately 2 to 3 weeks of
operation as this was likely the time required for the steady-state development of
a biofilm on the attached growth media to which NH3-N removal was attributed.
Therefore, an anaerobic environment was established near the surface of the
media, which was mainly responsible for denitrification in aeration basin after
approximately 1.5 months of operation at both the 5-day and 2-day HRTs. From
this study, it can be noted that HRT was a limiting factor affecting the
contaminants removal efficiencies of aeration basin. Therefore, an increase in
HRT would increase the removal of contaminants.
References
[1] Heavey, M., Low-cost Treatment of Landfill Leachate Using Peat. Waste
Management, 23, pp. 447-454, 2003.
[2] Kinsley, C., Crolla, A., & Fernandez, L., Treatment of Landfill Leachate
using a Peat Filter: Final Report. City of Ottawa, Ottawa, ON, 2003.
[3] Kennedy, P. & Van Geel, P., Hydraulic of Peat Filters Treating Septic
Tank Effluent. Transport in Porous Media, 41, pp. 47-60, 2000.
[4] Lyons, H.J. & Reidy, T.J., The Use of Peat in Treating Landfill Leachate.
Humic Substances in Soils, Peats and Waters: Health and Environmental
Aspects, eds. M.H.B. Haynes & W.S. Wilson, UK Royal Society of
Chemistry, pp. 475-485, 1997.
[5] Talbot, P., Bélanger, G., Pelletier, M., Laliberté, G., & Arcand, Y.,
Development of a Biofilter Using an Organic Medium for On-site
Wastewater Treatment. Water Science & Technol., 34, pp. 435-441, 1996.
[6] Viraraghavan, T. & Ayyaswami, A., Use of Peat in Water Pollution
Control; A Review. Canadian J. of Civil Engineering, 14, pp. 230-233,
1987.
[7] Rock, C.A., Brooks, J.L., Bradeen, S.A., & Struchtemeyer, R. A., Use of
Peat for On-Site Wastewater Treatment: I. Laboratory Evaluation. J.
Environ. Qual., 13(4), pp. 518-523, 1984.
[8] Bergeron, M., Peat. Canadian Minerals Yearbook, pp. 37.1-37.8, 1987.
[9] Buttler, A., Dinel, H., & Levesque, P.E.M., Effects of Physical, Chemical,
and Botanical Characteristics of Peat on Carbon Gas Fluxes. Soil Sci.,
158(5), pp. 365-374, 1994.
[10] McLellan, J.K. & Rock, C.A., Pre treating Landfill Leachate with Peat to
Remove Metals. Water, Air, & Soil Pollution. 37; pp. 203-215, 1988.
[11] Mitsch, W.J. & Gosselink, J.G., Wetlands, Van Nostrand Reinhold Co.,
New York, 1993.
[12] Sharma, D.C. & Forster, C.F., Removal of Hexavalent Chromium Using
Sphagnum Moss Peat. Water Research, 27(7), pp. 1201-1208, 1993.
[13] Mckay, G., Use of Adsorbents for the Removal of Pollutants from
Wastewater, CRC Press, 1996.
[14] Couillard, D., Review: The Use of Peat in Wastewater Treatment. Wat.
Res., 28(6), pp. 1261-1274, 1994.
Are crop and range lands being contaminated

with cadmium and lead in sediments
transported by wind from an adjacent
contaminated shallow lake?
H. O. Rubio1, T. R. Saucedo1, M. R. Bautista2, K. Wood3,
C. Holguin3 & J. Jiménez4
1
Campo Experimental la Campana-Madera del Instituto Nacional de
Investigaciones Forestales, Agrícolas y Pecuarias, México y Facultad de
Zootecnia de la Universidad Autónoma de Chihuahua, México
2
Universidad Juárez Autónoma de Tabasco
3
New Mexico Water Resources Research Institute, Las Cruces,
New Mexico, USA
4
Facultad de Zootecnia de la Universidad Autónoma de Chihuahua,
México
Abstract
Soil contamination by cadmium (Cd) and lead (Pb) in range and crop lands can
occur from polluted sediments carried by the wind. Soils samples were obtained
in three different areas on the leeward (east) side of a shallow lake called the
Laguna de Bustillos: (1) sediment in the shallow lake (S), (2) soil in the range
area adjacent to the shallow lake (RA), and (3) soil in agriculture land (AL)
adjacent to the range area. In each area, three composite soil samples were
randomly collected at depths of 0-15, 15-30 and 30-50 cm. Therefore, 45 soil
samples were analyzed as a 3 (area) x 3 (depth) factorial arrangement. The
samples were air dried, passed through a 2.0 mm sieve, ground and passed
through a 0.355 mm sieve, and digested with concentrated nitric acid. The metals
Cd and Pb were determined using ICP optical emission spectrometry. In addition
pH and CE were evaluated. Cadmium concentration was statistically different
among areas and showed a strong interaction between depth x area. Maximum
Cd concentration was found in S with 0.94 mg kg-1 in the surface horizon (0-15
cm). Lead concentration was different among areas and the interaction was not
as strong as the one noted for Cd. Highest Pb concentrations were observed in
the surface horizons (0-15 cm) of both S and RA with 74.19 mg kg-1 and 55.09
mg kg-1. Nevertheless, for the AL system the maximum Pb concentration was
found in the subsurface horizon (30-50 cm) with 40.23 mg kg-1. It is concluded
that Cd and Pb are contaminating the RA and AL through wind movement with
Pb contamination being the worst.
Keywords: cadmium, lead, soil contamination, laguna Bustillos, Chihuahua,
México.
doi:10.2495/GEO060141
1 Introduction
Cadmium (Cd) and lead (Pb) are considered to be at the top of the five most
important elements in terms of food-chain contamination. These two metals,
when present in highly contaminated soils may have a negative human
reproductive outcome [1]. It is generally recognized that the presence of these
metals in soils is usually thorough antrophogenic activities such as agriculture
(fertilizers, liming materials, agrochemicals), industry (atmospheric deposition,
wastes) and urban activities (sewage sludges, drainage deposition) whereas other
elements in soils are often derived from the parent rock material which the pedo-
genetic processes are carried out. Once in soil, the metals as contaminants, often
attach to soil particles that can migrate from one place to another.
Evaluation of heavy metals contamination is often based on comparison with
the background concentration in a given environment [2], evaluating the effect
on plant growth or soil utilization [3, 4] and determining the hazard to human
health [5, 6]. It is generally accepted that soil sediment is a more complex
medium than water. This is true because soils include water as well as solid and
gas phases, and soils do not move from one place to another as fast as does water
[7], and determining concentrations of heavy metals is a difficult task [8].
The water of the Laguna de Bustillos in Chihuahua, México is highly
contaminated [9, 10] and it is suspected that its soil sediments are also
contaminated. In the particular case of the Laguna de Bustillos environment, the
sediment is exposed because the water of the shallow lake is absent during large
periods of time. For example, during the year 2005 (January to July) most of the
shallow lake was completely dry. In addition, strong wind events are present in
the period from February to April every year. Therefore, the question that arises
is whether the soil sediment from the shallow lake is contaminating the close
rangeland and agriculture land. The objective was to evaluate the level of soil
contamination with Cd and Pb metals in three different environments; lake
sediments, rangeland, and agriculture land. To our knowledge, this is the first
report on contaminant risk assessment of rangeland and cropland soils from a
polluted soil in Mexico.

The study was conducted on the east side of the Laguna de Bustillos, located in
the state of Chihuahua, México. Its location is in the polygon: latitude 28° 58´
12”-28° 15´ 00” N; longitude 107° 09´ 36”-106° 15´ 00” W and at 2,300 meters
elevation (Figure 1). Annual precipitation averages about 480 mm mostly as rain
during the summer (July to September), but some snow events occur during the
winter (December to February). The source is a shallow lake, heavily silted due
to soil erosion coming from a denuded watershed which becomes dry during
drought periods. Wind storms moving from the west to the east side usually
occur every year from late winter (February) through the beginning of spring
(April).
Figure 1: A panoramic view of the Laguna de Bustillos, Chihuahua, Mexico.
Soil samples where obtained from three different areas on the leeward (east)
side of the Laguna de Bustillos. The first system was the sediment of the shallow
lake (S) where the water used to be, but at the time of collecting the sample was
dry. The second area was the rangeland (RL) adjacent to the shallow lake where
the domestic livestock graze during the whole year. The third area was the crop
land (AL) that is close to the RL. The domestic livestock belong to the
commonly owned communities (Ejidos) of Cuitlahuac, La Selva, Centro Calles
and Fabela. An “Ejido” is a community where the law established that any
grazing land must be collectively shared [11] which results in the worldwide
phenomena commonly called “The Tragedy of the Commons”; therefore, the RL
is an extremely deteriorated grassland. The most important crops in the AL are
beans (Phaseoluos vulgaris), corn (Zea mays) and oats (Avena sativa) growing in
the precipitation season under dryland condition.
Five points were randomly selected in the shallow lake (S). Then, at each
point of S a transectal line to the east was selected with five randomly located
points in RL as well as in AL. At each point, three composite soil samples were
taken at 0-15, 15-30, and 30-50 cm profile depths. Hence, 15 soil samples were
taken in each system, giving a total of 45 soil samples. The samples were air
dried and passed through a 2.0 mm sieve. After this, the samples were ground
and passed through a 0.355 mm sieve. They were evaluated for Cd and Pb
concentration as well as pH and EC. The digestion of soil samples for Cd and Pb
evaluation was realized with concentrated nitric acid in the laboratory of the
Faculty of Zootechnic of the Autonomus University of Chihuahua, Mexico,
following the sampling and analysis protocol of Canada (MAF). Cd and Pb
concentrations were determined using an ICP-OES (Inductively Coupled

Plasma-Optical Emission Spectrometry) 2100, Perkin Elmer. Values of pH and
EC were determined in a saturated paste using a standard glass electrode and a
conductivity meter, respectively. The statistical analysis used a factorial
treatment design 3 x 3; where Factor A was the system (S, RL, AL) and Factor B
was the depth (0-15, 15-30, 30-50 cm). The differences were noted using a 0.01
significance level.

The Cd concentration was different among systems. Not surprisingly the higher
Cd concentration was noted in S, with the high concentration noted in the upper
horizon (0-15) where it was detected at 0.946 mg kg-1 in comparison with 0.672
mg kg-1 found at the 15-30 cm depth and 0.614 mg kg-1 noted in the 30-50 cm
depth. In contrast, in RL and Al, the higher Cd concentration was not detected in
the upper profile (0-15 cm); in fact, the lowest concentration in these two areas
was found in the upper horizon (Figure 2). A strong interaction was noted in Cd
concentrations (Figure 2). These results are similar to the levels found in Cd
concentrations in a soil of West Bengal, India where it was detected at 0.37 mg
kg-1 [12], soil samples form England and Wales had a Cd concentration of 0.8
mg kg-1 [13] and China soils had a Cd concentration of 0.07 mg kg-1 [14]. In
other studies, Cd levels were 0.4-0.8 mg kg-1 in Spain [15], levels of 0.6-1.4 mg
kg-1 in asphalt roads in Germany [16] and levels of 0.41 mg kg-1 in Japan [17].
0.9
0.8
0.7
0.6
0.5
0.4
0.3
Agriculture
0.2 Sediment
0.1 Rangeland
0
0 10 20 30 40 50 60
X axis = Soil depth (cm)

Y axis = Concentration (mg kg-1)
Figure 2: Mean of Cadmium level under three systems and different soil depth.
A study in the United States of America [18] found 3,045 surface soil samples
contained Cd values from <0.01 to 2.0 mg kg-1 with a mean of 0.265 mg kg-1. In
any case, the levels of Cd found in the present study do not represent a hazard to
the soils in the three areas, because Dudka et al. [18] concluded that soils with
Cd concentrations of up to 30 mg/kg are still safe for production of some crop
plants. Moreover, the soil contamination due to wind-born Cd was insignificant,
being our hypothesis not tested for this element.
The Pb concentrations were different among areas and no statistical
differences were observed for depth and for the interaction (Figure 3). Maximum
levels of Pb concentration was noted in S with the higher concentration in the
profile 15-30 cm depth with about 75.99 mg kg-1 while the lesser amount of Pb
was observed in AL system in the upper profile with 33.49 mg kg-1 (Figure 3).
A study evaluating surface soils in a greenhouse [15] found Pb concentrations in
a range of 2.5 to 89.9 mg kg-1 which are similar to the results presented in this
study; but the concentration range was shorter from 33.49 to 75.99 mg kg-1.
Another study [19] reported a Pb concentration as high as 59 mg kg-1 in the 0-20
cm depth of the soil profile irrigated with wastewater. Moreover, in a study
carried out in India [12] they reported a mean concentration of 10.4 mg kg-1 in a
soil affected with arsenic. A study carried out in Japan [17] noted a 21 mg kg-1 of
Pb in an uncultivated soil. The values reported here are of practical importance,
because some of the soil fauna like the earthworm (Eisenia fetida Andrei) may
be severely affected with Pb concentration higher than 30 mg kg-1.
80
70
60
50
40
30
20 Agriculture
Sediment
10
Rangeland
0 10 20 30 40 50 60
X axis =Soil depth (cm)
Y axis =Concentration (mg kg-1)
Figure 3: Mean of lead concentration under three different systems and

different soil depth.
A study conducted on road dust in Sydney [20] suggested that deposition and
removal of road dust is not a static process. We could speculate that our findings
suggest that Cd and Pb had similar origin and this is a continuous process. This
means that an equilibrium may be reached when the rate of deposition in a
system is equal to the rate of removal.
4 Conclusion
This study showed that our hypothesis that sediments of the Laguna de Bustillos
are contaminating the rangelands and crop lands with Cd and Pb is true.
However, the Cd concentration does not represent a hazard at least in the area
tested. On the other hand, the Pb concentration in the contaminated area may
represent a potential hazard in the short term.
Acknowledgements
We are deeply grateful with the Produce Foundation of Chihuahua (Fundacion
Produce Chihuahua, A.C.) for partial financial support as well as with the
Faculty of Zootechnic of the Autonomous University of Chihuahua. The INIFAP
(National Research Institute of Forestry, Agriculture and Rangeland) deserves
special thanks for financial support and for facilities given to this research.
References
[1] Laudanski, T., Sipowicz, M., Modszelewski, P., Bolinski, J.,
Szamatowicz, J., Razniewska, G., and Akerlund, M. 1991. Influence of
high lead and cadmium soil content on human reproductive outcome. Int.
J. Gynaecol. Ostet. 36(4):309-315.
[2] Zhang, X.P., Deng, W., and Yang, X.M. 2002. The background
concentrations of 13 soil trace elements and their relationship to parent
material and vegetation in Xizang (Tibet), China. J. Asian Earth Sci.
21:167-174.
[3] Arao, T de N., Sugiyama, M., and Takahashi, M., 2003. Genotypic
differences in cadmium uptake and distribution in soybeans. Plant Soil
251:247-253.
[4] Sheppard, S.C. 2005. Assessment of long-term fate of metals in soils:
Inferences from analogues. A review. Canadian Jour. of Soil Sci. 85(1):1-
18
[5] Storelli, M.M., Barone, G., and Marcotrigiano, G.O. 2005. Cadmium in
cephalopod molluscs: implications for public health. Journal of Food
Protection 68(3):577-588.
[6] Linde, A.R., Sanchez-Galan, S., and García-Vazquez, E. 2004. Heavy
metal contamination of European Eel (Anguilla anguilla) and brown trout
(Salmo trutta) caught in wild ecosystems in Spain. Journal of Food
protection 67(10): 2332-2336.
[7] Peijnenburg, W.J.G.M. and Jager, T. 2003. Monitoring approaches to

access bioaccessibility and bioavailability of metals: matrix issues.
Ecotoxicol. Environ. Saf. 56:63-77.
[8] Chen, M., Ma, L.Q. and Harris, W.G. 1999. Baseline concentrations of 15
trace elements in Florida surface soils. J. Environ. Qual. 28:1173-1181.
[9] CNA, 2003. Comision Nacional del Agua. Estadísticas del agua en
México. Reporte.
[10] Rubio, A.H., Saucedo, R.A., Lara, C.R., Wood, K., and Jimenez, J. 2005.
Water quality in the Laguna de Bustillos of Chihuahua, México. Water
Resources Management III. Editors M. De Conceicao Cunha and C.A.
Brebbia. 155-160.
[11] Cornehls, J.V. 1964. Mexico´s rural road to progress: an analysis of
agrarian reform and agricultural development. Univ. of Texas, Austin,
USA.
[12] Roychowdhury, T., Uchino, T., Tokunaka, H., abd Ando, M. 2002.
Arsenic and other heavy metals in soils from an arsenic-affected area of
West Bengal, India. Chemosphere 49(6)605-618.
[13] McGrath, S.P. and Loveland, P.J. 1992. Heavy metals in soils. In the soil
geochemical atlas of England. B.J. Alloway (ed.). Blackie Academic and
Professional, Glasgow, Scotland.
[14] Wang, Y., and Wei, F.S. (eds.), 1995. Soil environmental element
chemistry. Chinese Environmental Science Press, Peking, People´s
Republic of China.
[15] Gil, C., Boluda, R., and Ramos, J. 2004. Determination and evaluation of
cadmium, lead and nickel in greenhouse soils of Almeria (Spain).
Chemosphere 55(7)1027-1034.
[16] Munch, D. 1992. Soil contamination beneath asphalt roads by polynuclear
aromatic hydrocarbons zinc, lead and cadmium. Science of the Total
Environments 126(1-2):49-60.
[17] Akira, T., Hirofumi, T., Masami, N., Yuichi, T., Toyokazu, U., Shun´ichi,
H., Jiro, I. 2005. Effect of long-term fertilizer application on the
concentration and solubility of major and trace elements in a cultivated
andisol. Soil Sci. Plant Nutr. 51(2)251-260.
[18] Dudka, S., Piotrowska, M., and Terelak, H. 1996. Transfer of cadmium,
lead, and zinc from industrially contaminated soil to crop plants: a field
study. Empoissonnement Pollution 94(2)181-188
[19] Mapanda, F., Mangwayana, E.N., Nyamangara, J., and Giller, K.E. 2005.
The effect of long-term irrigation using wastewater on heavy metal
contents of soils under vegetables in Harare, Zimbabwe. Agriculture,
Ecosystem&Environment 107(2-3)151-165
[20] Ball, J.E., Jenks, R., Aubourg, D. 1996. Dry weather build-up of
constituents on road surfaces. In proceedings of the 7th International
Conference on Urban Storm Drainage. Hanover, Germany 785-790.
The sorption characteristics of PAHs onto soils

in the presence of synthetic and bio surfactant
J.-H. Chang1, Z. Qiang2 & C.-P. Huang3
1
Department of Environmental Engineering and Management,
Chaoyang University of Technology, Taiwan
2
Environmental Research Center,
Department of Civil, Architectural and Environmental Engineering,
University of Missouri-Rolla, USA
3
Department of Civil and Environmental Engineering,
University of Delaware, USA
Abstract
Contaminants such as polycyclic aromatic hydrocarbons (PAHs) are of special
concern due to their high toxicities and hydrophobicities. The high hydrophobic
natures of PAHs enable their strong sorption onto soil or sediments, which
makes it difficult for their removal from the environment. This study was to
explore the sorption characteristics of PAHs with synthetic nonionic surfactant
and biosurfactant in the soil. Two surfactants (one synthetic and one naturally
occurring) Triton X-100 and rhamnolipid were used. Results show that the
partition coefficient of PAHs in the Triton X-100 system decreases with
increasing surfactant concentrations. Throughout the whole concentration range
of studied biosurfactant (0 ~ 30 CMC), the partition coefficient of PAHs in
rhamnolipid system is roughly constant, around 400.0 L/Kg.
Keywords: sorption, soils, PAHs, surfactant, biosurfactant.
1 Introduction
The polycyclic aromatic hydrocarbons (PAHs) are of special concern for the
natural environment due to their high toxicities and hydrophobicities. On the
basis of the characteristic of high hydrophobicity, PAHs are strongly sorbed onto
soil or sediments [1]. As such, it is difficult to remove PAHs from the soils.
Among many remediation techniques, surfactants have been induced in soil-
doi:10.2495/GEO060151
washing, soil-flushing, and pump-and-treat technologies to increase the

solubilization of sorbed hydrophobic contaminants [2]. The sorption-desorption
process has been considered as the most critical mechanism with respect to the
soil remediation. To study the sorption reaction of PAHs in the soil-water
system in the presence of surfactants one is able to gain insight into the
surfactant remediation process of contaminated soils. The purpose of this
investigation is to explore partition coefficients of different PAHs in the presence
of surfactants. In addition to the synthetic surfactant, the biosurfactant is also
evaluated in this study.
1.1 Surfactant fundamentals
Surfactants can be categorized into four groups including cationic, anionic,

nonionic and zwitterionic (both cationic and anionic groups). These
classifications are typically based on the nature of their head group (hydrophilic
head). In aqueous solution, surfactant molecules have the formation of
aggregates known as micelles when the surfactant concentrations exceed a
certain value. The certain concentration is called the critical micelle
concentration, namely, CMC. The CMC value of surfactant is a function of
surfactant type (nonionic type generally has lower CMC than ionic one), and
system conditions (e.g., temperature, hardness) [3]. Compared to the cosolvent
(e.g., alcohols), surfactants exhibit a much higher degree of surface activity and
present the high potential for remediation techniques.
In aqueous solution, micelles consist of two portions: one portion is the
hydrophilic exterior (the hydrophilic heads are oriented to water phase) and the
other is the hydrophobic interior (the hydrophobic tails are oriented to the center
of micelles). Since hydrophobic organic compounds can be dissolved into the
interior of micelle, the solubilities of organics are then enhanced in the solution.
The increased “aqueous solubility” of organic compounds in micelles is referred
to as solubilization. The degree of organic solubilization depends on the class
and the concentration of surfactants.
1.2 Biosurfactant
Biosurfactants are produced by many different bacterial genera. There are a

variety of biosurfactants manufactured and most of them can be classified as
anionic or nonionic biosurfactants. In addition, biosurfactants can also be
classified into several broad groups: glycolipids, lipopeptides,
lipopolysaccharides, phospholipids, and fatty acids/neutral lipids [4]. The largest
and best-studied group of the biosurfactant is glycolipids, which includes the
sophorose-, rhamnose-, trehalose-, sucrose-, and fructose-lipids. In this study,
the rhamnolipid biosurfactant has been chosen as the dissolution agent. The
production and purification of biosurfactants are difficult, which are controlled
by many factors including growth conditions, culturing medium nutrients,
temperature, pH, and agitation method [5]. Molecular weight of biosurfactant
ranges from approximately 500 to 1,500 mw and CMC value of biosurfactant
typically ranges from 1 to 200 mg/L [6].
1.3 PAHs sorption-desorpiton phenomena
PAHs sorption onto the soil in the presence of surfactants can be distributed in
three phases including water, micelles, and the soil. During the sorption-
desorption process, there are mainly three reactions occurring such as: (1) the
dissolution of PAHs in the micelle phase, (2) the sorption of surfactant in the soil
phase, (3) the partition of PAHs between aqueous and soil phase [7]. As all of
three reactions reach equilibrium, the partition coefficient is defined as the ratio
between the concentration of the PAHs in the soil phase and in the liquid phase
(water+micelle), respectively. Although the liquid phase also comprises the
other sub-phases including the dissolved organic matter (DOM) and colloids, the
effect of DOM and colloids in the liquid phase is relatively small and can be
neglected compared to the micelle concentration. The measured partition
coefficient is still an apparent rather than true coefficient between aqueous and
soil phase.
2 Methodology
The soil samples collected from a specific waste site were air-dried and sieved
through a No. 10 standard sieve (2 mm openings). Table 1 shows the physical-
chemical properties of this soil sample with corresponding analytical methods
adopted. The PAHs organic chemicals were purchased from Aldrich Co. The
purity of fluorine, phenanthrene, fluoranthene, and pyrene is 98.0% and that of
anthracene is 97.0%. Table 2 lists the major features of these PAHs organic
chemicals including molecular formula, molecular weight, saturation water
solubility, and logKow. The selected surfactants were Triton x-100 and
biosurfactant (rhamnolipid).
Table 1: Basic physical-chemical characteristics of the soil.
Physical-chemical Result Method

Characteristics
Sand (%) 14.0 Hydrometer
Silt (%) 38.0 Hydrometer
Clay (%) 48.0 Hydrometer
pH 7.6 In 0.01M CaCl2
Organic Matter (%) 1.7 Heating at 105 oC for 2 hours, then
at 360 oC for 2 hours
Batch sorption experiments were conducted with fluorene, phenanthrene,

anthracene, fluoranthene, and pyrene at various CMC values ranging from 1 to 5.
The soil samples were contaminated by PAHs in advance, which concentrations
in soils ranged from 25 µg/g to 200 µg/g. To a series of glass tubes, 2 g of PAHs
contaminated soil samples and 10 ml of various concentration (0 CMC to 30
CMC) surfactant solutions were added. Two surfactants were used to conduct
the sorption experiments including Triton x-100 and rhamnolipid. The tubes
were placed in a shaker and shaken constantly for 72 hours to reach the
equilibrium condition. The mixtures were then centrifuged at 2,500 rpm (1,000
g) for 10 minutes using a Precision Scientific Co. model K-9 centrifuge to
separate the coarse particles from the supernatant. Then, a Sorvall superspeed
refrigerated centrifuge model RC-5 was used to remove fine particles in the
supernatant at the rotation speed of 12,000 g. The concentration of PAHs
organic compounds in the centrate was analyzed with a HPLC/FLD (Hewlett-
Packard, model 1100 series), respectively.
Table 2: The physical properties of fluorene, phenanthrene, fluoranthene,

pyrene, and anthracene.
Physical fluorene phenanthrene anthracene fluoranthene pyrene

Property
Molecular C13H10 C14H10 C14H10 C16H10 C16H10
Formula
Molecular 166.2 178.2 178.2 202.3 202.3
Weight
Solubility 1.84 1.09 0.06 0.23 0.13
(mg/L)
logKow 4.18 4.57 4.54 5.22 5.13

Figure 1 shows the sorption isotherms at various Triton X-100 CMC values for
fluorene, phenanthrene, and anthracene, respectively. Results of fluoranthene
and pyrene present the similar trend (not shown here). All of these isotherms can
be fitted by linear equations. It implies that the partition process is the main
factor for the PAHs distribution in the micelle-soil-water system. The driving
forces of partition process are believed as the hydrophobicity of PAHs,
surfactant, and soils. Because the partition coefficient is used to describe the
competition of PAHs in the liquid and soil phases, the magnitude of various
PAHs partition depends not only on their water solubilities but also on the
sorption ability of the soils and surfactants. As such, no significant relationship
presents between partition coefficient and solubility in all figures.
Figure 1 also shows that partition coefficients of PAHs decreases with
increasing the surfactant concentrations. This observation is in agreement with
the results reported by Sun and Boyd [8]. It is noticed that the difference
between isotherms of blank and 5 CMC is insignificant for these five PAHs
compounds. It can be attributed to the sorption of surfactant on soil surface;
therefore, the solubilization capacity of PAHs is not significant. In addition,
results present that large amount of surfactants were sorbed in soil phase, the
sorbed surfactants may increase the sorption capacity of the soil.
Figure 2 shows sorption isotherms at various rhamnolipid CMC values for
fluorene, phenanthrene, and anthracene, respectively. Results of fluoranthene
and pyrene with a similar trend are not shown. In the figure, the differences
among various surfactant concentrations are insignificant.
200
blank
5 cmc
Fluorene in Soil (µ g/g), Qe

10 cmc
15 cmc
150
20 cmc
30 cmc
100
50
0
0 5 10 15 20 25
Fluorene in Solution (mg/L), Ce
200
Phenanthrene in Soil (µ g/g), Qe
150
100
blank
5 cm c
10 cmc
50 15 cmc
20 cmc
30 cmc
0
0 1 2 3 4 5 6 7 8
Phenanthrene in Solution (mg/L), Ce

Anthracene in Soil (µ g/g), Qe
5 cmc
250 10 cmc
15 cmc
20 cmc
30 cmc
200
150
100
50
0
0 2 4 6 8 10
Anthracene in Solution (mg/L), Ce
Figure 1: The PAHs isotherms with various CMC values in Triton X-100
solution-soil system such as fluorine, phenanthrene, anthracene.
160
140
Fluorene in Soil (µ g/g), Qe

120 5 cmc
10 cmc
100 15 cmc
20 cmc
30 cmc
80
60
40
20
0
2 3 4 5 6 7 8 9 10
Fluorene in Solution (mg/L), Ce
200
e
5 cm c
Phenanthrene in Soil (µ g/g), Q
10 cm c
150 15 cm c
20 cm c
30 cm c
100
50
0
0.1 0.2 0.3 0.4 0 .5 0.6 0.7
P h enanthren e in Solu tion (m g/L), C e

Anthracene in Soil (µ g/g), Qe
200
10 cm c
150 15 cm c
20 cm c
30 cm c
100
50
0
0 0.2 0.4 0.6 0.8 1
Anthracene in Solution (mg/L), Ce
Figure 2: The PAHs isotherms with various CMC values in rhamnolipid

solution-soil system such as fluorine, phenanthrene, anthracene.
Most equilibrium concentrations of PAHs in these data are below 1 mg/L,

which are far lower than their solubilities (10 ~ 20 mg/L) under the same
rhamnolipid concentration in water. This implies that most biosurfactant
molecules are sorbed in soils and just few in the solution to aid dissolving PAHs
compounds from soil to aqueous phase. However, these isotherms still can
roughly be fitted by linear equations. Throughout whole concentration range
studied of biosurfactant concentration (0 ~ 30 CMC), the partition coefficient of
PAHs in rhamnolipid system is constant, around 400.0 L/Kg.
4 Conclusion
The partition coefficient of PAHs between soil and micelle aqueous phase can be
described by the linear equations regardless of synthetic or bio surfactants. In
the Triton X-100 system, the partition coefficient of PAHs decreases with the
surfactant concentrations. In addition, Triton X-100 is sorbed in soil; the sorbed
surfactants may influence the sorption capacity of the soil. Throughout whole
biosurfactant concentration (0 ~ 30 CMC), the partition coefficient of PAHs in
rhamnolipid system can be regarded as a constant, around 400.0 L/Kg. The
biosurfactant also exhibits strong attraction onto the soil.
References
[1] Karickhoff, S.W., Organic Pollutant Sorption in Aquatic System. Journal
of Hydraulic Engineering, 10(6), pp. 707-735, 1984.
[2] Mackay, D.M. & Cherry, J.A., Groundwater Contamination: Pump-and-
Treat Remediation. Environmental Science and Technology, 23(6), pp.
630-636, 1989.
[3] Rosen, M.J., (eds). Surfactants and Interfacial Phenomena, John Wiley &
Sons Inc.: New York, 1989.
[4] Fiechter A., Biosurfactants: Moving towards Industrial Application.
Trends Biotech, 10, pp. 208-217, 1992.
[5] Miller R.M., Biosurfactant-Facilitated Remediation of Metal-
Contaminated Soils. Environmental Health Perspective, 103(Suppl 1), pp.
59-62, 1995.
[6] Lang, S. & Wagner, F., (eds). Structure and Properties of Biosurfactants.
In: Biosurfactants and Biotechnology, New York, 1994.
[7] Shaoo D., Smith, J.A., Imbrigiotta, T.E., & Mclellan, H.M., Surfactant-
Enhanced Remediation of a Trichloroethene -Contaminated Aquifer. 2.
Transport of TCE. Environmental Science and Technology, 32(11), pp.
1686-1693, 1998.
[8] Sun, S. & Boyd, S.A., Sorption of Nonionic Organic Compounds in Soil-
Water Systems Containing Petroleum Sulfonate-Oil Surfactants.
Environmental Science and Technology, 27(9), pp. 1340-1346, 1993.
Carbon dioxide sequestration in coal:

implications for CO2 disposal and
CH4 displacement from coal seams
M. Mirzaeian & P. J. Hall
Department of Chemical and Process Engineering,
University of Strathclyde, Glasgow, UK
Abstract
The sequestration of CO2 within unmineable coal seams is one of the most
attractive options for reducing atmospheric CO2 levels. Thus there is currently
considerable interest in the interactions of coal with CO2 for its long-term
disposal. This paper reports the analysis of coal / CO2 interactions at pressures of
up to 30 bar. The results obtained from differential scanning calorimetry (DSC)
show that the interactions of CO2 with coal leads to strongly bound carbon
dioxide on coal. It was also found that the temperature of the second order phase
transition of coal decreases with increase in CO2 pressure significantly,
indicating that high pressure CO2 diffuses through coal matrix, causes significant
plasticization effects, and changes the macromolecular structure of the coal.
Desorption characteristics of CO2 from coal were studied by temperature
programmed desorption mass spectrometry (TPD-MS). It was found that CO2
binds more strongly to coal and demands more energy to desorb from coal at
higher pressures.
Keywords: CO2 sequestration, coal, irreversible adsorption, high pressure
interactions, macromolecular structure.
1 Introduction
Increased atmospheric CO2 concentrations due to fossil fuel combustion cause
entrapment of solar radiation in the atmosphere and induce a gradual warming of
the Earth’s surface (Greenhouse effect). This problem is now recognized as one
of the most important environmental issues facing society. Therefore there is
doi:10.2495/GEO060161
currently considerable interest in the permanent disposal of CO2 and one option
is to sequester it into uneconomic coalfields.
Coal is a chemically heterogeneous solid containing mainly carbonaceous
material with very lower amount of mineral mater. It is a microporous
macromolecular material containing a wide range of highly reactive chemical
functional groups [1]. Its porosity results in the entrance of fluids into its
structure and its polymeric nature accompanied with the presence of various
functional groups, leads to the chemical interactions of fluids with its matrix.
It is well known that coal swells if it is in contact with solvents, such as
pyridine, which break hydrogen bonds [2]. There is some evidence that coal can
swell in high pressure CO2 [3], presumably due to the quadrupolar nature of the
molecule disrupting weak electrostatic bonds within the coal structure.
Dilatometric studies on coals in contact with CO2 showed significant increase in
sample size [4]. It was suggested that CO2 swells coal because the solubility
parameter of CO2 is close to the solubility parameter of coal. In the most
significant study of coal behaviour in high pressure CO2 atmospheres Reucroft
and Sethuraman [3] have shown that coals swell after exposure to CO2 and the
amount of swelling increases with increasing pressure. It is supposed that the
increased swelling effect with increased pressure might be due to the solubility
parameter of CO2 approaching a value closer to that of the coals.
In the present study we have applied differential scanning calorimetry (DSC)
for the first time to the investigation of coal/ CO2 interactions and the effect of
high pressure CO2 on the coal structure. Temperature programmed desorption
mass spectrometry (TPD-MS) has been also applied to the study of the
desorption characteristics of CO2 from coal at pressures of up to 30 bar.
2 Experimental
Wyodak coal obtained from the Argonne sample bank was used in this study.
2.1 The DSC procedure
The DSC measurements were performed with a Mettler DSC 30 instrument in

conjunction with Mettler software TA72PS for data acquisition and processing.
Standard aluminium pans were used, with two pinholes in order to minimise
mass transfer limitations in evaporation of water or contact of gas with sample
during DSC scans. Nitrogen flowing at 10 ml/min was used as a carrier gas to
keep the cell free of oxygen during measurements. Typically 10 mg of sample
was used in an experiment. The DSC measurements were performed at a heating
rate 10ºC/min. Cooling of the furnace between consecutive heating scans was
carried out using a liquid nitrogen cooling accessory directly beneath the furnace.
2.1.1 Adsorption of N2 and CO2 on Wyodak coal
To study the irreversible adsorption of CO2 on the coal two different series of
scans were conducted on the Wyodak coal sample.
1- A sample of fresh coal was purged with N2 flowing at 10 ml/min for 10

min in DSC chamber then heated to 110ºC at 10ºC/min and held for 30 min,
cooled at nominal rate of 100ºC/min to -60ºC. The sample was then heated from
-60 to 200°C at 10ºC/min three times in succession with a cooling rate of
10ºC/min between heating runs. In all scans the sample always remained in the
DSC chamber under nitrogen flowing at 10 ml/min.
2- A sample of fresh coal was purged with N2 flowing at 10 ml/min for 10
min in DSC chamber. then heated to 110ºC at 10ºC/min and held for 30 min,
cooled at nominal rate of 100ºC/min to -60ºC under N2 flowing at 10ml/min. At
this point gas was switched to CO2 and the sample was then heated from -60 to
200ºC under CO2 flowing at 10 ml/min at 10ºC/min three times in succession
with a cooling rate of 10ºC/min between heating runs.
2.1.2 Effect of high pressure CO2 on the coal structure

Differential scanning calorimetry was also employed to study the effect of high
pressure CO2 on the macromolecular structure of Wyodak coal:
1- A sample of Wyodak coal was purged with N2 flowing at 10 ml/min for 10
min in DSC chamber. Then heated from 30ºC to 110ºC at a heating rate of
10ºC/min, held at 110ºC for 30 min and cooled to 30ºC in N2 atmosphere at a
flow rate of 10 ml/min. This was performed to remove water from coal sample
before determining its glass transition temperature. The dried sample was heated
from 30ºC to 200ºC at 10°C/min in N2 atmosphere flowing at 10 ml/min three
times in succession. This experiment was carried out to determine the phase
transitions of Wyodak coal in N2 atmosphere.
2- A sample of Wyodak coal was purged with 20 bar Ar three times in
succession, to flush adsorbed gases present in the pores in coal and also oxygen
form the high pressure cell. Then heated to 110ºC, held for 30 min at 110ºC, and
cooled to 30ºC at 10°C/min under Ar atmosphere in the high-pressure cell. The
sample was loaded with CO2 to the desired pressure at room temperature. It was
exposed to this CO2 environment for 24 h. After that, the CO2 pressure was
rapidly released, the sample was purged with 20 bar Ar and transferred to the
DSC chamber. The sample was purged with N2 with a flow rate of 10 ml/min for
10 min in the DSC chamber and DSC was carried out from 30ºC to 200ºC at
10°C/min in N2 atmosphere three times in succession. The purpose of this
sequence was to determine the effect of high-pressure CO2 atmospheres on the
phase transition of the coal sample.
2.2 The TPD-MS procedure
Approximately 150 mg of sample was placed in a sample tube and loaded with
CO2 to the desired pressure at room temperature in a high-pressure cell. The
sample was exposed to this high-pressure CO2 atmosphere for a certain period of
time. Then CO2 pressure was rapidly released and the sample was transferred to
the desorption chamber and purged with high purity helium for 10 min at 298 K,
before commencing the TPD run. The gas flow rate was 100 ml/min. To perform
a TPD –MS scan, the sample was heated by linearly increasing the temperature,
20 K/min, and the evolution of CO2 from the sample was monitored by a Hiden
Analytical HAL/HPR20 Quadrupole Mass Spectrometer (QMS).
3.1 Irreversible adsorption of CO2 on Wyodak coal
The results of the calorimetric measurements of the adsorption process for

nitrogen and carbon dioxide on Wyodak coal are shown in figures1-2.
Heat flow (W/g) Heat flow (W/g)

8 8
exo
6 endo 6
4 4
2 2
0 first scan 0
second and third scans

-2 -2
-4 -4
-50 0 50 100 150 200
o
Temperature ( C)
Figure 1: DSC scans for adsorption of N2 on Wyodak coal sample.
Heat Flow (W/g) Heat Flow (W/g)

8 8
exo
6 endo 6
4 4
2 2
0 0
second and third scans
-2 -2
first scan
-4 -4
-50 0 50 100 150 200

o
Temperature ( C)
Figure 2: DSC scans for adsorption of CO2 on Wyodak coal sample.
DSC results for the adsorption of N2 on coal sample in figure 1 show that
coal/N2 interactions are very weak and the adsorption of N2 on coal occurs
physically and reversibly.
Exothermic peaks for the adsorption of CO2 on coal in figure 2 are associated
with the uptake of CO2. This is an activated process and presumably at the
temperature of the exotherms there is enough thermal energy to overcome the
activation energy for diffusion. A comparison between figure1 and figure 2
shows that exotherms evident at low temperatures are absent when the
experiments were conducted under N2. This suggests that interactions between
coal and CO2 are much stronger than those between coal and N2.
The integrated values for the exotherms associated with the adsorption of CO2
on Wyodak coal are given in table 1. These values are indicative of the amount
of CO2 sorbed during the experiment. The reduction in the value of the exotherm
between the first and second runs suggests that some CO2 is irreversibly bound
to the structure even after heating to 200ºC.
Table 1: Differential enthalpies of the adsorption of CO2 on Wyodak coal.
Irreversible
First scan Second scan Third scan Sorption capacity
∆ H (J/g) 27.40 20.60 18.72 6.8
3.2 Structural change in coal caused by high pressure CO2
Figure 3 shows the DSC thermograms for dried Wyodak coal from 30ºC to
200°C. The first scan show an irreversible process, which might be attributed to
the structural rearrangement and relaxation in the coal when heated above its
glass transition temperature. Subsequent two scans after the first scan show a
reversible second order process. This process has the characteristics of glass
transition [5]. Before the transition coal is a glassy solid with severely restricted
macromolecular motions and diffusion of gases and liquids in its structure is
slow. When heated to a certain temperature which is called glass transition
temperature, a significant increase in coal’s macromolecular motions happens.
Above the transition coal becomes rubbery and diffusion into its structure
becomes much faster.
Figure 4 shows the DSC thermograms for dried Wyodak coal held under 30
bar CO2 atmosphere for 24 hours prior to DSC measurements. In this case, the
first scan illustrates two endothermic effects. The first one might be related to the
evaporation of moisture adsorbed by sample during transferring from the high
pressure cell to the DSC chamber and also continuous release of sorbed CO2
since desorption process is endothermic. The second effect might be attributed to
the fast release of sorbed CO2 from the coal sample at the vicinity of its glass
transition temperature. As the coal in the DSC chamber is heated it will
continuously release CO2. At the vicinity of glass transition temperature the
desorption rate may suddenly be accelerated since the chain mobility of the coal
suddenly increases. This process is irreversible and has disappeared on the
second and third scans.
1.5 1.5
exo
1 1
endo
0.5 0.5
0 0
-0.5 -0.5
-1 third scan -1
second scan
-1.5 -1.5
first scan
-2 -2
40 60 80 100 120 140 160 180 200
o
Temperature ( C )
Figure 3: DSC for dried Wyodak coal from 30ºC to 200ºC in N2 atmosphere.

3 0.5
exo
2
0
endo
1
-0.5
first scan
0
second scan -1
-1 Delta H = 13.2 J/g
-1.5
-2
third scan -2
-3
-4 -2.5
-5 -3
40 60 80 100 120 140 160 180 200
Temperature ( oC )
Figure 4: DSC for dried Wyodak coal (held under 30 bar CO2 atmosphere
for 24 hours) from 30ºC to 200ºC in N2 atmosphere.
Figure 5 shows the change in the glass transition temperature of the coal with
CO2 pressure. Depression in glass transition temperature of coal at CO2
atmosphere might be due to the solubility of CO2 into the coal matrix and
plasticization of coal by CO2.
3 3
exo
2 2
endo
1 1
30 bar
0 0
20 bar
-1 10 bar -1
5 bar
-2 -2
-3 -3
40 60 80 100 120 140 160 180 200
o
Temperature ( C )
Figure 5: The dependence of glass transition temperature of Wyodak on the

pressure of CO2.
Characteristics of the glass transition of Wyodak coal at various pressures of

CO2 are given in table 2. The further decrease in the glass transition temperature
with CO2 pressure might be due to the further solubility of gas into the coal
which is partly due to the solubility parameter of CO2 approaching to a value
closer to that of the coal at higher pressures.
It is supposed that the acidic and basic properties of CO2 which allow it to
form hydrogen bonds or other acid-base bonds play a major role in its solubility
in coal [3]. It dissolves in the organic coal matrix, thus modifying the physical
and possibly the chemical structure of the coal matrix by disrupting hydrogen
bonds in the coal structure during its interaction with coal [6].
Table 2: Glass transition temperature of Wyodak coal at different pressures

of CO2.
Second scan Third scan

P CO2 (bar) Tg (°C) Tg (°C)
0 121.1 121.6
5 86.6 81.7
10 83.2 78.2
20 82.4 77.8
30 81.1 75.7
3.3 Desorption characteristics of CO2 from Wyodak coal
Figure 6 shows the TPD-MS spectrum of CO2 desorption from Wyodak coal.
The coal sample was held at 5 bar CO2 atmosphere for 24 hours prior to the
TPD-MS measurements. The spectrum has two regions of interest. The first one
at low temperature region (in the range of 300-350 K) and second one after the
peak temperature where desorption rates decrease with increasing temperature.
-10
2 10 -10
2 10
Desorption rate (arb.un.)

1.5 10 -10
-10
1.5 10 1 10
-10
5 10 -11
E xp . Da ta
Redhead Eq u. Res idua l

-10
1 10 0
300 310 320 3 30 340 350
Tem pe rature (K)
-11
5 10 Exp. Data
Redhead Equ.
0
300 350 400 450
Temperature (K)
Figure 6: TPD-MS spectrum of CO2 adsorbed on Wyodak coal at 5 bar.
To analyze the desorption data, the low temperature part of the TPD-MS
spectrum has been modelled using the assumption of a first order desorption
process with a single activation energy for desorption (Redhead Equation) [7].
2
N  Edes  1 1  T   Edes  1 1 
ln =  −  +   exp −  −  − 1
 NP

 R  T TP   TP   R  T TP 
NP is the maximum desorption rate at peak temperature, TP, and N is the

desorption rate at any temperature.
The embedded graph in figure 6 shows that the first order model provides a
good fit to the experimental data in the range of 300-350 K. The value of the
activation energy for desorption, Edes, is calculated from the fitting of the low
temperature part of the desorption spectrum to the model. The value of Edes for
desorption of CO2 from Wyodak coal is estimated as Edes ≈ 15 kcal/mole. This
value is much smaller than the ~ 30 kcal/mol value for activation energy for
desorption of pyridine from Illinois #6 coal measured by Hall and Larsen [7].
Pyridine is an excellent hydrogen bound acceptor and a strong organic Lewis
base [8], it interacts with coal through strong noncovalent interactions and
hydrogen bonding with coal hydroxyls [9]. A comparison between these values
shows that the estimated value for activation energy of CO2 desorption from
Wyodak coal is sensible.
The trail of the spectrum after the peak shows that desorption rate does not
follow the first order desorption model. The process of CO2 desorption from coal
is a mass transfer process involving several elementary processes. These
processes may include diffusion through macropores, diffusion through
micropores and diffusion through coal matrix. In small micropores, diffusional
resistance limits mass transfer process. The mass exchange can also be limited
due to the large energy barrier for desorption. At higher surface coverages, the
diffusivity is high and the rate of desorption reaches a maximum. As temperature
increases and the desorption process proceeds, high energy sites in the
microporous structure are released. Due to the large energy barriers the rare of
desorption form these sites is very slow [10]. Therefore the deviation of
desorption spectrum from the fist order kinetic model might be due to the
contributions of activated diffusion effects, micropore diffusional resistance,
barrier resistance of high energetic sites and experimental error in the instrument.
TPD-MS spectra of CO2 adsorbed on Wyodak coal at various pressures are
compared in figure 7. The total area under a TPD-MS spectrum is proportional to
the amount of adsorbed CO2 [11]. The values of Edes and the integrated areas
under the spectra of CO2 for various pressures are shown in Table 3. It can be
noted that the desorption intensities increases with pressure indicating the
amount of CO2 sorbed in the coal is greater at higher pressures. Consequently at
higher pressures the relatively larger amount of CO2 would be desorbed.
1 1 0 - 10
No CO2
- 11
8 10 5 bar CO2
15 b ar C O 2
- 11 20 b ar C O 2
6 10
- 11
4 10
- 11
2 10
0
300 35 0 40 0 45 0
Te m p erature (K )
Figure 7: TPD-MS spectra of CO2 adsorbed on Wyodak at various pressures.
Table 3: Desorption parameters for CO2 adsorbed on Wyodak coal at

various pressures for 24 hours.
P (bar) E des (cal/mol) Integrated area

5 14521 1.5744 × 10 -9
15 15210 3.341 × 10 -9
20 15399 4.673 × 10 -9
4 Conclusions
It has been found that CO2 binds to the structure of Wyodak coal strongly and
irreversibly even after heating to 200°C. Glass transition of coal decreases with
CO2 pressure significantly, so there is a strong suggestion that high pressure CO2
diffuses through coal matrix, causes significant plasticization effects and changes
the macromolecular structure of coal. Desorption characteristics of coal loaded
with high pressure CO2 show that the amount of CO2 uptake increases with
increasing CO2 pressure. Increase in activation energy for CO2 desorption from
coal with pre-adsorbed CO2 pressure suggests that high pressure interactions will
demand more energy to desorb from coal probably due to the further access of
CO2 to the coal microporous structure.
The results of this study prove that Wyodak coal have a great affinity and high
irreversible sorption capacity for CO2 and could be excellent for CO2
sequestration.
References
[1] Green, T. L.; Kovac, J.; Brenner, D.; Larsen, J. W. In Coal Structure;
Meyers, R. A. , Ed., Academic Press, New York, pp. 199-282, 1982.
[2] Larsen, J. W.; Mohammadi, M. Energy & Fuels, 4, pp. 100-106, 1990.
[3] Reucroft, P. J; Sethuraman , A. R. Energy and Fuels, 1, pp.72-75, 1987.
[4] Reucroft, P. J.; Patel, H., Fuel, 65, pp. 816-820, 1986.
[5] Mackinnon, A. J.; Hall, P. J. Energy & Fuels, 9, pp. 25-32, 1995.
[6] Glass, A. S.; Larson, J. W. Energy & Fuels, 8, pp. 629-636, 1994.
[7] Hall, P.J.; Larsen, J.W. Energy & Fuels, 7, pp. 47-51, 1993.
[8] Arnett, E. M.; Joris, E.; Murty, T. S. S. R.; Gorrie, T. M.; Schleyer, P. v.
R. J. Am. Chem. Soc. 92, pp. 2365-2377, 1970.
[9] Larsen, J. W.; Baskar, A. J. Energy & Fuels, 1, pp. 230-232, 1987.
[10] Reid, C. R.; Thomas, K. M. Langmuir, 15, pp. 3206-3218, 1999.
[11] Habenschaden, E.; Kuppers, J. Surf. Sci, 138, pp. L147, 1984.
Section 4
Climatological processes
Response of dissolved organic carbon in a

shallow groundwater ecosystem to a simulated
global warming experiment
K. P. Wilson & D. D. Williams
Department of Life Sciences,
University of Toronto at Scarborough, Canada
Abstract
Dissolved organic carbon (DOC) in marine and freshwater ecosystems represents
an immense reservoir of organic matter with varied and significant ecological
value. Global warming poses a significant threat in that it has the capacity to
alter the concentration and distribution of DOC. Since groundwater constitutes
approximately two-thirds of the available freshwater on earth, it is crucial to
determine how global warming may affect its DOC balance. However, in higher
latitudes carbon cycling is poorly understood, and ecosystem-scale studies are
urgently required. We conducted an in situ temperature manipulation of a
shallow groundwater system in Ontario, Canada that simulated temperature
increases predicted by general circulation models for this region. Specifically,
treatment block temperatures in spring, summer, and fall were elevated 3.9 ± 0.6
SD °C, whereas winter temperatures were elevated 5.0 ± 0.6 °C compared with a
control block. We found no significant difference in DOC between control and
treatment blocks during the pre-manipulation study period. However, there was a
significant increase in DOC with groundwater depth in both blocks: 4.54 ±
0.25 mg/l at –20 cm to 5.79 ± 0.24 mg/l at –100 cm. During this period there was
also a difference in DOC among seasons: fall and winter concentrations were
lower than spring and summer. During the manipulation period there was also no
difference in DOC between the control and treatment blocks, however, a positive
trend in the treatment block was observed for all collections. Also, seasonal and
depth differences between blocks were still apparent. Although during the
manipulation period nitrate and total phosphorus showed no difference between
control and treatment blocks, ammonia showed a significant decrease in the
treatment block. We discuss the implications of these findings to the
biogeochemistry and ecology of shallow aquifers.
Keywords: global warming, dissolved organic carbon, DOC, shallow
groundwater, Canada, temperature manipulation, ecosystem experiment.
doi:10.2495/GEO060171
1 Introduction
The importance of dissolved organic carbon (DOC) as a potential source of
carbon and energy for subsurface and surface metabolism has been widely
studied. Fisher and Likens [1] first examined how subsurface waters can
effectively transport DOC from terrestrial ecosystems to stream ecosystems.
Hynes [2] showed that subsurface DOC contributes significantly to stream
ecosystem metabolism. These influences of DOC are due to its ability to affect a
variety of physical, chemical, and biological processes. DOC entering shallow
groundwater can be from terrestrial environments via soils (Chappelle [3]),
through streams via the hyporheic zone (Williams and Hynes [4]), or laterally
from riparian zones (Meyer and Tate [5], Storey et al. [6]). Once DOC enters
shallow groundwater ecosystems, it can be oxidized to CO2 (Chappelle [3]), or if
consisting of labile substances it can be rapidly utilized by microorganisms
(Wetzel [7], Storey et al. [6]). DOC interacts with dissolved nutrients influencing
nutrient concentrations and can also act as a buffer by affecting pH (Pace and
Cole [8]). There is growing concern of the fate of DOC in higher latitudes
because these temperate regions are expected to respond to global warming more
so than lower latitudes (Schindler [9]).
General circulation models (GCM) predict increases in temperature to vary
between 1.5 and 5.0 °C by the year 2040 for southern Ontario, Canada
(Hengeveld [10]). These models also project a differentially higher temperature
increase in winter than in summer in latitudes >30° (Hengeveld [10]).
Temperature is one of the most important factors that affect life history
characteristics and distributions of bacteria (Chapelle [3]) and aquatic insects
(Sweeney [11], Vannote and Sweeney [12]). General circulation models have
also shown that increased CO2 will change precipitation and temperature
patterns, and therefore likely impact the abundance and distribution of species
(McCarty [13]). Changes in the community structures of shallow groundwater
ecosystems may affect DOC concentrations through changes in metabolic
activity and hence aspects of the global carbon-cycle.
Global warming studies examining DOC concentrations have been conducted
mainly in terrestrial ecosystems, such as soils (MacDonald et al. [14]), bogs and
fens (Pastor et al. [15]), and arctic tundra and sedge ecosystems (Neff and
Hooper [16]). Some aquatic ecosystem studies have also examined the response
of DOC to global warming and climate change. For example, Schindler et al.
[17] examined the physicochemical properties of boreal lakes, with emphasis on
DOC and Clair et al. [18] modelled DOC loss from a small temperate wetland
under a doubling of CO2. To our knowledge there has not been any direct whole-
ecosystem manipulation of a shallow groundwater ecosystem that has examined
DOC concentrations both spatially and temporally. To address the possible
effects of global warming on DOC concentrations in a shallow groundwater
ecosystem, we conducted an in situ temperature manipulation of a shallow
groundwater ecosystem in accordance with global warming projections for
southern Ontario, Canada. Specifically, we examined DOC concentrations: 1)
spatially along a natural vertical gradient from surface water to –100 cm
sediment depths; 2) temporally by comparing seasons; and 3) by perturbing the

groundwater temperature over a period of 13 months. We predict two possible
responses of DOC concentrations to our ecosystem manipulation: 1) DOC
concentrations will not change between control and treatment blocks (ambient
DOC concentrations maintained by soil C storage); or 2) DOC concentrations
will decrease in the treatment block compared with the control block
(consumption of DOC through increased metabolism).
2 Methods and materials
2.1 Study site
This study was conducted on a small first order spring-brook (Valley Spring)
located in southern Ontario, Canada (43°45’ N, 79º15’ W). Valley Spring is
located at an elevation of 152 m and is approximately 60 m in length and 0.5-
1.5 m wide with a discharge ranging between 1800-2300 l/hr. The surrounding
vegetation consists of coniferous and deciduous trees and mixed grasses and
shrubs. For further descriptions see Williams and Hogg [19] and Hogg and
Williams [20].
2.2 Temperature manipulation
The temperature manipulation study area was located approximately 5 m from

the spring source in an 8 m x 4 m grid. The study area was separated into a
control and experimental treatment block, each 4 m x 4 m. Heating pipes were
installed on 8 October 2003. Eight pipes (110 cm long, 4 cm outer diameter)
pinched at the bottom (last 5 cm), were placed at 0.5 m intervals perpendicular to
the stream channel along four transects at 1 m intervals longitudinally with the
stream channel (n = 32 for each of the control and treatment blocks, fig.1). A
treatment divide (5 m in length and 1 m deep) was installed along a natural
groundwater divide in October 2003 using sixteen gauge galvanized steel sheet
metal.
To increase the temperature in the treatment block, a 61 m long
EASYHEAT® roof and gutter de-icing cable (Model ADKS, 120 V, 1000 W)
was installed along each transect (n = 4). Starting at the end of each cable, 6 m
of cable was wrapped length wise and tied together with cable-ties and placed
into each pipe with the lead end running to the next pipe (in series). In order to
control seasonal temperatures, each cable was attached to a variable transformer
Powerestat® model # 3PN117C, 120 V, 12 A. Heat was applied on 5 March
2004 and continued until 5 May 2005. From general circulation models our
targeted temperature differences between the control and treatment blocks for
summer, spring, and fall range from 3.5-4.5 °C and in winter from 5-6 °C.
2.3 Collecting regime
Bimonthly samples of DOC concentrations, nutrients, and water chemistry were

collected in June, August, October, December 2002, and February, April, June,
August, October, and December 2003 and 2004, and February, and April, 2005
(n = 18 collection dates). Collections between June 2002 and February 2004
(n = 11) were used to generate pre-manipulation data to determine if there were
differences between the control and treatment blocks. Collections between April
2004 and April 2005 (n = 7) were used as manipulation response samples to
examine possible temperature effects on DOC concentrations.
Control Treatment
o o o o o o o o o o o o o o o o
o o o o o o o o
o o o o o o o o
Groundwater Divide
Figure 1: Schematic of experimental design showing the control block (no heat)
and treatment block (heating). Bold circles (o) represent heating
pipes, regular circles (o) represent pipes with no heating. Dashed line
is location of groundwater divide. Large circles with an (X) are the
location of nested piezometers. Arrows are direction of surface water
flow.
2.4 DOC sampling protocol
For each sampling date, duplicate samples of DOC concentrations were collected
from five depths, –20, –40, –60, –90, and –100 cm and from the surface. Water
samples were collected from nested mini-piezometers (Freeze and Cherry [21])
located within the control and treatment blocks. DOC determinations were
performed on a Tekmar Dohrmann™ Phoenix 8000 UV-persulfate oxidation
TOC analyzer.
2.5 Statistical analyses
All data were analyzed using JMP-start statistics software, (SAS Institute Inc.).
A Shapiro-Wilks W-test of normality was used to test for homogeneity of
variances (Zar [22]). DOC and temperature data were analyzed using two-way
analysis of variance (ANOVA). The response variables (DOC and temperature)
were compared with the predictor variables of depth, season, control and
treatment blocks, and pre-manipulation and manipulation collections. To
determine differences among depths and seasons Tukey-Kramer HSD
comparison tests were conducted. Winter seasons included December and
February collections, spring consisted of April collections, summer included

June and August collections, with fall including October collections. Stepwise
regression analyses were conducted on the response variable (DOC) to 9
predictor variables (total phosphorus O-PO43-, mg/l), nitrate (NO3—N, mg/l),
ammonia (NH3-N, mg/l), sulfide (S2-, mg/l), dissolved oxygen (DO, mg/l),
conductivity (µS), pH, total dissolved solids (TDS), and temperature (°C).
3 Results
3.1 Temperature data: pre-manipulation
Temperature showed no statistical difference for all depths between control and
treatment sites during the pre-manipulation. Fall and winter temperatures were
most uniform among depths compared with spring and summer. Fall and winter
temperatures increased with depth with fall values ranging from 12.6 ± 0.3 SD to
13.2 ± 0.3 °C and winter values ranging from 7.4 ± 0.3 to 8.0 ± 0.3 °C at –20 and
–100 cm, respectively. In contrast, spring and summer temperatures decreased
with depth. Spring temperatures ranged from 8.2 ± 0.3 at –20 cm to 6.3 ± 0.3 °C
at –100 cm and summer temperatures at –20 cm ranged from 14.3 ± 0.3 to 12.9 ±
0.3 °C at –100 cm.
3.2 Temperature data: manipulation
Temperature showed statistical differences between control and treatment blocks

during the experimental manipulation. Temperatures between control and
treatment blocks were significantly different for each depth in each season
(p < 0.0001, n = 468, F = 2314.5). In winter, mean temperature for the control at
–20 cm was 7.9 ± 0.1 °C, compared with 11.9 ± 0.1 °C for the treatment, an
average difference of 4 °C, fig. 2. In winter, temperature differences between
control and treatment plots were slightly higher with depth. At –40 cm the
control was 9.5 ± 0.1 °C and the treatment was 14.6 ± 0.1 °C, a difference of
5.1 °C. Differences between control and treatment at –60, –80 cm, and –100 cm
were 5.5, 5.5, and 4.9 °C, respectively, fig 2.
Spring, summer, and fall temperature differences between control and
treatment blocks also varied with depth. Differences at –20 cm for spring,
summer, and fall were 3.2, 3.3, and 3.3 °C, respectively. At –40 cm, differences
were 4.1, 4.1, and 4.0 for spring, summer, and fall, respectively. At –60, –80, and
–100 cm, differences were also close among seasons; –60 cm differences were
4.3, 4.3, and 4.0 °C, at –80 all were 4.2, and at –100 cm differences were 3.7,
3.5, and 3.7 °C. In fall, highest temperatures for the treatment block were 18.6
and 18.7 °C at –80 and –100 cm, respectively.
3.3 DOC concentration patterns: pre-manipulation
There was no significant difference in DOC concentrations between control and

treatment blocks for surface and subsurface pre-manipulation collections.
However, seasonal patterns were detected for most collection sites. Pre-
manipulation concentrations of DOC at the surface were significantly higher in
spring (4.17 ± 0.63 SD mg/l) compared with summer (1.21± 0.27 mg/l), fall
(1.35 ± 0.40 mg/l), and winter (1.51 ± 0.30 mg/l; p = 0.002, n = 3, F = 6.43; fig.
3A).
20
18 -20 cm
16 Treatment
14
12
10
8
6
Temperature ( 0C)
Control
4
Heat turned
2
off
0
20
-100 cm Treatment
18
16
14
12
10
8
6
Control
4 Heat turned
2 off
0
Spring 2004 Summer 2004 Fall 2004 Winter 2005
Figure 2: Temperature (°C) at –20 and –100 cm sediment depths, for control
and treatment blocks during the heat-manipulation phase. Heat was
turned off on 5 May 2005.
Furthermore, seasonal patterns during the pre-manipulation collections were

not consistent with depth. At –20 cm, winter DOC was significantly lower (3.48
mg/l ± 0.31 SE) than spring (5.28 ± 0.68), summer (4.75 ± 0.30 mg/l), and fall
(4.60 ± 0.37 mg/l) concentrations (p = 0.016, n = 3, F = 3.91; fig. 3A). At –40
cm, spring and summer DOC concentrations were significantly higher than both
fall and winter (p < 0.0001, n = 3, F = 10.60). Spring and summer DOC were
5.18 ± 0.58 mg/l and 5.24 ± 0.29 mg/l, respectively, compared with 3.77 ± 0.37
mg/l for fall and 3.06 ± 0.30 mg/l for winter. In contrast, seasonal trends were
not significant for the –60 and –80 cm depths. Mean DOC concentrations at –60
cm ranged from a low of 4.22 ± 0.57 mg/l in winter to a high of 6.46 ± 0.91 mg/l
in spring, and at –80 cm, DOC concentrations ranged from a low of 4.36 ± 0.34
mg/l in winter to a high of 5.52 ± 0.30 mg/l in spring. However, at –100 cm
seasonal trends were similar to –20 and –40 cm. Spring, summer, and fall DOC
concentrations were all higher and significantly different from winter
concentrations (p = 0.0003, n = 3, F = 7.73).
Pre-manipulation (A)
8
-100 cm
6 -60 cm
-80 cm
4
Dissolved Organic Carbon (mg/l)
-40 cm
-20 cm
2
Surface
0
8
Manipulation (B)
6
0
Winter Spring Summer Fall
Season
Figure 3: Seasonal concentrations of dissolved organic carbon (DOC, mg/l)
for (A) pre-manipulation and (B) manipulation collections.
3.4 DOC concentration patterns: manipulation
Increased experimental temperatures did not have a significant statistical effect

overall between control and treatment blocks. However, manipulation collections
showed similar patterns to pre-manipulation collections. Similar to pre-
manipulation collections, surface DOC concentrations were highest in the spring
(3.56 ± 0.08 mg/l) and significantly different from winter (2.95 ± 0.08 mg/l),
summer (1.03 ± 0.10 mg/l), and fall (1.00 ± 0.11 mg/l; p < 0.0001, n = 3,
F = 213.84; fig. 3B). However, different patterns were detected at several
subsurface depths. At –20 cm, spring showed the highest DOC levels (6.09 ±
0.28 mg/l) which were significantly different from winter (4.25 ± 0.28 mg/l),
summer (4.18 ± 0.28 mg/l), and fall (3.05 ± 0.40 mg/l; p < 0.0001, n = 3,
F = 15.27). In contrast with pre-manipulation results, manipulation patterns
from –40 to –100 cm showed statistical differences. At –40 cm, DOC
concentrations for spring, summer, and winter seasons were highest and
significantly different from the fall (p = 0.27, n = 3, F = 3.64). Seasonal patterns
during the manipulation were also detected at –60 and –80 cm unlike pre-
manipulation collections. At –60 cm, fall was significantly different from the
other seasons (p = 0.010, n = 3, F = 4.69; Table 1). In contrast, at –80 and –100
cm, summer DOC concentrations were significantly higher (6.21 ± 0.35 mg/l,
and 6.61 ± 0.35 mg/l, respectively) than fall (3.22 ± 0.5 mg/l, 4.41 ± 0.51 mg/l,
respectively; p = 0.005, n = 3, F = 8.44), whereas winter and spring

concentrations were similar at these depths.
Although, there were no statistically significant differences between control
and treatment blocks detected during the manipulation, a positive trend was
detected. At –20, –40, and –60 cm, DOC concentrations were slightly higher in
the treatment block compared with the control block. Differences between
control and treatment block concentrations were 4.3 to 4.7 mg/l at –20 cm, 3.7 to
4.2 mg/l at –40 cm, and 4.5 to 5.1 mg/l at –60 cm. Positive increases in DOC
concentrations were 0.4 mg/l, 0.5 mg/l, and 0.6 mg/l for –20, -40, and –60 cm,
respectively, for all manipulation collections combined. However, this trend was
not detected at –80 and –100 cm.
3.5 Stepwise regression: pre-manipulation and manipulation
The response variable DOC was analyzed with several predictor variables
(nutrients and water chemistry, n = 9) to examine which predictor(s) best explain
DOC concentrations. Examining pre-manipulation data for all collections and
depths combined, the highest r2 was 0.36 for all predictors (n = 9). After
examining all possible models, NH3-N mg/l, total-phosphorus (TP, mg/l), and
temperature (°C) were the best predictors, r2 = 0.33. The same analysis using
manipulation data (all collections and depths) had an r2= 0.31 for all predictors.
However, the best predictors for the manipulation period were NO3 mg/l, NH4
mg/l, and total dissolved solids (TDS), with an r2 = 0.27, n = 3.
Interestingly, predictor variables explain more variation when examining
individual depths. For example, at –100 cm for the manipulation period, all
predictors produce r2 = 0.73, n = 9. With the best model, including NO3-N mg/l,
NH3-N mg/l, pH, and TDS, r2 = 0.71, n = 4.
4 Discussion
Surface concentrations of DOC were temporally variable during the pre-
manipulation and manipulation study periods. Highest seasonal DOC
concentrations at the surface occurred in spring compared with other seasons, fig
3A, 3B. A similar pattern was detected in an alpine catchment in Colorado,
U.S.A. Boyer et al. [23] determined that during spring snowmelt, stream DOC
concentrations increased and were highest (approximately 4.5 mg/l) during this
time compared with the rest of the year. The primary source of increasing DOC
concentrations in spring was attributed to subsurface flow of water through soils
in the Deer Creek catchment (Boyer et al. [23]). In contrast, Bernal et al. [24],
determined that surface DOC concentrations were highest in a transition period,
dry to wet (September to November, our fall season), and lowest in a wet and
dormant period (December to February, our winter) and a vegetated period
(March to May, our spring). In the Bernal et al. study [24], seasonal differences
may be attributed to the study having been conducted in an intermittent
Mediterranean stream with autochthonous and allochthonous carbon sources
being available at different times of the year.
Shallow groundwater (–20 to –100 cm) concentrations of DOC below the

Valley Spring streambed were more seasonally variable than surface
concentrations. Seasonal pre-manipulation and manipulation concentrations of
DOC in the subsurface were always higher (3-7 mg/l) than the surface (< 4.2
mg/l), fig. 3. Highest subsurface DOC concentrations were also during spring
and summer, fig 3A, 3B. Rutherford and Hynes [25] sampled DOC from three
agriculturally impacted Ontario streams from the surface to a depth of –140 cm.
Variations in DOC concentrations were highly variable with depth and over
time. Concentrations were typically higher at the surface and at –20 cm, but
some stations had higher concentrations at the deepest depths (–140 cm) than
intermediate depths (–40 to –60 cm; Rutherford and Hynes [25]). Kaplan and
Newbold [26] determined from the literature that concentrations of DOC in
shallow groundwater can either decrease or increase with depth. These patterns
are suggested to be due to extensive abiotic and biotic processing of terrestrial
DOC sources in the vadose zone (zone of water limited above by the land surface
and below by the water table) creating low phreatic zone DOC levels, whereas
high phreatic zone DOC implies the opposite, low processing rates (Kaplan and
Newbold [26]).
Temperature differences between the control and treatment blocks during the
temperature manipulation were successfully maintained. The spring through fall
difference, combining all depths, was 3.9 ± 0.5 SD °C, and combining all depths
in winter the difference was, 5.0 ± 0.6 °C, fig. 1. Although targeted temperatures
were reached, no significant differences in DOC concentrations were found
between control and treatment blocks. Thus, our first of two predictions that no
change in DOC concentrations would occur was supported. There could be a
couple of reasons for this. Firstly, DOC concentrations in aquatic environments,
especially in hyporheic and shallow groundwater ecosystems can be very
heterogeneous. For example, Pabich et al. [27] examined DOC concentrations
below the water table of a shallow estuary on Cape Cod, U.S.A. DOC
concentrations of shallow groundwater were spatially variable, but temporally
stable suggesting that local heterogeneity plays an important role in DOC
delivery to shallow groundwaters (Pabich et al. [27]). Rutherford and Hynes [25]
also suggest that the heterogeneity of DOC concentrations may be due to the
complex flow patterns of hyporheic zones, and the typically mixed nature of bed
sediments. Our data also suggest that subsurface DOC concentrations are
heterogeneous due to the spatial and temporal variability in our system, thus
possibly masking temperature-induced differences, fig. 3B. Secondly, the
molecular structures that make up DOC are varied and highly complex. DOC is
usually categorized into two main groups, non-humic or labile substances and
humic substances (Wetzel [7]). Our study site is located in a temperate forest in
southern Ontario, Canada. Such headwater streams are typically characterized by
plant matter that accumulates after leaf-drop in the fall. Humic substances form
most (70-80%) of the organic matter in soils and water, are structures with high-
molecular-weight (HMW) and are typically the result of microbial activity on
plant material (Wetzel [7]). Humic substances released from leaves may decrease
turnover rates of DOC and provide a continuous supply of DOC to the shallow
groundwater ecosystem of Valley Spring. Further, since humic substances are

recalcitrant to biological activity and tend to have low turnover rates, the effects
of increased metabolic activity on DOC concentrations under the increased
thermal regime induced in Valley Spring may not have been detectable over the
short time period of 13 mo. In contrast, labile DOC typically consists of
carbohydrates, proteins, amino acids, and other low molecular weight (LMW)
substances. These LMW substances are readily used by microorganisms,
creating conditions for rapid flux in aquatic ecosystems.
Manipulation concentrations of DOC were correlated most with NO3-N mg/l,
NH3-N mg/l, and total dissolved solids (TDS), r2 = 0.27, p < 0.0001, n = 3. DOC
was positively correlated with ammonia, both concentrations increased with
depth, and DOC was negatively correlated with nitrate; DOC concentrations
increased with depth while nitrate concentrations decreased. This shows a tight
coupling between DOC and nitrate and ammonia. For example, if there are small
changes in DOC concentrations there is the possibility on non-linear effects on
watershed N retention (Goodale et al. [28]). These latter authors examined the
spatial patterns of nitrate and DOC concentrations of 100 northeastern U.S.
streams, and showed that as DOC concentrations increase, nitrate concentrations
decrease.
Our study did not see a change in DOC concentrations under a simulated
global warming experiment, although a trend of slightly higher DOC
concentrations was detected in the treatment block. We advocate that further
large-scale ecosystem manipulations should be conducted to more fully
understand the role and transformations of DOC in streams and shallow
groundwaters. In particular, these studies should examine the molecular fractions
or species of DOC, in order to detect changes in allochthonous and
autochthonous sources (e.g. Sachse et al. [29], Sobczak and Findlay [30]). For
example, it has been shown that allochthonous DOC is the most important
determinant of thermocline depth in small boreal lakes (Perez-
Fuentetaja et al. [31]). Further, if the type of DOC entering such aquatic systems
is not known (allochthonous verse autochthonous) correct management policy
decisions may be very difficult to make.
References
[1] Fisher S.G. & Likens G.E., Energy Flow in Bear Brook, New Hampshire:
an integrative approach to stream metabolism. Ecological Monographs
43(4), pp. 421-439, 1973.
[2] Hynes, H.B.N., The stream and its valley. Internationale Vereinigung fur
Theoretische und Angewandt Limnologie, 9(1), pp. 1-15, 1983.
[3] Chapelle, F.H., Ground-water microbiology and geochemistry, John
Wiley & Sons, Inc.: New York, pp. 263, 2001.
[4] Williams D. D. & Hynes H.B.N., The occurrence of benthos deep in the
substratum of a stream. Freshwater Biology, 4(1), pp. 233-256, 1974.
[5] Meyer J.L. & Tate C.M., The effects of watershed disturbance on
dissolved organic carbon dynamics of a stream. Ecology, 64(1), pp. 33-
44, 1983.
[6] Storey, R.G., Howard, K.W.F. & Williams, D.D., Factors controlling
riffle-scale hyporheic exchange flows and their seasonal changes in a
gaining stream: A three-dimensional groundwater flow model. Water
Resources Research, 39(2), pp. 2003.
[7] Wetzel, R.G., Limnology: Lake and River Ecosystems, Academic Press:
San Diego, pp. 735, 2001.
[8] Pace, M.L. & Cole, J.J., Synchronous variation of dissolved organic
carbon and color in lakes. Limnology and Oceanography, 47(2), pp.
333-342, 2002.
[9] Schindler, D.W., The cumulative effects of climate warming and other
human stresses on Canadian freshwaters in the new millennium.
Canadian Journal of Fisheries and Aquatic Sciences, 58(1), pp. 18-29,
2001.
[10] Hengeveld, H.G., Projections for Canada’s climate future. Environment
Canada Document CCD 00-01, pp. 1-27, 2000.
[11] Sweeney, B.W., Bioenergetic and developmental response of a mayfly to
thermal variation. Limnology and Oceanography, 23(2), pp. 461-477,
1978.
[12] Vannote R.L & Sweeney B.W., Geographic analysis of thermal equilibria:
a conceptual model fro evaluating the effects of natural and modified
thermal regimes on aquatic insect communities. The American Naturalist,
115(5), pp. 667-695, 1980.
[13] McCarty J.P., Ecological consequences of recent climate change.
Conservation Biology, 15(2), pp. 320-331, 2001.
[14] MacDonald N.W., Randlett, D.L. & Zak, D.R., Soil warming and carbon
loss from a lake states spodosol. Soil Science Society of America Journal,
63(1), pp. 218-221, 1999.
[15] Pastor, J., Solin, J., Bridgham, S.D., Updergraff, K., Harth, C.,
Weishampel, P. & Dewey, B., Global warming and the export of
dissolved organic carbon from boreal peatlands. Oikos, 100(2), pp. 380-
386, 2003.
[16] Neff, J.C. & Hooper, D.U., Vegetation and climate controls on potential
CO 2, DOC and DON production in northern latitude soils. Global
Change Biology, 8(9), pp. 872-884, 2002.
[17] Schindler, D.W., Bayley, S.E., Parker, B.R., Beaty, K.G., Cruikshank,
D.R., Fee, E.J., Schindler, E.U. & Stainton, M.P., The effects of climatic
warming on the properties of boreal lakes and streams at the Experimental
Lakes Area, northwestern Ontario. Limnology and Oceanography, 41(5),
pp. 1004-1017, 1996.
[18] Clair, T.A., Arp, P., Moore, T.R., Dalva, M. & Meng, F.R., Gaseous
carbon dioxide and methane, as well as dissolved organic carbon losses
from a small temperate wetland under a changing climate. Environmental
Pollution, 116(1), pp. S143-S148, 2002.
[19] Williams D.D. & Hogg I.D., The ecology and production of invertebrates
in a Canadian coldwater spring. Holarctic Ecology, 11(1), pp. 41-54,
1988.
[20] Hogg, I.D. & Williams, D.D., Response of stream invertebrates to a
global-warming thermal regime: An ecosystem-level manipulation.
Ecology, 77(2), pp. 395-407, 1996.
[21] Freeze, R.A. & Cherry J.A., Groundwater, Prentice-Hall, Inc.: New
Jersey, pp. 23-24, 1979.
[22] Zar J.H., Biostatistical Analysis, Prentice Hall: New Jersey, pp. 88-89,
1999.
[23] Boyer, E.W., Hornberger, G.M., Bencala, K.E. & McKnight, D.M.,
Response characteristics of DOC flushing in an alpine catchment.
Hydrological Processes, 11(12), pp. 1635-1647, 1997.
[24] Bernal, S., Butturini, A. & Sabater, F., Seasonal variations of dissolved
nitrogen and DOC: DON ratios in an intermittent Mediterranean stream.
Biogeochemistry, 75(2), pp. 351-372, 2005.
[25] Rutherford, J.E. & Hynes, H.B.N., Dissolved organic carbon in streams
and groundwater. Hydrobiologia, 154(1), pp. 33-48, 1987.
[26] Kaplan, L.A. & Newbold, J.D., Surface and subsurface dissolved organic
carbon (Chapter 10). Streams and Ground Waters, ed. J.B. Jones &
Mulholland P.J., Academic Press: San Diego, pp. 237-253, 2000.
[27] Pabich, W.J., Valeila, I. & Hemond, H.F., Relationship between DOC
concentration and vadose zone thickness and depth below water table in
groundwater of Cape Cod, U.S.A. Biogeochemistry, 55(3), pp. 247-268,
2001.
[28] Goodale, C.L., Aber, J.D., Vitousek, P.M & McDowell, W.H., Long-term
decreases in stream nitrate: Successional causes unlikely; Possible Links
to DOC? Ecosystems, 8(3), pp. 334-337, 2005.
[29] Sachse, A., Henrion, R., Gelbrecht, J. & Steinberg C.E.W., Classification
of dissolved organic carbon (DOC) in river systems: Influence of
catchment characteristics and autochthonous processes. Organic
Geochemistry, 36(6), pp. 923-935, 2005.
[30] Sobczak, W.V. & Findlay, S., Variation in bioavailability of dissolved
organic carbon among stream hyporheic flowpaths. Ecology, 83(11), pp.
3194-3209, 2002.
[31] Perez-Fuentetaja A., McQueen D.J. & Ramcharan C.W., Predator-induced
bottom-up effects in oligotrophic systems. Hydrobiologia, 317(2), 163-
176, 1999.
Regional analysis of climate and bioclimate

change in South Italy
A. Capra, P. Porto & B. Scicolone
Department of Agro-forestry and Environmental Sciences and
Technologies, Mediterranean University of Reggio Calabria, Italy
Abstract
In this paper, monthly values of rainfall (P) and temperature (T) recorded in
Southern Italy (Calabria and Sicily) during the period 1921-2000 are
investigated. In particular, a series of 211 raingauge and 53 temperature stations
are analysed for evidence of trend by using the linear regression and the Kendall
non-parametric test. The tests are applied at a seasonal and annual scale; a spatial
analysis is also carried out at both a regional and sub-regional scale in order to
check the effect of different sub-areas on these trends. An additional
investigation useful for checking the climate change effects on vegetation is also
included analysing bioclimatic parameters such as evapotranspiration and aridity
index. The results obtained confirmed, for the two investigated regions, the
importance of the climatic analysis carried out at a regional scale. In fact, the
tests showed for both the P and T records a strong difference between the two
investigated regions. In particular, the total annual P showed a decreasing and
increasing trend with -344 and 197 mm/100 years respectively in the Tyrrhenian
and Ionian sub-regions of Calabria and a decreasing trend with -179 mm/100
years in Sicily. A decreasing trend is evident in the Tyrrhenian sub-region for the
mean annual values of Tmax (-3.2 °C/100 years) and Tmin (-2.9 °C /100 years); a
different trend is shown in the Ionian sub-region where the mean annual values
of T showed an increase of 2.2 and 0.9 °C /100 years for Tmax and Tmin
respectively. In contrast, a clear increasing trend (1.8 and 2.2 °C /100 years) is
shown for both Tmax and Tmin in Sicily. A detailed analysis involving the
calculation of 10-year moving averages, showed a significant change of trending
after the 1950s for P and after 1970s for T.
Keywords: climate change, precipitation, temperatures, bioclimatic parameters.
doi:10.2495/GEO060181
1 Introduction
It has been observed that during the last 100-150 years the Italian climate has
changed and it has resulted in a rise of temperature and aridity. According to
Brunetti et al. [1], during the period 1865-2000, the mean annual temperature
(Tya) indicates a 0.4°C/100 years rise within the northern areas (N) of the country
(continental zone) and a 0.7°C/100 years rise in the central (C) and southern (S)
Italy (peninsular zones). More particularly, at seasonal scale, the slopes of the
regression line are greater during the winter season, ranging from 0.7°C/100
years (N) to 0.9°C/100 years (S), while for the summer season they are lower
and in some cases not significant. A negative trend of the annual rainfall (Py) is
evident within both N and S areas with a slope equal to -47 mm/100 years and -
104 mm/100 years respectively which mean, for the investigated period, 7% of
the mean rainfall for the northern area and 18% for the South. This decreasing
trend occurred particularly after 1950 since then the number of wet days (NP) has
also decreased. The maximum annual temperature (Tymax) series show a slope of
the regression line ranging between 0.4°C/100 years for N and 0.6°C/100 years
for S while the minimum annual temperature (Tymin) series show slope values
between 0.3°C/100 years for N and 0.5°C/100 years for S. According to Brunetti
et al. [1], the most important contribution to this positive trend is due to roughly
the last 20 years (before 1996) for N and to roughly the last 50 years for S. In
respect to extreme events, it has resulted an increase of rainfall intensity within
both N and S areas [1], [2] and a tendency toward an increase in drought [3], [4],
[5], [6].
Other investigations carried out in Italy [4], [7] showed a different behaviour
of the same climatic series if a subdivision of the N and S areas into smaller sub-
regions is taken into account.
For this reason, it seems necessary to investigate the climate change on a
regional scale where different geographical factors (e.g. distance from the sea,
elevation, aspect) are likely to influence the magnitude of these trends. Such
investigation can be useful not only in terms of improving scientific knowledge
but also to provide the Italian government with useful information in order to
make the right choices in planning future activities.
Calabria and Sicily (Fig. 1) are two regions located in South Italy and are
particularly prone to be investigated at local scale because of their geographic
characteristics and local orographic features. In particular, Calabria (Ca) is a
narrow (with a width ranging from 30 to 95 km) and long peninsula extending
from North to South for about 250 km; a mountain range (Apennine) runs in
latitude and divides the region into two opposite areas: the Tyrrhenian and the
Ionian zones. This mountain range is located almost perpendicularly to the
direction of the dominant moisture-bearing winds and for this reason it causes a
very strong variability in terms of rainfall and temperature patterns considering
also the different altitudes and aspects [8]; that is why it is easy to find flat and
semi-arid coastal areas (Tya and Pya equal to 17.4°C and 683 mm respectively) as
well as mountain zones with Tya equal to 9.1°C and Py equal to 1242 mm (the
highest in the South). Locally, the annual rainfall Py ranges from 1107 mm over
the Tyrrhenian area (CaT) to 945 mm over the Ionian zone (CaI).
Figure 1: Location of investigated regions, sub-regions and weather stations.

The two Calabrian sub-regions named Tyrrhenian and Ionian are
marked with CaT and CaI respectively; SiN, SiS, and SiE represent
North, South, and East zones of Sicily.
Sicily (Si) is the largest island of the Mediterranean Sea. According to

traditional geographical distinctions for making sub-regional analysis, Sicily was
divided into three homogeneous sub-regions (North, East, and South in Fig. 1).
Each sub-region is separated from the others by mountain ranges and for this
reason the rainfall pattern is different from area to area. Locally, the annual
rainfall Py is 591 mm in the South area (SiS), 735 mm over the East area (SiE)
and 717 within the North sub-region (SiN). Because of its particular geographic
location in the centre of the Mediterranean Sea, Sicily was often investigated as a
key region in order to explain the climate evolution within the Mediterranean
basin [8].
To date, the studies carried out in this area are limited on a few stations (5 for
temperatures and 11 for rainfalls – see Brunetti et al., [1], [2]) and for the reasons
explained above it could be of interest to extend this investigation on a greater
number of stations in order to give useful information on climate change in
Mediterranean areas.
The study proposed here aims at analysing temperatures and precipitations
over the two regions (Calabria and Sicily) for the period 1921-2000, using
weather stations with high spatial resolution. The analysis also includes the use
of bioclimatic indexes because in agro-forestry environments the consequences
of climatic change depend on the interaction between temperatures (maximum,
Tmax, minimum, Tmin, and mean, Ta) and rainfalls (Pm), that can be summarised by
appropriate indexes which account for aridity and plant water demand.
2 Data and methods

This paper investigates monthly values of temperature (Tmax, Tmin, and Ta) and
rainfall (Pm) collected by the Italian Hydrographic Service (IHS) during the
periods extending respectively from 1926 to 2000 and 1921 to 2000. The IHS
database concerning the two investigated regions includes about 120 temperature
and 600 raingauge stations.
Even if a considerable number of stations was considered, many of these sites
were neglected because long periods of malfunctioning.
The analysis was firstly carried out on a dataset including 53 temperature and
raingauge stations (25 of which in Ca and 28 in Si). These sites were selected
considering the dataset continuity as well as geographic and altitude
representativeness; datasets with more than 20% of lacks (with some exceptions)
were also neglected.
An additional database including only rainfall series (monthly, seasonal and
annual) and composed of 158 Calabrian stations was also used in the analysis in
order to get more details about the spatial variability over the investigated
regions. For this database an additional analysis involving the number of wet
days, Np, was carried out for the period 1951-2000. The analysed weather
stations are located as in figure 1.
Although no specific tests were applied, we suppose that the series used here
are homogeneous because they were collected using the same criterion that did
not change during the analysed period; also, specific tests involving Sicilian data
only, did not show particular anomalies due to malfunctioning [6].
The monthly values of Tmax, Tmin, Ta and Pm were also used to calculate the
following indexes:
a) the monthly reference evapotranspiration, ET0m, estimated by Hargreaves
formula [9] that, for sites where no direct radiation measurements are available,
assumes the following form:
ET0 m = 0.0023 Ra (Ta + 17.8 ) Tmax − Tmin d (1)
where Ra (mm d-1) is the extraterrestrial radiation, calculated on the basis of

latitude and calendar day [10]; d is the number of days of the month. The
Hargreaves equation was chosen because it represents a good compromise
between simplicity (because based on temperature data only) and goodness of
results in many environments [10] including Sicily [11];
b) the aridity index, AI, calculated using the following equation:
Py
AI = 12 (2)
∑ ET0 m, j
j =1
The monthly series of Tmax, Tmin, Ta and Pm were also analysed at sub-regional
scale within the two Calabrian (CaT and CaI) and the three Sicilian (SiN, SiE
and SiS) homogeneous sub-regions.
Seasonal and annual mean values of temperature and rainfall for each sub-
region were checked with the Mann-Kendall non-parametric test, as described in
Hirsch et al. [12] to look for a trend. The slope of the trends was calculated by
least-square linear fitting. An additional analysis involving the 10-years running
averages was carried out with the aims to show the long period tendencies.
The indices ET0m and AI were analysed by splitting the study period into two
sub-periods: from 1926 to 1962 and from 1963 to 2000. For each period, we
calculated the slope of the regression lines and the percentage of stations for
which the t-test [12] at the 0.05 level was significant.
3 Results
3.1 Yearly and seasonal temperature and precipitation analysis
The mean values b of the calculated slopes of the regression lines for the
investigated stations are listed in Table 1 together with the proportion of weather
stations for which the Mann-Kendall test showed significant values (at the 0.05
level of significance).
The Tya values showed increasing trends in all zones and sub-zones, with b
values ranging from +0.1 to +2.2 °C/100 years, with the only exception of CaT
(where b = - 1.4 °C/100 years). In general, the b values calculated for Si were
greater than those resulted in Ca. The higher increments occurred during the
winter season in Si and during the spring in Ca where Tya also showed a
decreasing trend in autumn and, only for CaT, in summer and winter.
The analysis carried out for Tmin showed in most cases b values greater than
those related to Tmax. The Tmin increase was similar in all seasons; conversely, the
Tmax increase was greater in autumn and in winter.
At seasonal scale, the analysis showed the highest b values in Si and in
particular for SiS.
The percent of weather stations for which the Mann-Kendall test (K, %)
showed significant values was always greater than 50% for the annual values of
Ta, Tmax and Tmin. At seasonal scale, the highest percentages occurred in Si, with
values greater than 80% particularly for Tmin.
Table 1: Mean values of linear regression coefficients (b), and percent of

stations with significant trend (Kendall’s test, K, %) for
temperatures (T) and rainfalls (P) of Sicily (Si) and Calabria (Ca).
T
P
zone Tmax Ta Tmin
b K b K b K b K
Spring Si 1.1 39 1.7 46 2.3 71 -7 4
SiN 1.2 29 1.4 43 1.6 57 2 0
SiE 1.1 22 1.8 56 2.5 89 -30 11
SiS 1.0 58 1.8 42 2.5 67 5 0
Ca 1.4 44 2.1 40 0.8 56 -43 15
CaT -1 43 2.3 43 -1.2 43 -62 13
CaI 2.7 44 0.5 22 1.1 39 7.9 0
Summer Si 1.3 29 1.9 57 2.4 75 8 7
SiN 1.7 29 1.5 43 1.3 57 3 0
SiE 1.2 22 2.0 56 2.7 89 2 0
SiS 1.2 33 2.0 67 2.9 75 15 17
Ca 0.7 48 1.4 36 0.3 56 37 22
CaT -3.2 43 -0.5 43 -0.3 43 30 42
CaI 2.1 50 0.1 33 1.1 33 106 21
Autumn Si 2.1 54 1.9 68 1.6 64 -82 36
SiN 1.5 29 1.4 43 1.4 57 -80 29
SiE 2.1 67 2.1 78 2.0 87 -97 22
SiS 2.4 58 2.0 75 1.5 67 -72 50
Ca 1.1 44 -0.4 56 -0.1 32 -128 39
CaT -0.1 29 -3.9 71 -2.6 14 -135 40
CaI 1.9 44 -0.4 50 0.4 11 43 1
Winter Si 2.8 71 2.6 75 2.5 68 -98 57
SiN 1.8 43 1.8 43 1.7 43 -105 57
SiE 2.6 89 2.6 89 2.6 78 -91 56
SiS 3.4 75 3.1 83 2.8 75 -100 58
Ca 1.1 28 0.8 28 0.6 48 -139 37
CaT -1.5 29 -1.0 29 -1.7 43 -176 47
CaI 2.2 22 0.6 22 1.1 28 41 0
Year Si 1.8 61 2.0 57 2.2 79 -179 43
SiN 1.5 57 1.5 57 1.5 71 -181 43
SiE 1.7 56 2.0 89 2.4 89 -212 44
SiS 2.0 67 2.2 75 2.4 75 -153 42
Ca 1.1 56 0.1 56 1.1 68 -272 47
CaT -3.2 43 -1.4 57 -2.9 43 -344 49
CaI 2.2 51 0.2 44 0.9 33 197 1
b = °C/100 years for T and mm / 100 years for P; N = North; E = East;
S = Sud; T = Tyrrhenian; I = Ionian.
The annual values of rainfall showed decreasing trends in all zones and sub-
zones, with b values ranging from -153 to -344 mm/100 years; the decrease was
greater for Ca while it covered about 21% of the mean annual rainfall of the
investigated period for both the regions. Where b was positive (CaI) the percent
of stations with significant trends was very low (1%) and can be neglected.
Negative trends occurred in autumn and winter, with b values ranging from -
72 to -176 mm/100 years, with the only exception of CaI. During these seasons,
although almost all the stations showed decreasing trends, these were significant,
for more than 50% of the sites, only for winter rainfalls of Si.
A greater variability occurred for spring rainfalls, while the summer ones
showed positive values of b (between 2 and 106 mm/100 years).
The rainfall decreasing trend seems to depend principally on the decrease of
Np (number of wet days); in particular, for Ca, during the period 1951-1998, Np
decreased of 32 days (b = 67 days/100 years); this decrease occurred especially
during winter and spring.
Referring to long-period tendencies, the 10-years running averages showed
complex and different behaviours from zone to zone and between the
investigated climatic variables. Tya showed a strong increase since the 70s
(Fig. 2) for Si; the trends are very similar from zone to zone. Conversely, the
corresponding trends in Ca, since the same period, seem to be decreasing
(Fig. 2). In this case, if the Tay series are divided into two sub-periods of equal
duration, it can be seen that in CaI, b is increasing during the first sub-period
(1926-1962) and almost equal to zero during the second one (1963-2000); in
CaT, b was almost equal to zero during the first period and negative during the
second one.
18
17.5
17
Mean temperature, °C
16.5
16
Calabria
15.5 Sicily
CaI
15 CaT
14.5
14
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Year
Figure 2: Ten-years running averages of annual mean temperature.
Py showed a decreasing trend, more clearly since the 50s, in both Si and Ca.
Even in this case, the 10-years running averages showed similar behaviours
within the Sicilian sub-regions; conversely, in Ca the regional results seem to
hide the increasing trend of CaI (Fig. 3).
1600
1400
1200
Precipitation, mm
1000
800
Calabria
Sicily
600
CaI
CaT
400
200
0
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Year
Figure 3: Ten-years running averages of annual precipitation.
Table 2: Mean linear regression coefficients (b), and percent of stations with
significant trend (t-test, %) for Evapotranspiration (ET0) and
Aridity index (AI) for Sicily in the sub-periods 1926-62 and 1963-
2000.
ET0
AI
Spring Summer Autumn Winter Year
1926-62 b 29 82 21 22 154 -0.2
t, % 25 25 25 25 25 18
1963-2000 b 58 35 7 40 150 -0.3
t, % 50 25 39 46 46 14
b = mm/100 year for ET0; mm.mm-1/100 years for AI.
3.2 Bioclimatic indices analysis
The annual values of ET0 in Si showed b values always positive and almost the
same for the two investigated sub-periods during which its increase was about
5% of the ET0 values calculated for the same sub-periods. At seasonal scale, the
greatest rise occurred in summer, during the first sub-period (1926-1962) and in
spring during the second one (1963-2000).
The proportion of weather stations that showed significant trends (t-test at
0.05 level of significance) was greater during the second sub-period and this
behaviour can be explained by the greater temperature increase occurred during
the same period. The aridity index AI showed a decrease in both the sub-periods,
particularly during the second one (1963-2000), where it assumes a mean value
equal to 0.62 (typical for sub-humid dry climate) that is less than the first period
where its calculated mean value resulted equal to 0.70 (typical for sub-humid
climate). The trends are significant (at the 0.05 level) for less than 20% of the
considered weather stations.
4 Conclusions
The analysis carried out in this study which involved temperature and
precipitation data covering a period of about 80 years, showed an increase of
temperature (ranging from 0.1 and 2.2°C/100 years for the mean annual
temperature) and a decrease of precipitation (ranging from 153 to 344 mm/100
years for the annual rainfall) over the two investigated regions (Calabria and
Sicily) located in South Italy. The precipitation decrease can be explained by the
decrease of the number of wet days NP according to other studies performed in
different Italian regions. The calculated values of evapotranspiration ET0, which
accounted for mean (Ta), maximum (Tmax) and minimum (Tmin) values of
temperature, showed an increasing trend which was lower than that related to the
temperature because Tmax and Tmin did not increase with the same magnitude.
The aridity conditions are not encouraging during the analysed period: the
calculated Aridity Index (AI) showed values which tend to decrease from a first
sub-period (1926-1962) to a second one (1963-2000). This resulted in a climate
change in Sicily from a sub-humid climate to a sub-humid dry climate.
Because of the geographic location and the particular orographic features of
the two investigated regions which extend over 3° of latitude, a strong difference
in terms of climate change occurred. These main differences can be summarised
as follows: a strong increase in temperature has occurred in Sicily since the
1970s, while in Calabria the same variable showed steady values or a light
decrease; Tmax and Tmin values have clearly decreased during all the seasons over
the Calabrian Jonian sub-region (-3.2 e -3.9 °C/100 years respectively for Tmax
and Tmin); rainfalls have increased over the Calabrian Ionian sub-region (197
mm/100 years for Py).
The overall results showed the importance of the climate change analyses at a
regional scale and provide the Italian authorities with useful information to begin
to assess the impacts of such climate changes on human activities.
References
[1] Brunetti, M., Buffoni, L. & Mangianti, F., Temperature, precipitation and
extreme events during the last century in Italy, Global and Planetary
Change, 40, pp. 141-149, 2004.
[2] Brunetti, M., Buffoni, L., Maugeri & M, Nanni, T., Trend of minimum
and maximum daily temperatures in Italy from 1865 to 1996, Theor. Appl.
Climatol., 20, pp. 1017–1031, 2000.
[3] Capra, A., Indelicato, S., Li Destri Nicosia, O. & Scicolone, B.,
Evaluation de la sécheresse d’aprés les données de précipitation. Une
application au Sud d’Italie, Proc. of the 16th European Regional Conf.,
ICID, Budapest, Hungary, June, pp. 41-51, 1992.
[4] Capra, A., Li Destri Nicosia, O. & Scicolone B., Application of fuzzy sets
to drought classification, Proc. of the 2nd Int. Conf. On Advances in Water
Resources Technology and Management, Eds. G. Tsakiris & M.A. Santos,
Balkema, Rotterdam, pp. 479-483, 1994.
[5] Piccarreta, M., Capolongo, D. & Boenzi, F., Trend analysis of
precipitation and drought in Basilicata from 1923 to 2000 within a
southern Italy context, Int. Journal of Climatology, 24, pp. 907-922, 2002.
[6] Rossi, G. & Cancelliere, A., Problemi e prospettive del monitoraggio e
delle mitigazione della siccità, Editoriale Bios, Cosenza, Quaderni di
Idrotecnica, pp. 9-44, 2002 (in Italian).
[7] Capra A., Malara, S.L. & Scicolone, B., (2004), Analisi delle temperature
e delle piogge mensili in Calabria nell’ultimo cinquantennio, Economia
Montana- Linea Ecologica, 3, pp. 31-36, 2004 (in Italian).
[8] ENEA, Cambiamenti climatici e rischi di siccità e desertificazione in area
mediterranea ed in Sicilia, Progetto speciale clima globale, Roma, Italy,
giugno, 2002 (in Italian).
[9] Hargreaves, G.H. & Samani, Z.A., Reference crop evapotranspiration
from temperature. Transactions of the ASAE, 1(2), 96–99, 1985.
[10] FAO, Crop evapotranspiration. Guidelines for computing crop water
requirements, FAO Irrigation and drainage Paper, 56, 1998.
[11] Di Stefano, C. & Ferro, V., Estimation of Evapotranspiration by
Hargreaves Formula and Remotely Sensed Data in Semi-arid
Mediterranean Areas, J. Agric. Engng Res., 68, 189–199, 1997.
[12] Hirsch, R.M., Helsel, D.R., Cohm, T.A. & Gilroy, E.J., Statistical analysis
of hydrological data, Handbook of Hydrology, Ed. Maidment, D.R.,
McGraw-Hill, pp. 17.1-17.55, 1993.
Section 5
Geo-environment in
urban settings
The urban geo-science model:

an essential tool to support planning and
sustainable development
D. McC. Bridge, B. L. Morris & J. R. A. Giles
British Geological Survey, Nottingham, UK
Abstract
New technology for viewing and sharing digital information means that national
geological surveys can present geo-science information in innovative ways that
take better account of the third dimension. The British Geological Survey is
piloting the development of 3-D geo-science models in four conurbations across
the UK. These models are being developed with a view to supporting sustainable
development, particularly in the context of urban groundwater management and
development planning. The work is research based but is being guided by the
very specific needs of regulatory bodies and local authorities. The success of this
initiative will depend ultimately on creating greater awareness amongst the user
community of the value of 3-D models, and ensuring that relevant information is
made available in advance of major infrastructure projects.
Keywords: urban geoscience, 3-D modelling, groundwater, development
planning.
1 Introduction
Most of the 60 million people that make up the population of Britain live in
urban and suburban areas that comprise only 10 per cent of the countries’ surface
area. The principal conurbations (Greater London, West Midlands, Greater
Manchester, West Yorkshire, and the Central Lowlands of Scotland) expanded
dramatically during the industrial revolution, in part, due to the ready availability
of local raw materials (coal, water, ironstone). Some of these conurbations
directly overlie productive aquifers where groundwater is a resource; others are
underlain by less permeable rocks where urban drainage and the disposal of
water from the urban infrastructure to the subsurface is a problem. Both require
doi:10.2495/GEO060191
an understanding of subsurface geology and hydrogeology to guide urban

planners, inform urban infrastructure investment and enable sustainable
development.
The Government’s promotion of urban regeneration and redevelopment of
brownfield sites places an additional premium on understanding the conditions
of the ground before development, and the likely impacts both at the construction
stage and subsequently, when the land is restored to use.
2 The challenge
The provision of information on ground conditions is not a new idea. William
Smith’s first geological map, published in 1820, was compiled to meet a
practical need – to document the principal mineral resources of the country.
Subsequent refinements by the British Geological Survey (BGS) over the past
170 years mean that the UK is now covered by the most detailed set of
geological maps of any country in the world. One of the problems facing any
national survey, however, is how best to make information available at a level of
detail that meets the needs of decision makers, not all of whom are necessarily
geologists.
The applied mapping programme of urban conurbations, commissioned by
Central Government during the 1980s and 1990s, was an ambitious attempt to
address this problem. It aimed to provide accessible advice on the relevance of
geological conditions to strategic land use planning e.g. Forster et al [1].
Although the initiative was successful at the technical level, the hard copy
outputs were expensive to produce and there was no provision for updating the
underlying databases or the map outputs.
Recent advances in modelling software and web delivery options make the
task of integrating and delivering geo-science information a practical reality.
However, raising awareness amongst policy makers and planners of the role of
geo-science information remains a significant challenge.
3 The 3-D approach

As part of its remit to make geological information available to a wider audience,
the British Geological Survey (BGS) has embarked on a national programme of
3-D modelling. Within this programme, work has begun in four major
development areas (Thames Gateway, Manchester, Clyde Basin and Belfast) to
build high-resolution geo-science models capable of supporting decisions on a
range of geological, engineering and hydrogeological topics.
These urban models concentrate on the shallow subsurface (zone of human
interaction) and are constructed primarily from downhole information collected
in the course of site investigation studies. A typical urban model may incorporate
information from several thousand boreholes, fig 1. Modelling of Quaternary
sediments is carried out using proprietary software developed at the University
of Cologne, Kessler and Mathers [2]. In more complex, faulted bedrock
sequences, Gocad® is the preferred package. There is complete interoperability

between the two systems and outputs can be combined and exported in standard
format.
Figure 1: 3-D model of part of Trafford Park, Manchester showing selected

Quaternary elements.
4 The role of the 3-D model

The outputs from the 3D geological model can be tailored to a range of end uses,
and have direct application in several key areas, including urban groundwater
management, planning for development and preliminary site appraisal. Three
case studies illustrate these different applications.
4.1 Groundwater management, Trafford Park, Manchester
One consequence of the European Water Framework Directive is that, as surface

and groundwater safeguards become more stringent, there will be a greater need
to manage urban water resources in a more sustainable manner. The use of
geological models to inform and better represent the complexities of the
subsurface in groundwater flow models and contaminant studies is now
beginning to interest regulatory bodies and hydrogeological consultants. In the
Trafford Park industrial zone of Manchester, the sensitivity of the Triassic
sandstone bedrock aquifer to pollution and the extent to which recharge may
occur have been analysed through detailed characterisation of the underlying
superficial deposits. Potential hydrogeological pathways from ground surface to
the sandstone have been identified, and thematic outputs show the importance of
the Manchester Ship Canal and related waterways as potential sources of
recharge and pollution of the bedrock aquifer, figs. 2 & 3.
Figure 2: Synthetic section derived from 3-D model showing inferred

recharge pathways.
Figure 3: High vulnerability /recharge zones (shown cross-hatched) in

Trafford Park, Manchester.
4.2 Development planning
‘Unforeseen ground conditions’ are frequently cited as the main cause of project
overruns and escalating development costs. In Scotland, Glasgow City Council,
working in collaboration with BGS, has taken the innovative step of making
available their detailed borehole databases for incorporation in a 3-D superficial
and bedrock model now being built for the city centre. This model, when
completed, will provide a representation of ground conditions at local and
regional scale and inform planning of subsequent investigation phases. Emphasis
is being placed on the acquisition of mining information with a view to
delineating areas within the city where development may be compromised by
unrecorded, shallow coal workings.
Figure 4: Property damage caused by shallow undermining for coal. Belmont

Street, Glasgow.
4.3 Preliminary site assessment
At project concept stage, modern visualisation techniques and virtual models are
used increasingly by design consultants to illustrate the environmental impact of
major new infrastructure. This technology can be equally applied to the
subsurface, where the 3-D model can provide a picture of probable ground
conditions to facilitate preliminary design and costing. The example, fig. 5,
shows how, by model interrogation, a range of geological and geotechnical
information can be quickly assembled to plan a focused and cost-effective site
investigation programme.
Figure 5: Rapid site assessment based on interrogation of 3-D model,

attributed with geotechnical data.
5 Information management: the real cost of 3-D modelling

3-D geological models are only as good as the data that underpins them. The cost
of acquisition of new data is prohibitive, therefore effort is focused on prior
information, Wood and Curtis [3]. Borehole information is acquired by a wide
range of exploration industries (e.g. hydrocarbon, coal, water, site
investigation). Such information has to be collected from the companies, either
under statute (Water Resources Act), or by donation from the company after
careful negotiation. Once the original records have been acquired, the
information management process begins. This is a non-trivial task, which
requires considerable resources and a wide range of skills from a number of non-
geological disciplines. Because of the volumes of data involved (over one

million records held in the National Geoscience Records Centre) data
management has to follow clear procedures as outlined by Feinemann [4].
From the 3D modelling viewpoint, the key requirements are to:
• populate the databases
• manage the domains associated with the databases to ensure
interoperability
• make the data available via appropriate applications
6 Emerging issues
The primary output of the modelling process is a 3-D geological model with a
resolution designed to support the urban map scale of 1:10 000. For many
applications, the geological model is sufficient in itself to provide answers to
many development problems. The extent to which other information
(geotechnical, hydrogeological, geochemical) is incorporated in the model is
likely to be driven by legislative requirements. In the context of groundwater, the
emerging topic include:
6.1 Infiltration from rainfall, surface drainage and the urban pipe
infrastructure
One of the major uncertainties of urbanisation is the urban water budget, which
is much more complicated to calculate than its rural equivalent. Get it wrong and
the resulting uncertainty can be costly as groundwater levels rise beyond their
historical levels, causing major problems of flooding (tunnels, basements,
underground car parks) and foundation stability.
6.2 Point-source pollution from contaminated land
As brownfield sites are targeted for redevelopment, planners are increasingly

interested in the ability of the subsurface to detain, attenuate or dilute the
products of former polluting activities, especially if the contaminants are toxic.
6.3 Sustainable urban drainage systems
SUDS technologies are the subject of much discussion in urban engineering

fields, being seen by protagonists as a cost-effective way to limit the complexity
and consequences of runoff control in built areas and by critics as having the
potentially harmful side-effect of dispersing pollutant loads. The answer to many
questions over their use lies not merely in the mechanics of the diverse
engineering structures being developed as SUDS systems but in the ability of the
subsurface (soils, regolith/unsaturated zone and saturated zone) to remove,
attenuate or mediate substances of concern.
7 Conclusions
3-D urban geo-science models complement and enhance the information
portrayed on traditional 2-D geological maps. Although access to, and use of 3-D
urban geo-science models is currently limited, there is increasing recognition
among users of the added value offered by acquiring information in 3-D rather
than 2-D. The challenge is to make the models available at an early stage in the
urban regeneration process so that they may better inform strategic planning
options, ground investigation and reclamation strategies. This is reliant on the
models being appropriate in terms of content and scale, and that inherent
uncertainties associated with the model are appropriately quantified.
Acknowledgement
This paper is published with the permission of the Executive Director of the
British Geological Survey (NERC).
References
[1] Forster, A., Arrick, A., Culshaw, M. G. & Johnston, M., (eds) A geological
background for planning and development in Wigan, British Geological
Survey Technical Report No. WN/95/3: Keyworth Nottingham, 1993.
[2] Kessler, H. & Mathers, S. Maps to models. Geoscientist, 14(10) pp. 1- 6.
[3] Wood, R. & Curtis, A. Geological prior information and its application to
geoscience problems. Geological prior information: informing science and
engineering , Wood, R. and Curtis, A. (eds) Geological Society, London,
Special Publication, 239, pp.1-14, 2004.
[4] Feineman, D. R., Data management : Yesterday, today and tomorrow. Petex
‘92 Conference. London, 1992.
Developing design-oriented strategies to

combat regional scale climate change
B. Stone, Jr
Georgia Institute of Technology, City and Regional Planning Program,
USA
Abstract
This paper presents a study of residential parcel design and surface heat island
formation in a major metropolitan region of the southeastern United States.
Through the integration of high-resolution multispectral data (10 m) recorded by
the National Aeronautical and Space Administration (NASA) with property tax
records for over 100,000 single-family residential parcels in the Atlanta, Georgia
metropolitan region, the influence of the size and material composition of
residential land use on an indicator of surface heat island formation is reported.
In contrast to previous work on the urban heat island effect, this study derives a
parcel-based indicator of surface warming to permit the impact of land use
planning regulations governing the density and design of development on the
excess flux of heat energy to be measured. The results of this study suggest that
the contribution of individual land parcels to regional surface heat island
formation could be reduced by approximately 40% for average sized parcels
through the adoption of specific land use planning policies, such as zoning and
subdivision regulations, and with no modifications to the size or albedo of the
residential structure.
Keywords: climate change, urban heat island effect, land use, urban planning.
1 Introduction
The significance of land use to climate change has only recently begun to receive
serious attention within the technical literature. As reported by Kalnay and Cai
[1], analyses of both surface and atmospheric temperature trends over the last
half century suggest that land use change may be responsible for as much as 50%
of the observed reduction in the mean diurnal temperature range – a phenomenon
doi:10.2495/GEO060201
largely attributed to increasing minimum temperatures throughout the

continental United States. Two principal causes of this increase in minimum
temperatures, agriculture and the urban heat island effect, are identified as
significant drivers of ongoing changes in climate. In short, an increase in surface
temperatures resulting from land cover changes, independent of ongoing
increases in greenhouse gas concentrations, appears to be a significant driver of
climate change. In a similar vein, Foley et al [2] conclude from a global analysis
of land use that while “[l]and use has generally been considered a local
environmental issue … it is becoming a force of global importance” (p. 570).
Despite a growing recognition of the importance of land use to climate
change, the suite of mitigation strategies embodied in the still-evolving Kyoto
framework remain almost entirely technological in nature, with little attention to
strategies rooted in the design of human settlements. Because the process of
climate change is driven by two distinct mechanisms – an increase in the heat
absorbing potential of the atmosphere and an increase in surface heat emissions
– there is a critical need to focus not only on greenhouse gas reductions but also
on strategies to reduce surface heat emissions at the local level.
To address this issue, this article presents the results of a study of residential
parcel design and surface warming patterns in the Atlanta, Georgia (USA)
metropolitan region. With the aid of high resolution thermal infrared (IR)
imagery and a parcel level geographic information system, the thermal properties
of over 100,000 single family parcels are associated with the spatial structure and
material composition of residential development throughout a 48 km2 study
region. In measuring the thermal properties of individual land parcels, an
“emissivity adjusted” indicator of the parcel black body flux is developed to
account for the influence of variable surface emissivity and sky radiation
captured in the thermal IR imagery. Furthermore, the predevelopment radiant
flux is estimated and subtracted from the observed, post-development radiant
flux to differentiate the warming impacts of urban development from natural,
background sources of surface heat.
2 Background
In order to associate remotely sensed thermal IR data with land use planning
policies it is essential that a “policy-relevant” indicator of surface warming be
developed. While remote sensing technologies can provide a valuable tool for
assessing the thermal performance of urban environments, these must be used
with great caution to yield reliable information for land use planning. Most
importantly, there is a critical need to isolate the influence of specific classes of
land use on surface warming patterns. While previous work has demonstrated a
clear distinction between the thermal properties of “urban” and “rural” land
cover types, each of these broad categories of development is composed of a
number of specific land use classes that are subject to an array of development
regulations. If we are to mitigate heat island formation through the modification
of specific zoning and subdivision regulations, we must be able to isolate the
influence of these individual policies on surface warming. To be policy-relevant,
an indicator of surface warming must be capable of quantifying the thermal

characteristics of distinct and legally defined classes of land use.
In this section of the paper, I discuss three criteria that must be satisfied to
derive a policy-relevant measure of heat island formation. These include: 1) the
measurement of surface warming at the level of the individual land parcel; 2) the
adoption of a flux-based rather than a temperature-based indicator to capture the
effects of land area on warming; and 3) the adjustment of the radiant flux to
more closely correspond to the sensible flux of heat.
2.1 Measuring thermal performance at the parcel level
In assessing the surface thermal properties of urban and rural land covers, many
studies have adopted a unit of analysis consistent with the resolving dimension
of a satellite radiometer, such as 1.1 km or 120 m (e.g., Roth et al [3]; Nichol
[4]). An important limitation of adopting such a spatially uniform unit of
analysis, however, is that it fails to conform to the irregularly shaped boundaries
of individual land parcels, the unit of area at which land use is controlled in the
United States. For example, in dense and well-mixed urban districts, a spatially
uniform unit of analysis is likely to capture both commercial and residential land
uses, confounding any attempt to isolate the thermal properties of a single land
use class. Such an outcome may be categorized as a special instance of mixed
pixel error – mixed, that is, in terms of land use class rather than land cover type.
The limitation of this type of mixed pixel error for planning analysis is that it
obscures the underlying policy drivers of environmental performance. If we are
interested in evaluating the environmental impacts of a particular planning
policy, such as maximum impervious area regulations, we must first identify
those zones subject to the policy. For example, within a square kilometer of
urban land there is likely to be a number of unique land use classes, each with its
own zoning regulations governing the permissible area of impervious cover per
parcel. In measuring the environmental characteristics of an area of land subject
to multiple zoning regulations, it is difficult to assess the relative influence of
any single policy on environmental performance. For the purposes of planning
research, a more methodologically sound approach entails the measurement of
surface warming at a dimension compatible with land use regulation, such as a
zoning district or individual land parcel. As adopted herein, a parcel-based
approach to surface warming analysis permits parcel-specific land use policies,
such as zoning regulations, subdivision regulations, and building codes, to be
associated with thermal performance at the same spatial dimension.
2.2 Accounting for the effects of land area on surface warming
While parcel-based measures of surface warming are needed to associate heat

island formation with land use policies, the adoption of a spatially irregular unit
of analysis complicates the use of surface temperature as an indicator of
warming. As noted by Price [5] over 25 years ago, temperature-based
measurements, such as the urban—rural temperature differential, often fail to
reflect the magnitude of heat islands created by urban areas of different sizes.
Price states, “[f]or investigation of urban heating, the peak temperature is less
significant than a summation (area integral) of the excess power radiated as a
result of the surface temperature elevation” (p. 1557). In other words, a measure
of the excess flux of thermal energy from a city provides a more accurate
indicator of the total magnitude of warming within an urbanized area than does
the urban—rural temperature differential.
In this sense, the magnitude of surface warming between urban and rural
zones (defined herein as the excess flux of sensible heat energy) is more relevant
to urban planning policy than heat island intensity (the urban – rural temperature
differential). The reason for this is that the area-integrated flux of heat energy
emitted from a city or parcel provides a more accurate indicator of the volume of
the near surface atmosphere influenced by elevated surface temperatures than
does the surface temperature differential. In this sense, two cities of similar
climate and development type but dissimilar land area are likely to exhibit
similar average surface temperatures but different fluxes of energy. As the area
of a city expands, the volume of the heat island “dome” – a volume subject to
elevated air temperatures and enhanced pollution formation – expands as well.
Measures of heat island intensity across cities may not reflect a differential in the
volume of heat island domes. While smaller in scale, the same is true when
comparing two parcels of different area.
2.3 Approximating the sensible heat flux
While aerial and satellite radiometers provide the most viable means of obtaining
continuous surface thermal data across a large study region, a principal limitation
of remote sensing studies of the urban heat island effect is a failure to account for
the variable distribution of surface emissivity. Given the wide array of surface
material types found in urban areas, ranging from heavily canopied areas to the
impervious materials of driveways and rooftops, radiant flux densities recorded
by thermal sensors are prone to systematically underestimate thermodynamic
temperatures from the heterogeneous surfaces of urban environments. As a
result, the surface radiant flux density must be adjusted to more closely
approximate the black body flux density, which is a better indicator of sensible
heat.
In summary, to be useful in planning research, an indicator of the urban heat
island effect must be spatially compatible with the legal dimensions of land use
control, and must accurately approximate the quantity of energy contributed by a
single land parcel to elevated temperatures within cities. What follows in the
remainder of the paper is an application of this conceptual approach to the
Atlanta, Georgia metropolitan region and a presentation of policy insights that
may be gleaned from this analysis.
3 Methods
In May of 1997, the National Aeronautic and Space Administration (NASA)
obtained high resolution multispectral imagery over the Atlanta, Georgia
metropolitan region for the purpose of investigating the influence of land use on
urban heat island formation. Recorded from an aerial platform with the
Advanced Thermal and Land Applications Sensor (ATLAS) at a ground
resolution of 10 meters, 15 channels of multispectral data were obtained over a
48 km2 region by NASA during the course of two flights and made available to a
range of research efforts known collectively as “Project Atlanta.” The
exceptionally high spatial resolution of this data permits the thermal
characteristics of high-density urban parcels to be reliably measured and
integrated into a parcel level geographic information system. With the aid of the
Project Atlanta data, a measure of the parcel “net black body flux” was derived
for every single family residential lot in the Atlanta study region. What follows
is a brief overview of the methods employed to derive a measure of the net black
body flux based on remotely sensed data; for a full discussion of this
methodology, please see Stone and Norman [6].
3.1 Deriving the net black body flux
The first step in estimating the parcel net black body flux requires that the
thermal characteristics of individual land parcels be measured. To do so, parcel
boundary information for the approximately 104,000 single-family parcels
located in the study region was obtained from the City of Atlanta and Fulton
County offices of tax assessment. By registering the multispectral imagery and
parcel geography data layers to the same coordinate system, it was possible to
spatially overlay the two datasets in a geographic information system. Once
integrated, the average apparent radiant flux density of each parcel was derived
through a zonal summary function.
In addition to the apparent radiant flux density, information on the composite
parcel emissivity and emitted sky radiation was needed to derive the black body
flux density. Employing the method of Carlson et al [7], the fraction of
vegetative cover for each land parcel was estimated through the derivation of a
normalized difference vegetation index (NDVI) with ATLAS bands in the red
and near infra-red spectral regions. Based on published data for vegetative and
impervious materials characteristic of residential zones, an emissivity value of
0.96 was assigned to the vegetative component of each parcel and a value of 0.92
was assigned to the impervious component (Jensen [8]; Oke [9]). The composite
parcel emissivity was then calculated based on the area of pervious and
impervious materials per parcel and used, in concert with the average value of
emitted sky radiation for the day of ATLAS data collection, to compute the
estimated parcel black body flux density.
As a final step, the quantity of radiant energy attributable to predevelopment
land characteristics – the rural “background” radiation – was subtracted from the
measured average parcel black body flux density. In deriving an indicator of the
urban heat island effect, we are interested in the thermal differential between
developed and undeveloped land features. When measuring heat island intensity,
this variable is quantified as the temperature differential between developed and
undeveloped zones. When measuring heat island magnitude, this variable
becomes the flux differential between developed and undeveloped zones. To
calculate this flux differential at the parcel level, we first calculate the average
black body flux density for a developed parcel and then subtract from this
quantity an estimate of the predevelopment black body flux density that would
result were the same parcel occupied by a natural land cover, such as forest
canopy cover. This flux density differential is then multiplied by the parcel area
to estimate the parcel net black body flux.
3.2 Measuring parcel design
Following the development of the parcel net black body flux measure, four
parcel design variables were derived for this study based on city and county
property tax records and the multispectral ATLAS imagery. These included the
area of impervious materials, such as driveways and building footprints, the area
of lawn and landscaping, the proportion of the parcel overlaid by tree canopy,
and the number of bedrooms in the residential structure. This final variable was
used to control for residential capacity. In short, I am interested in comparing
the thermal characteristics of parcels varying in size and material composition
but similar in the number of residents the parcel was designed to accommodate.
The inclusion of this variable in our analysis ensures that differences in thermal
performance by region of the city are not attributable to differences in residential
capacity.
Information on the area, footprint size of residential buildings (houses and
detached buildings), and number of bedrooms in the residential structure was
obtained from the City of Atlanta and Fulton County offices of tax assessment.
As driveway areas are not recorded in the property assessment process, a sample
of parcels stratified by age and size was selected to estimate the area of these
paved surfaces. The driveways of selected parcels were then measured directly
to determine how paved areas scale with parcel size throughout the study region.
In combination, the estimated area of driveway paving and the area of the
building footprint constitute the impervious component of the parcel. The area
of lawn and landscaping was derived by subtracting the estimated impervious
area from the total parcel area. Finally, the proportion of each parcel overlaid by
tree canopy was estimated from the NDVI.
4 Analysis and results

With the aid of the Statistical Package for the Social Sciences (SPSS), an
analysis was performed to assess the influence of parcel design on the parcel net
black body flux. The average single family residential parcel in the database is
characterized by a three bedroom house and is approximately 1,360 m2 in area,
of which 208 m2 or 15% is occupied by the impervious components of the house
and driveway, with the remaining 85% occupied by lawn and landscaping.
Approximately 45% of the average single-family residential parcel is overlaid by
tree canopy cover.
To evaluate the graphic correlation between parcel area, material
composition, and surface warming, values of increasing lot size are plotted
against impervious area, lawn area, and the net black body flux in Figure 1. As
illustrated in this figure, the magnitude of surface warming scales closely with
lot size, with the mean net black body flux increasing by a factor of about six
between the highest and lowest density classes. While this descriptive evidence
supports my hypothesis of a negative relationship between the density of single-
family residential development and surface warming, it is important to note that
this simple covariation does not account for the distribution of residential
capacity or tree canopy cover throughout the region. If larger lot sizes are also
associated with greater residential capacities (e.g., four and five bedroom houses)
and a less mature tree canopy cover, then the relationship between the density of
development and thermal performance may be attributable to incompatible
design objectives or to the age of development in different areas of the city. In
order to control for these important influences, a multivariate statistical model is
developed and evaluated in the following section.
4000 8000
3500 7000
Net Black Body Flux (W)

3000 6000
Area of Land Cover
2500 5000
(sq. meters)
2000 4000
1500 3000
1000 2000
500 1000
0 0
0 - 500 500 - 1,000 1,000 - 2,000 2,000 - 4,000 > 4,000
Lot Size (sq. meters)
Imperv Area Lawn Area Net Black Body Flux
Figure 1: Parcel attributes by lot size.
4.1 Explanatory analysis
In order to quantify the potential thermal benefits of specific design based

strategies, ordinary least squares was used to develop a predictive model linking
parcel design to the net black body flux. Four independent variables were
incorporated into the model to explain the net black body flux. These included
the area of impervious cover, the area of pervious cover or lawns and
landscaping, the percentage of the parcel overlaid by tree canopy, and the
number of bedrooms in the residential structure.
With the aid of SPSS, a set of parameter estimates and model summary
statistics were generated for the approximately 104,000 single family residential
parcels located in the Atlanta study region. The results of this modeling process
are presented in Table 1.
The results of the regression analysis indicate that the various parcel design
attributes are significantly related to the parcel net black body flux. As expected,
the area of impervious and pervious materials were found to exhibit a significant
positive relationship with parcel warming when controlling for other parcel
design attributes believed to be associated with the dependent variable. Also as
expected, the presence of tree canopy cover was found to significantly decrease
the magnitude of warming per parcel. These relationships were found to hold
when controlling for the number of bedrooms in the residential structure,
suggesting that the elevated surface warming is not solely attributable to
variation in residential capacity.
Table 1: Regression analysis results for parcel net black body flux (W)*.
Variable B-Coefficient Standardized Significance

Coefficient
Impervious Area (m2) 0.079 0.274 < .001
Lawn Area (m2) 0.013 0.553 < .001
Tree Canopy Cover (%) -23.93 -0.419 < .001
Number of Bedrooms 0.438 0.019 < .001
Summary Statistics Adj. R-Square F-Statistic F Significance

0.57 35,217 < .001
* The square root of the net black body flux was used in this model as a variance
stabilization transformation.
5 Discussion and conclusions

The results of this analysis provide compelling evidence that the size and
material composition of single family residential parcels are significantly related
to the magnitude of surface warming in the Atlanta study region. Specifically,
smaller, higher density parcels were found to be associated with a lower net
black body flux than larger, lower density parcels when controlling for the class
of land use and the number of bedrooms in the residential structure. While both
the area of impervious materials and lawn and landscaping were found to be
positively related to the parcel net black body flux, the area of lawn and
landscaping – a strong correlate of parcel size – was found to be the strongest
predictor of excess surface warming.
In contrast to previous work on the urban heat island (e.g., Hoyano [10]), the
results of this study support the hypothesis that lower density, dispersed patterns
of urban residential development contribute more surface energy to regional heat
island formation than do higher density, compact forms. While specific to the
Atlanta metropolitan region, due to the uniformity of building materials and
development patterns found throughout North American cities, these findings are
generalizable to a number of large, mid-latitude cities characterized by a similar

climatological and physiographic characteristics. For such cities, this work
illuminates at least three specific planning strategies that may be employed to
offset the excess surface heat energy emitted from both existing and future
residential development, including reductions in the area of residential lawn
space, reductions in the area of impervious cover, and an increase in tree canopy
cover. Each of these strategies is discussed briefly in turn.
The results of this analysis suggest that the most effective design-oriented
strategy for mitigating the thermal impacts of new housing development is a
reduction in the area of the residential lawn. For the average parcel, a reduction
in the lawn area of 25%, holding other design attributes constant, would be
associated with a 14% reduction in the parcel net black body flux. By serving to
increase the density of land use, a reduction in the lawn area of new residential
development would offset enhanced surface warming through at least two
mechanisms. First, by reducing the average lot size of development at the urban
periphery, the area of rural land converted from forest, Atlanta’s natural land
cover, to urban land uses is reduced, constraining the zone of enhanced warming.
Second, by limiting the total zone of urban expansion, higher density
development limits the extension of roadways and other thermally intensive
infrastructure, also serving to preserve rural landscapes.
The finding of a positive relationship between the area of impervious cover
and the parcel net black body flux confirms the well established role of mineral-
based building and paving materials in urban heat island formation. The most
widely advocated approach to mitigating the thermal impacts of impervious
materials is through the application of high albedo surface treatments to increase
surface reflectivity – an approach that is cost effective and applicable to both
new and existing development. For new construction, a reduction in the total
area of impervious cover provides an additional tool for offsetting an increase in
surface warming. These findings suggest that a 25% reduction in the impervious
cover of an average single-family parcel would be associated with a 15%
reduction in the net black body flux. When combined with an equivalent
reduction in lawn area, parcel warming was found to be 27% lower.
While the potential to reduce the thermal impacts of new growth are
significant, it is important to note that changes to a city’s land development
regulations will have only a limited influence on existing development. As the
vast majority of the Atlanta region’s 2020 built area is already in place, changes
in future peripheral development will serve to offset continued growth rather
than to abate present day heat island formation. In light of this observation, the
most effective design-oriented strategies for reducing total regional warming
must address both new and existing development. As noted above, one strategy
for doing so would entail the replacement of traditional driveways with driveway
runners at the time of routine resurfacing. A second critical strategy for existing
development entails the planting of trees. As indicated by our analysis, for the
average single family parcel, an increase in tree canopy cover from 45 to 60%
reduces the parcel net black body flux by 13%. For trees strategically planted
along roadways and in proximity to houses, the thermal benefits are likely to be
greater. In combination, a 25% reduction in the areas of pervious and

impervious materials, and an increase in parcel tree canopy cover from 45 to
60%, were found to reduce the net black body flux from a parcel of average size
by approximately 40%.
References
[1] Kalnay, E. & Cai, M., Impact of urbanization and land use change on
climate. Nature, 423, 528-531, 2003.
[2] Foley, J., Defries, R., et al, Global consequences of land use. Science,
309, 570-574, 2005.
[3] Roth, M., Oke, T., Emery, W., Satellite-derived urban heat islands from
three coastal cities and the utilization of such data in urban climatology.
International Journal of Remote Sensing 10, 1699-1720, 1989.
[4] Nichol, J., High-resolution surface temperature patterns related to urban
morphology in a tropical city: A satellite-based study. Journal of Applied
Meteorology 135, 135-46, 1996.
[5] Price, J., Notes and correspondence: Assessment of the urban heat island
effect through the use of satellite data. Monthly Weather Review 107,
1554 -1557, 1979.
[6] Stone Jr., B., Norman, J., Land use planning and surface heat island
formation: A parcel based radiation flux based approach. Atmospheric
Environment, forthcoming. Study methodology reprinted with permission
from Elsevier.
[7] Carlson, T., Capehart, W., Gillies, R., A new look at the simplified
method for remote sensing of daily evapotranspiration, Remote Sensing of
Environment 54, 161-167, 1995.
[8] Jensen, J., Remote sensing of the environment: An earth resource
perspective. Prentice Hall, Upper Saddle River, NJ, 2000.
[9] Oke, T., Boundary layer climates, Routledge, New York, NY, 1987.
[10] Hoyano, A., Relationships between the type of residential area and the
aspects of surface temperature and solar reflectance (based on digital
image analysis using airborne multispectral scanner data). Energy and
Buildings 7, 159-73, 1984.
Land subsidence evolution and controlling

mechanisms near Mexico City:
environmental planning and management
M. A. Ortega-Guerrero
Centro de Geociencias,
Universidad Nacional Autónoma de México, México
Abstract
Mexico City is situated in the Basin of Mexico on a highly compressible
lacustrine aquitard that overly highly productive aquifers of both volcanic and
sedimentary origin. This aquitard contributes with regional recharge through
leakage to the underlying alluvial-pyroclastic regional aquifer, from which about
50 m3/sec are pumped to supply water for the Metropolitan Area of Mexico City,
to nearly 25 million inhabitants. The extensive lacustrine Chalco Plain in the
southeastern part of Mexico City is underlined by an aquitard up to 300 m thick.
This area was a shallow lake until the 1940s when it was drained for agricultural
use and human habitation. In the middle 1980s, when major groundwater
extraction began with the onset of the Santa Catarina Well Field in the middle of
the Chalco Plain, subsidence of 2 m occurred between 1984 and 1989, causing a
shallow lake to form and gradually expand. Land subsidence in the central part
of the Chalco Basin has increased to 0.4 m/yr since 1984 and by 1991 total
subsidence had reached more than 10 m at present. The rapid land subsidence in
this area is causing the accumulation of meteoric waters during the rainy season
resulting in extensive flooding of farmland and potentially urban areas. Three
different stages of lake growth were identified: (i) ponding (1984-1989), (ii) lake
formation (1990-1993), and (iii) bifurcation (1994-present). In this last stage, the
local presence of a basalt layer in the middle of the Santa Catarina Well Field is
causing differential consolidation and forming a local topographic elevation that
permits the development of two separated lakes. Since the middle of the 1990s,
the new lake conditions are permitting the growth of few aquatic native plants
and arrival of different species of migratory birds: therefore, environmental
planning and management for the southernmost end of the Basin of Mexico is
proposed through the development of a series of connected artificial lakes to
control urban growth, to improve environmental conditions of the MAMC, to
develop ancient agricultural production systems such as the floating gardens or
“Chinampas” and recreational facilities with proportional benefits to the social
and economical sectors.
Keywords: land subsidence, consolidation, artifitial lake development,
environmental planning.
doi:10.2495/GEO060211
1 Introduction
Most of Mexico City is constructed over this highly compressible aquitard within
the Basin of Mexico and is surrounded by volcanic mountains (Figure 1). The
Basin is hydrologically closed and as such, a quiescent series of interconnected
lakes occupied the basin floor prior to the construction of an artificial drainage
system in 1789 (Bribiesca [1]). Groundwater is extracted from a regional
alluvial-pyroclastic aquifer and from a fractured volcanic aquifer for supplying
almost 25 million inhabitants and some 30% of the nation’s industry in the
Metropolitan Area of Mexico City (MAMC) INEGI and INE [1]. Both aquifers
underlie a thick sequence (between 50-300 m) of clayey lacustrine sediments,
considered in hydrogeological terms as a regional aquitard Ortega and Farvolden
[2]. Approximately 52 m3/s of groundwater has been extracted from these
aquifers, out of a total of 65 m3/s consumed in the MAMC [1] and NRC-AIC-
ANI [3].
Figure 1: Location of the Chalco Basin within the Basin of Mexico and
former lakes.
Groundwater extraction from the regional alluvial-pyroclastic aquifer

underneath the lacustrine aquitard started in the XIX century and became
extensive during the 1930s Hiriart and Marsal [4]. Extensive land subsidence
was documented since the early 1940s in downtown Mexico City over several
decades due to depressurization and consolidation of the aquitard caused by
heavy pumping, particularly in the middle of the city Carrillo [5]. Indeed, the
magnitude of the subsidence became so extreme in some locations that severe
problems arose related to building foundations, sewer drainage, and subway
transportation system.
To minimize further damage from additional subsidence within the MAMC

many of the downtown wells were taken out of production and new well fields
were constructed on the outskirts of the city and groundwater from the basin was
also imported to satisfy water demands. In the southern part of the Basin of
Mexico, the most recent well field to be added to the urban supply system is
located within the Chalco Basin (Figure 3), where one of the ancient lakes once
existed (Figure 1). This new well development has resulted in the partial transfer
of the land subsidence problem from within the MAMC to the nearly agricultural
and urban areas of the Chalco Plain.
The main goal of this work is to update the land elevation survey up to 2005
and to present hydrological evolution of the development of the New Chalco
Lake and environmental conditions recorded since 1988 to 2005. A plan for
environmental improvement is suggested throughout the development of a series
of artificial lakes, in the southern end of the Basin, by controlled groundwater
extraction from the regional aquifer, with different environmental, social and
economic impacts.
2 Hydrogeology of the Chalco Basin

The Chalco Plain was a shallow lake until the 1940s when it was drained for
agricultural use and human habitation. Historic information indicates that the
Chalco Plain was an area of fresh ground-water discharge prior to the onset of
heavy groundwater extraction from the semiconfined aquifer, in contrast with the
former Texcoco Lake that was recognized since the Aztecs by its high salinity
content.
The thickness of the lacustrine aquitard increases from the edges of the
former lake to a maximum thickness of about 300 m in the centre of the plain
(Figures 2 and 3) Ortega-Guerrero et al. [6]. The lacustrine sediments are
interbedded with Quaternary basalts and pyroclastic deposits near Sierra
Chichinautzin and in the vicinity of the Santa Catarina Well Field (SCWF). Of
particular importance, for this work, is the basalt interbedded with the lacustrine
sequence in the middle of the SCWF, which has influenced differential land
subsidence and that will be discussed below. A regional granular aquifer
underlies the lacustrine aquitard; consisting primarily of alluvial and pyroclastic
material and ranging in thickness between 200 and 400 m. Tertiary volcanic
bedrock is below the granular aquifer (Figures 2 and 3).
Detailed investigations of transient groundwater flow and land subsidence in
the Chalco Basin reveal sever exploitation of the regional aquifer system,
resulting in progressive lowering of piezometric levels in the aquifer and a fast
decline in land surface. Aquifer depressurization and land subsidence began
throughout the Chalco Basin prior to the construction of the SCWF in the middle
of the plain in 1984. Since pumping started from the SCWF the regional
piezometric surface in the aquifer has been drooping at a rate of about 1.5
m/year. The present rate of land subsidence in the middle of the clay plain is near
0.40 m/year, the highest within the Basin of Mexico. Consequently,
accumulation of rain and surface water is occurring in this topographic

depression Ortega-Guerro et al. [7].
Figure 2: Location of the Santa Catarina Well Field and hydrogeologic cross-
section (Modified from Ortega-Guerrero et al. [6]).
Figure 3: Hydrogeologic cross-section showing the distribution of the

lacustrine clayey aquitard and the interbedded basalts (Modified from
Ortega-Guerrero et al. [6]).
Land subsidence has caused elevation gradients to change, such a degree, that
part of the surface water drainage network is not longer functional and a series of
pumps are used to drain surface water to the nearby Texcoco area (Figure 1).
Consequently, more of the land surface is progressively inundated, particularly
during the rainy season and flooding has not only taken agricultural land out of
production but has also begun to threatened urban areas that are expanding from
the mountain slopes and edges of the former lakes onto the plain. In 2000,
extraordinary rain events caused the northeastern urban area of Valle de Chalco,
to become flooded with a mixture of rainwater and industrial-urban wastewater,
affecting the health of people and their property.
Numerical simulations of land subsidence based on numerous observation
piezometer sites carried out by Ortega-Guerrero et al. [6], suggest that under
current pumping rates, total land subsidence in the area of thickest lacustrine
sediment will reach 15 m by the year 2010 (Figure 4). If pumping were reduced
to the extent that further decline in the potentiometric surface is prevented, total
maximum subsidence would be significantly less, about 10 m, and the rate would
nearly cease by 2010 (Figure 4). This analysis demonstrates that the new shallow
lake conditions will prevail for several decades with enormous environmental
effects. How can this hydrologic, landscape, ecologic and environmental
situation can be controlled for environmental benefits is the topic of this work.
Figure 4: Predicted transient evolution of the total subsidence in the middle of

the Chalco Plain if the rate of drawdown is reduced by, 0% (case 1),
25% (case 2), 50% (case 3), 75% (case 4) and 100% (case 5).
(Modified from Ortega-Guerrero et al. [6]).
Of concern of this region therefore is the total magnitude of the land

subsidence and the future subsidence rate that is likely to result from the current
groundwater extraction system. There are also multiple geo-environment
concerns at the Chalco Basin particularly on: the new hydrological and landscape
ecology, environmental planning and management based on sustainable

groundwater resources exploitation, flooding hazards, between others.
The information published in Ortega-Guerro et al. [7] and Ortega-Guerrero et
al. [6] was augmented with new data such as land elevation surveys, evolution of
hydraulic response of the aquitard and evolution of lake growing, identification
of environmental modifications derived from the process of pumping-land
subsidence, to provide elements for environmental planning and management.
3.1 Evolution of land subsidence and lake growth
Based on physical observations of the extension of the new Chalco Lake with
time, three different periods of lake evolution are identified: (1) Ponding stage,
(2) New lake stage, and (3) Bifurcation of the lake.
(1) Ponding stage. In 1987 a series of nests of drive point piezometers

where installed in the southern edge of the Chalco Plain to measure hydraulic
heads in the upper 20 meters of the lacustrine sequence. In 1988 this site and
other isolated sites became flooded as shown in Figure 5. Those sites were
located at particular area that had a total drop in land surface elevation of 2 m in
the period 1984 to 1989. The network of surface water canals to drain surface
water became no functional, and in order to avoid flooding to the nearby urban
areas on the Chalco Plain, a system of electric pumps were installed. The Xico
ponds are due to the load of the Xico Volcano and existed before operation of the
SCWF.
(2) New Lake stage. Progressive consolidation of the aquitard due to

aquifer pumping in the SCWF caused a total land subsidence of about 7 m in the
central part of the plain. In August 1991 the lake extended to an area of about 4
km2 (400 ha) as shown in Figure 5. The roads parallel to the SCWF were
progressively elevated to provide access to the wells.
(3) Bifurcation of the lake. In 1994 a second lake began to form in the
northern part of the Chalco Plain. The bifurcation is associated with the presence
of a basalt unit interbedded with the lacustrine sequence in that part of the plain
(Figure 3). This figure also shows the position of the basalt unit and the present
position of the two lakes. This basaltic layer is causing differential consolidation
of the aquitard and potentially will maintain the lakes separated for some period
of time until a lateral connection may be reached. The extension of the lakes in
2005 is also shown in Figure 5, where development of two individual lakes is
observed, covering a total area of about 6 km2 (600 ha).
Figure 5: Evolution of the shallow lakes. The bifurcation of the lake in 2005 is
controlled by a basalt unit interbedded with the lacustrine sequence in
the middle of the Santa Catarina Well Field Modified from Ortega-
Guerrero et al. [7].
3.2 Environmental situation
In the first stage of the New Chalco Lake formation, land subsidence in the
middle of the Chalco Plain caused elevation gradients to change and part of the
surface water drainage network was not longer functional and the wastewater
from urban areas and industry, without previous treatment, accumulated towards
this shallow lake, causing sever problems in water quality with total absence of
wild life. However, as land subsidence increased and extraordinary rainwater
events flooded the topographic depression, in the 1990s, more favorable water
quality conditions developed, permitting the growth of few aquatic native plants
and arrival of different species of migratory birds. This unexpected situation may
open new possibilities for environmental planning and management in the Basin
of Mexico.
Water balance at the New Chalco Lake is not well understood, it captures
between 7E+06 to 10E+06 m3/yr of rainwater; real evaporation and
contributions from sewage are not well known. Groundwater-surface water
interaction and the new climatological conditions in addtion to chemical and
isotopic characterization of the new lake are the focus of present research.
3.3 Developing a series of connected lakes in the southern end of the Basin
of Mexico
There are different well fields in the southern lacustrine area of Mexico City. A
well-planned and controlled extraction of groundwater from the underlying
alluvial-pyroclastic and basaltic aquifers can, progressively, generate adequate
amounts of consolidation of the overlying compressible aquitard, to develop a
series of topographic depressions (Figure 6) that can be filled with rainwater,
runoff and perhaps some treated water. The wells would continue to provide
drinking water to the City while causing progressive consolidation.
Groundwater extraction must be designed to progressively reduce hydraulic
conductivity at the bottom of the aquitard as demonstrated by Rudolph and
Frind [8] and avoid an excess of stress that may cause fracturing of the aquitard
as demonstrated by Aguilar-Perez et al. [9]. Fracturing of the aquitard may cause
important leaks.
The area proposed for lake development is shown in Figure 6. The basic idea
is to connect the Xico ponds, with the New Chalco Lake, Mixquic, Tlahuac and
Xochimilco. The Mixquic and the Tlahuac areas require further consolidation
development to connect them with the other lakes.
In the Texcoco former lake, the Nabor Carrillo Lake was developed, in the
1960s, by pumping a 2-3 m thick sandy layer located about 20 m below a clayey
aquitard to form a topographic depression, which was filled with water. This
method required a large number of shallow wells that produced high salinity
content in the water, which localy was used to produce caustic soda Marsal [10].
However cost benefit analysis may be needed to choose the method for
consolidating the aquitard.
Figure 6: Proposed development of a series of connected lakes towards the

southern end of the Basin of Mexico to improve environmental
conditions.
This development of a series of connected lakes by controlled consolidation

would progressively improve environmental conditions and ultimately improve
social and economic situation of the inhabitants of this region of the Basin
through payment of environmental services and ecotourism. Food production
using the system of Floating Gardens or “Chinampas” from the Aztec tradition
and still use in Xochimilco and Mixquic may represent an additional attraction of
this project.
4 Conclusions
The extensive lacustrine Chalco Plain in the southeastern part of Mexico City is
underlined by an aquitard up to 300 m thick. Land subsidence in the central part
of the Chalco Basin has increased to 0.4 m/yr since 1984 and by 1991 total
subsidence had reached 8 m and more than 10 m at present. The rapid land
subsidence in this area is causing the accumulation of meteoric waters during the
rainy season resulting in extensive flooding of farmland and potentially urban
areas. Three different stages of lake growth were identified: (i) ponding (1984-
1989), (ii) lake formation (1990-1993), and (iii) bifurcation (1994-present). In
this last stage, the local presence of a basalt layer in the middle of the Santa
Catarina Well Field is causing differential consolidation and forming a local
topographic elevation that permits the development of two separated lakes. This
area is still susceptible to the highest potential land subsidence effects as a result
of groundwater extraction of anywhere in the basin. Environmental planning and
management for the southernmost end of the Basin of Mexico is proposed
through the development of a series of connected artificial lakes to control urban
growth, to improve environmental conditions of the MAMC, to develop ancient
agricultural production systems such as the floating gardens or “Chinampas” and
recreational facilities with benefits to social and economical sectors.
References
[1] Bribiesca, C.J.L. Hidrología Histórica del Valle de México [Historic

hydrology of the Valley of Mexico]. Ingeniería Hidráulica, Vol. XIV, No
1, México . pp. 107-125, 1960.
[2] Ortega, G.M.A. and Farvolden, R.N. Computer analysis of regional
groundwater flow and boundary conditions in the Basin of Mexico.
Journal of Hydrology 110, pp. 271-294, 1987
[3] NRC-AIC-ANI. National Research Council, Academia de la Investigación
Científica and Academia Nacional de Ingeniería. México City´s water
supply. National Academic Press, Washington DC, 1995.
[4] Hiriart, F. and Marsal, R.J. The subsidence of México City. In volumen
Nabor Carrillo, Secretaría de Hacienda y Crédito Público. México, Vol II,
pp. 109-147, 1969.
[5] Carrillo, N. Influence of Artesian Wells in the Sinking of Mexico City.
Comisión Impulsora y Coordinadora de la Investigación Científica,
Anuario 47. In: Volumen Nabor Carrillo, pp. 7-14, Secretaría de Hacienda
y Crédito Público, México, 1969.
[6] Ortega-Guerrero, M.A., Rudolph, D.L. and Cherry, J.A. Análisis of long-
term land subsidence near México City: Field investigations and
predictive modeling. Water Resources Research, 25(11), pp. 3327-3341,
1999.
[7] Ortega-Guerrero, M.A., Cherry, J.A. and Rudolph, D.L. Large-scale
aquitard consolidation near Mexico City. Ground Water, 31(5), pp. 707-
718, 1993.
[8] Rudolph, D.L. and Frind, E.O. Hydraulic response of hyghly compressible
aquitards during consolidation. Water Resources Research, 27(1), pp. 17-
30, 1991.
[9] Aguilar-Perez, L.A., Ortega-Guerrero, M.A., Hubp, J.L. and Ortiz, D.C.
Análisis nemérico acoplado de los desplazamientos verticales y
generación de fracturas por extracción de agua subterránea en las
proximidades de la Ciudad de México. Revista Mexicana de Ciencias
Geológicas. In Press, 2006.
[10] Marsal, R.J. Development of a lake by pumping-induced consolidation of
soft clays. In volumen Nabor Carrillo, Secretaría de Hacienda y Crédito
Público. México, Vol II, pp. 229-266, 1969.
Stability of slopes of municipal solid waste

landfills with co-disposal of biosolids
M. Chopra, D. Reinhart, M. Vajirkar & B. Koodhathinkal
University of Central Florida, Orlando, Florida, USA
Abstract
This paper deals with the impact of the addition of biosolids on the geotechnical
properties of class I landfills with Municipal Solid Waste (MSW). The properties
are estimated from a field exploration program using the Cone Penetration Test
(CPT). The geotechnical shear strength parameters (angle of internal friction and
cohesion) of MSW and biosolids mixture are estimated by correlation with the
CPT results. These parameters are then used to study the impact of the addition
of biosolids on the stability of the slope of the landfill. The slope stability
analysis is conducted on various landfill models using the software SLOPE/W.
The waste is assumed to act similar to a cohesionless soil. Based on the field
explorations, the angle of internal friction was found to be about 29°. Previous
research indicated that the most suitable approach to introducing biosolids into
the landfill was in the form of trenches. From the slope stability analysis, it was
found that the factor of safety reduces significantly with the introduction of
biosolids due to a reduction in the shear strength and with increase in the overall
moisture content.
Keywords: landfills, biosolids, slope stability analysis, cone penetrometer.
1 Introduction
Landfills are well engineered facilities that are used for disposal of Municipal
Solid Waste (MSW) and are located, designed, monitored, operated and financed
to ensure compliance with federal regulations. Disposal of water treatment
facility sludge and wastewater treatment facility biosolids presents significant
challenges to facility operators, as this practice often violates loading of metals
and exceeds the allowable levels of pathogens for land application. Thus,
doi:10.2495/GEO060221
landfilling of the biosolids is promising as it results in minimal contact with

humans and can provide moisture in the landfill for it to act as a bioreactor.
With the addition of sludge and biosolids, significant changes in the composition
and characteristics of landfill may take place. The stability of waste slopes is
particularly important and depends on the strength properties of the MSW with
biosolids. The complexity of the problem is further increased by the
heterogeneous nature of waste, placement conditions and level of decay of
various constituents of the landfills. Proper analysis and design of the landfills
have pushed the boundaries of geotechnical engineering practice, in terms of
proper identification and assessment of strength and deformation characteristics
of waste materials (Sharma et al. [1]; Sadek et al. [2]).
Typical laboratory approaches to measure the geotechnical parameters of the
waste such as density, moisture content, cohesion, and shear strength have
limitations due the heterogonous nature of the MSW. The scarcity of data is due
to the difficulties in sampling and testing the refuse. This difficulty is further
compounded by the fact that refuse composition and properties are likely to
change within a landfill and the waste is likely to decompose with time. An
alternate method for estimating landfill geotechnical properties is using in-situ
testing devices that provide accuracy with minimal efforts and costs. In the field
of geotechnical engineering, a device that meets these criteria for soil testing is
the Cone Penetrometer. The unit weight, moisture content, friction angle and
cohesion influence the stability of landfill side slopes and interfaces among
landfill components. True cohesion between particles is unlikely in landfills but
there may be significant cohesion that results from interlocking and overlapping
of the landfilled constituents (Singh and Murphy [3]).
This paper aims to evaluate the performance of slopes in landfills that may
allow the co-disposal of sludges and biosolids, using standard slope stability
analyses. Field tests using a Cone Penetrometer are carried out, followed by a
slope stability analysis on landfill models using the computer software
SLOPE/W (Geo-Slope [4]).
2 Field site description

The field study was conducted at the Highlands County Landfill, located in
Highlands County, Sebring, Florida. Two test MSW pads, namely Pilot Area
(PA) and Control Area (CA) were constructed at this location. Each test pad was
about 120 by 100 feet in plan dimensions at it base and 40 by 60 feet at the top.
The height of each test pad was about 10 feet. The compacted waste below the
test pads had an average thickness of about 13.2 feet. Liquid biosolids from the
City of Sebring Wastewater Treatment Plant were added to the waste in the PA.
Two tanker loads of unstabilized biosolids in the form of sludge were transported
from the City of Sebring Wastewater Treatment Plant to the site. The liquid
sludge had a solids content of 23 g/l or about 3%. The sludge volume was about
1,202 cubic feet which occupied about six inches of depth on an average. The
measured decrease in the surface elevation of the impounded sludge due to
infiltration of sludge moisture into the underlying and surrounding MSW over 48
hours was about 0.8 inches on an average. Based on the rate of evaporation for
the site area for the month of May, approximately 0.45 inches of the measured
decrease in sludge level over 48 hours could be attributed to evaporation losses
from the surface. The resulting average thickness of the remaining sludge in the
entire impoundment was about 4.75 inches.
The bulldozer operator had no problem operating the machine over the top of
the sludge. The bearing capacity for the wide track, low ground pressure
bulldozer (36-inch wide tracks) was excellent with no machine tilting or soft
spots were encountered. The three-foot thick layer of MSW was compacted into
a layer about 1.5 feet thick, without any observed extrusion of sludge to the
surface. No soft spots were encountered and no pumping action was observed as
the compactor moved back and forth. Additional MSW was then placed, spread,
and compacted on the pad. All four side slopes were constructed at a final grade
of about 1:3. The entire surface of the test pad was covered with 24 inches of
loosely spread intermediate cover soil. The Control Area (CA) was prepared in a
similar manner without the addition of sludge.
A total of 28 cone penetration tests were then conducted at these areas, five at
CA and twenty three at PA. CPT locations were taken as close to edge of the
slope as the truck could be placed.
3 Methodology
3.1 Cone penetration tests
The CPT utilizes electrical transducers rather than analog gauges to obtain a
nearly continuous profile of point (tip) resistance, sleeve friction and pore
pressure with depth. Specialized data acquisition hardware and software was
used to record readings from the transducers at a frequency of approximately five
readings per second. Electrical signals reading were then converted to
engineering units of stress using device-specific calibration factors. Immediately
after the sounding was completed, the borehole was filled with bentonite clay to
the top of the intermediate soil cover to prevent water and gas transport from the
landfill.
The desired depth of soundings in pilot area and control area was about 20
feet from the intermediate soil cover. Ten soundings had to be terminated at
depths shallower than the intended depth due to buried obstructions encountered
in the path of the cone in the test areas. The data was stored and used in a
subsequent statistical data reduction analysis to determine the geotechnical
properties for use in subsequent slope stability modeling. A frequency
distribution curve of the values of tip resistance was created for each sounding.
In order to eliminate outliers in the tip resistance values, the cut off frequency
values was selected as 20. Unrealistic tip resistance and pore pressure values are
possibly due to the heterogeneous material and obstructions.
Based on the field results for tip resistance and sleeve friction, an attempt was
made to estimate the location and potential effect of the biosolids layers within
the MSW mass. However, this layer could not be identified explicitly, as the
biosolids added had only about 3% solids content; it is likely that the wet
material got very well mixed with the MSW.
The CPT data provides a repeatable index of the aggregate behavior of the in
situ soil in the immediate area of the probe (Robertson [5]). Based on the
average values of the MSW parameters, the landfill material is seen to behave
similar to “Coarse Grained Sandy-Silty Soil”. Since the material is classified as
sand, charts proposed by Robertson and Campanella [6] may be used to estimate
the friction angle for the MSW. The resulting friction angle values were adjusted
proportionally for variation in density. This is a conservative approach and the
actual friction angle (φ) may be somewhat higher. From the generalized
equation for friction angle, the average value of friction angle was found to be
29º.
3.2 Slope stability analysis
The behavior of MSW is assumed to be frictional in nature and is governed by

the Mohr-Coulomb criteria. The slope stability analysis was conducted using the
software SLOPE/W. Daily or intermediate cover was not taken into account in
the slope stability analysis. The biosolids or sludges were placed in trenches.
The geotechnical properties needed for modeling the slope stability were
determined in the field using CPT testing as discussed in the previous section.
The landfills were analyzed under circular failure and block failure conditions,
and the factors of safety against these modes of failure were evaluated. This
paper only presents the results of the study with circular failure. A factor of
safety is an index indicating the relative stability of a slope. It does not imply the
actual risk level of the slope due the variability of input parameters. With
probabilistic analysis, two useful indices are available to quantify the stability or
the risk level of a slope. These two indices are known as the probability of
failure and the reliability index. The input parameters required for the Mohr-
Coulomb Failure Model are unit weight, angle of internal friction and cohesion
with associated standard deviation (SD) values for each parameter.
Table 1 shows the properties of each material used as input in the model. In
Table 1, MSW+BS1, MSW+BS2 and MSW+BS3 refer to layers of different
moisture content, which indicate the gradual migration of the biosolids’ moisture
into the MSW. This is reflected in the gradation of shear strength parameters
with addition of biosolids to the MSW. The shear strength parameters were
estimated from the field data obtained using the CPT.
For landfills modeled using the circular failure model, the radius of failure
plane and the center of the circular failure plane are required to be defined. The
radius of the failure plane was specified such that it covers a large range of radii,
while the grid for the center of the circular failure plane was chosen so that the
minimum factor of safety lies within the grid. Once the landfill model is
developed and the grid and radii for the failure plane are specified, slope stability
analysis is carried out. A typical result from the program is shown in Figure 1
indicating the failure plane and the minimum factor of safety for the chosen grid.
Table 1: Landfill material properties used in modelling.
Material Unit Weight Cohesion Friction Angle

Source
Description (pcf) (psf) (degrees)
MSW 70 (SD=5) 0 29 (SD=5) Field CPT
BS 35 (SD=5) 0 0 Ref [7]
Extrapolated
MSW + BS1 70 (SD=5) 0 7 (SD=5)
Results
Extrapolated
MSW + BS2 70 (SD=5) 0 15 (SD=5)
Results
Extrapolated
MSW + BS3 70 (SD=5) 0 22 (SD=5)
Results
Sand Liner 30 (SD=3) 0 40 (SD=5 Ref [8]
Figure 1: Typical slope stability analysis - landfill of slope 1:3 with MSW only
– circular failure.
Next, a landfill model was generated incorporating the placement of biosolids

in trenches as shown in Figure 2. The width of the trench was about 2-2.5 feet
and spacing between the trenches was about fifteen feet. The depth of the
trenches was about six feet. Biosolids were added in the trenches. The trenches
were then filled with regular MSW and compacted with regular effort. The
compaction effort needed to compact the waste layer was reduced based on
discussions with the landfill operators. Slope stability analysis was then
conducted on landfill models with side slope varying from 1:2 to 1:4. Although
side slope 1:2 is not permitted in Florida state regulation, it is studied as an
extreme condition.
Figure 2: Typical landfill profile with biosolids placed in trenches.
4 Results
The results are reported in two parts - results from the field tests and those from
the subsequent slope stability study. Field tests conducted consisted of Cone
Penetration Test on areas receiving biosolids, areas without biosolids. The
average tip resistances range from 60 to 100 tsf, the average sleeve resistance is
in the range of 0.7 to 2.4 tsf. These are similar to the range of tip resistance and
sleeve resistance values from previous literature on CPT used in landfills. The
region of solid waste is characterized by a highly variable set of test results.
The CPT results indicated that the cone frequently encountered stiff objects,
which produced sharp peaks in the tip resistance measurements resulting in
highly variable readings. However, a trend of increasing lower bound tip
resistance with depth, was apparent in most the tests. It is important, therefore,
to reduce these data by eliminating unrealistic outliers based on cone tip
resistance and pore pressure values. From the CPT probes, daily or interim
cover soil could not be distinguished from the refuse.
Using the method proposed by Robertson and Campanella [6], it is possible to
derive the profile of friction angle as a function of depth from the piezocone
penetration data for the landfill. The average value of friction angle is 29º. This
value of friction angle (φ) is used for slope stability analyses.
Slope stability analysis was conducted using commercially available software

SLOPE/W on models of landfills with and without biosolids. Biosolids were
placed in trenches and different values of side slope such as 1:2, 1:3 and 1:4
were modeled. Three scenarios are considered.
1. Landfill with MSW only.
2. Landfill with MSW and biosolids which were placed in trenches.
3. Landfill with MSW and biosolids in trenches taken close to the edge of
the slopes.
Table 2 presents the results of the factor of safety (FOS) of landfill slopes for
the different conditions that are modeled for the circular failure mode.
Table 2: Factor of safety for landfills with different side slopes.
Side MSW only MSW and Biosolids MSW and

Slope Biosolids with
trenches close to
side slope
FOS Reliability FOS Reliability FOS Reliability
Index Index Index
1:2 1.94 16.3 1.72 19.0 1.26 7.0
1:3 2.51 19.3 2.17 17.4 1.28 6.4
1:4 2.96 21.8 2.31 3.4 1.75 11.1
5 Discussion
Flattening the slope not only reduces the sum of the driving forces, but also tends
to force the failure surface deeper into the ground. The change in length of the
failure surface increases in the resisting forces because the shear strength is
distributed over a wider area, thereby enhancing stability. Results for FOS
suggest that the stability of the landfill slope has been reduced with the addition
of biosolids.
The factor of safety is clearly reduced if the biosolids trenches are placed
close to the edges of the slopes. This scenario allows a weak plane to develop
close to the side slope and encourages the failure plane to pass through this weak
layer. This situation is considered unstable and needs to be avoided in the field.
These results are similar to the previous work by Koodhathinkal [7] who tested
MSW with biosolids in a laboratory study. The lower values may be attributed to
a lower friction angle for biosolids that are placed as discrete layers and may not
be completely mixed in with the MSW.
Disposing of biosolids in trenches is a feasible solution from both slope
stability point of consideration and ease of field application practice. However,
trenches should not be close to the edge side of the landfill as the factor of safety
reduces significantly.
Berms at the toe of slopes contribute to resisting forces and as a result there is
an increase in the factor of safety when considering local failures.
6 Conclusions
Based on the field explorations, the angle of internal friction was found to be
about 29°. The most suitable approach to introducing biosolids into the landfill
was in the form of trenches. From the slope stability analysis, it was found that
the factor of safety reduces significantly with the introduction of biosolids due to
a reduction in the shear strength and with increase in the overall moisture
content.
References
[1] Sharma, H. D., Dukes, M.T. and Olsen, D.M. (1990). “Field Measurement
of Dynamic Moduli and Poisson's Ratios of Refuse and Underlying Soils at
a Landfill Site.” Geotechnics of Waste Fills - Theory and Practice ASTM
STP 1070.
[2] Sadek, S., Abou-Ibrahim, A., Manasseh, C. and El-Fadel, M. (2001).
“Stability of Solid Waste Sea Fills: Shear Strength Determination and
Sensitivity Analysis.” International Journal of Environmental Studies 58:
217-234.
[3] Singh, S., and Murphy Bruce (1990). “Evaluation of the Stability of Sanitary
Landfills.” Geotechnics of Waste Fills - Theory and Practice.
[4] Geo-Slope (2001). SLOPE/W for Slope Stability Analysis Version 5, Geo-
Slope.
[5] Robertson, P. K. (1990). “Soil classification using the Cone Penetration
Test.” Canadian Geotechnical Journal (27): 151-158.
[6] Robertson, P.K. and Campanella, R.G. (1983). “Interpretation of Cone
Penetration Tests, Part I: Sands,” Canadian Geotechnical Journal, Ottawa,
Vol. 20, No. 4, pp. 718-733.
[7] Koodhathinkal, B. R. (2003). Stability of Slopes in a Class I Landfills with
Co-Disposal of Sludges and Biosolids. MS Thesis, University Of Central
Florida.
[8] Das, B. M. (2000). Principles of Geotechnical Engineering, Brooks/Cole
Publishers.
Section 6
Geoecology
Net photosynthetic O2 evolution and calcium

precipitation in Chlamydomonas reinhardtii
Y. Y. Wu1, P. P. Li1, B. L. Wang2 & C. Q. Liu2
1
Institute of the Agricultural Engineering, Jiangsu University, Zhenjiang,
Jiangsu, People’s Republic of China
2
The State Key Laboratory of Environmental Geochemistry,
Institute of Geochemistry, The Chinese Academy of Sciences, Guiyang,
People’s Republic of China
Abstract
The relationship between carbonic anhydrase activity, net photosynthetic O2
evolution rate and calcium precipitation in Chlamydomonas reinhardtii was
studied. By adding bovine carbonic anhydrase and inhibitor of carbonic
anhydrase-acetazolamide to SE medium to alter the activity of external carbonic
anhydrase, the variation of net photosynthetic O2 evolution rate and calcium
precipitation in Chlamydomonas reinhardtii was determined. Calcium
precipitation of Chlamydomonas reinhardtii was determined by the difference of
algal calcium content during the culture period. The result shows that net
photosynthetic O2 evolution rate and calcium precipitation in Chlamydomonas
reinhardtii were enhanced when bovine carbonic anhydrase was added to
medium. Acetazolamide can inhibit calcium precipitation and photosynthetic O2
evolution. Carbonic anhydrase may be an important factor to accelerate algal
calcium precipitation. Moreover, the geological significance of carbonic
anhydrase in algal calcium precipitations and the relationship between algal
calcium precipitation and biomineralization of carbonate precipitation are
discussed.
Keywords: calcium precipitation, carbonic anhydrase, Chlamydomonas
reinhardtii, net photosynthetic O2 evolution rate.
1 Introduction
The biogeochemical cycle, which regulates all the biochemical reactions, is the
important “driving force” of the geo-environment. The carbon cycle is one of the
doi:10.2495/GEO060231
most important element cycles. CO2 is the important link and ligament of global
carbon cycle [1]. The photosynthetic fixation of CO2 and precipitation of CaCO3
are closely connected both spatially and temporally. Photosynthetic carbon
assimilation is commonly thought as the cause of calcium carbonate precipitation
in algae [2]. External HCO3- is inorganic carbon resource of calcification in the
algal cells [3, 4]. Carbonic anhydrase (Carbonic anhydrase, CA; carbonate
hydrolyase, EC 4.2.1.1) is a ubiquitous enzyme catalyzing the reversible
conversion of CO2 to bicarbonate. The ability to use HCO3- for photosynthesis
has been associated with the activity of external CA [5, 6]. Hence, there are the
putative relationship between CA, the photosynthetic fixation of CO2 and
calcification in algae. This work is intended to investigate the effect on
calcification and photosynthesis by CA, and the significance of CA in
biomineralization.
2.1 Culture growth
Chlamydomonas reinhardti, obtained from the Institute of Hydrobiology,

Chinese Academy of Sciences, was grown axenically in continuously stirred
artificial freshwater medium SE, which was provided by the Freshwater Algae
Culture Collection (http://www.ctcccas.ac.cn). The flasks were closed with
bacteriological cotton plugs. Two-day preculture was conducted in medium SE
free of calcium (NaCl 20 mgl-1 instead of 20 mgl-1 CaCl2⋅2H2O) before
treatment. It was harvested by means of centrifugation(1,700×g for 10 min);
The pellet was suspended in medium SE. Growth of culture was initiated by
introducing inoculum containing about 108 alga cells. Treatment 1: Cells were
cultured in the medium SE containing an inhibitor of extracellular carbonic
anhydrase--acetazolamide (AZ) (30 mM), at 25.0°C ±1.0°C, 150 µmol m-2 s-1
light on a 16-h/8-h day/night cycle. The control culture was in the same
condition except that AZ was omitted and distilled water was used. Treatment 2:
Cells were cultured in the medium SE containing bovine carbonic anhydrase
(BCA) (10µg ml-1) (Sigma C2522), at 25.0°C±1.0°C, 150 µmol m-2 s-1 light on
a 16-h/8-h day/night cycle. The control culture was the same above. After 6, 12,
and 24 hours, the packed cell volume, the algal content of calcium, and the net
photosynthetic O2 evolution rate were determined, respectively.
2.2 Determination of packed cell volume
Packed cell volume was determined by centrifugation of 5 mL of cell suspension

in hematocrit tubes for 15 min at 2,200×g.
2.3 Assay for the algal content of calcium
100 ml algal solution is filtered through analytical filter paper (Whatman), the
filter paper with algal cells is dried at 80oC. The dried sample is ashed at 500oC.
The ash is dissolved in a nitric acid solution (50 ml final volume), and analyzed
on an ICP spectrometer for calcium.
2.4 Measurements of photosynthetic O2 evolution
The net photosynthetic O2 evolution rate was measured with a Clark-type oxygen
electrode (YSI-5300, USA) under a photon flux density of 150µmol m-2 s-1 and a
temperature of 25°C. The effect of acetazolamide on the cells O2 evolution was
estimated after 6, 12, and 24 hours’ incubation in medium SE containing the
acetazolamide (30 mM). The effect of bovine carbonic anhydrase (BCA) on the
cells O2 evolution was estimated after 6, 12,and 24 hours’ incubation in medium
SE containing the BCA (10µg ml-1). Cells separated from the medium SE (by
centrifugation at 5000 × g for 5 min), were re-suspended in 2 ml of 25 mM
HEPES-KOH (pH 8.2) and transferred into an electrode chamber. Before
determining the dissolved inorganic carbon-dependent O2 evolution, cells were
allowed to photosynthesize to deplete possible intracellular pool of “CO2” until
no net O2 evolution was observed. Following the addition of the concentration
of 2 mM bicarbonate (final concentration), or the concentration of 2 mM
bicarbonate with 30 mM AZ (final concentration), or the concentration of 2 mM
bicarbonate containing 10 µg ml-1 the BCA (final concentration), the rate of O2
evolution was measured.
2.5 Statistics
Each treatment consisted of five replicates. The mean and standard deviation are
calculated for each treatment. One-way ANOVA and LSD tests are conducted
for each group.
3 Results
3.1 The effect on the algal content of calcium
After 6, 12, and 24 hours’ incubation in medium SE containing BCA, the algal
contents in calcium of Chlamydomonas reinhardtii are significantly higher than
those of the control at the same period. After 24 hours’ incubation in medium SE
containing AZ, the algal contents in calcium of Chlamydomonas reinhardtii are
significantly lower than those of the control at the same period (Table 1).
Table 1: The algal contents of calcium in Chlamydomonas reinhardtii with
time. Values are means ± SD of five replicates.
Time Control +BCA +AZ

Hours mg ml -1 PCV mg ml -1 PCV mg ml -1 PCV
6 7.44 ±0.67 10.36±1.12* 6.54±0.63
12 8.72 ±0.76 12.85±1.03* 7.25±0.75
24 10.36 ±0.83 14.78±1.34* 7.72±0.81*
*
The mean difference is significant between the culture containing BCA/AZ and
the control culture at the same period (P＜0.05).
3.2 The effect on the net photosynthetic O2 evolution rate
After 6, 12 and 24 hours’ incubation in medium SE containing BCA, the net

photosynthetic O2 evolution rate is significantly higher than that of the control at
the same period (Table 2). After 12 and 24 hours’ incubation in medium SE
containing AZ, the net photosynthetic O2 evolution rates are significantly lower
than those of the control at the same period (Table 2).
Table 2: The net photosynthetic O2 evolution rate of Chlamydomonas

reinhardtii with time. Values are means ± SD of five replicates.

Hours nmol O2 s-1 ml -1 nmol O2 s-1 ml -1 nmol O2 s-1 ml -1
PCV PCV PCV
6 15.45±1.74 26.35±3.25* 12.82±1.73
12 18.56±1.69 24.70±2.33* 10.87±1.52*
*
24 16.48±1.95 25.26±2.20 10.25±1.29*
*
The mean difference is significant between the culture containing BCA/AZ and
the control culture at the same period (P＜0.05).
3.3 The effect on the net photosynthetic O2 evolution rate
From tables 1 and 2, the ratio of the Ca precipitation quantity (CaPQ) to the net
photosynthetic O2 evolution rate (Pn) in Chlamydomonas reinhardtii can be
calculated (Table 3). During 6-12 hours, the ratio of Pn/CaPQ is smaller than that
during 12-24 hours. It shows that the Ca precipitation efficiency during 6-12
hours is greater than that during 12-24 hours.
Table 3: The ratio of the Ca precipitation quantity (CaPQ) to the net

photosynthetic O2 evolution rate (Pn) in Chlamydomonas
reinhartii.

Spaces Pn/ CaPQ Pn/ CaPQ Pn/ CaPQ
6-12 h 12.5:1 8.6:1 13.2:1
12-24 h 17.4;1 22.7:1 36.9:1
4 Discussion
Calcification is one of the most important biological processes in the living
world. Although, it has been proved that the calcification of corals provides CO2
source for its photosynthesis, the role of CA isn’t confirmed yet [7]. This
experiment confirmed the relation between the activity of CA and calcium
precipitation.
The difference of the algal content of calcium in Chlamydomonas reinhardtii
with time reflects the calcium precipitation and calcification. The inhibitor of
extracellular carbonic anhydrase—AZ decreases the calcium precipitation and

the net photosynthetic O2 evolution. Extraneous bovine carbonic anhydrase can
accelerate the calcium precipitation and the net photosynthetic O2 evolution. i.e.
external CA facilitates the calcium precipitation and the net photosynthetic O2
evolution.
Respiration Photosynthesis Plant
Ca
CO2 H2O Ca
CA
Water bodies
HCO3- CaCO3
Calcification
Figure 1: The sketch map regulation of photosynthesis and calcification by

CA.
The role of CA on the calcium ion precipitation in cell membrane and
carbonate transportation as well as biomineralization of calcium carbonate has a
vital significance. The activity of CA expresses the ability to regulate the carbon
cycle in the ecosystem. The carbon cycle in the ecosystem is related to the
transportation of calcium ion and carbonate. The variation of the ratio of the
calcium precipitation to the net photosynthetic O2 evolution is great. i.e. the
variation of rate of calcification is very great. It is similar to some other results
[8, 9]. From this experiment, we can conclude that the calcium carbonate
precipitation reacts rapidly at first, gradually slows down afterwards. It is
because no calcium carbonate crystals such as coccoliths can be produced not to
be capable of holding more calcium carbonate on the cell surface of
Chlamydomonas reinhardtii. The rates of absorption and the precipitation on
calcium were gradually decreased, even to zero. The calcium carbonate
precipitation reaction is known to produce CO2 according to the following
equation: Ca 2++2HCO3- →CaCO3 +CO2. Therefore, algal calcification exerts a
vital effect on calcium transportation and carbon cycle. CA influenced carbon
cycle by regulating the reaction: H2O+CO2↔HCO3-+H+[10]. Algal calcification
is also related to the concentration of HCO3-. Therefore, the relationship between
algal calcification, the net photosynthetic O2 evolution and the CA activity can be
found in aquatic ecosystem (Figure 1). From Figure 1, Calcium transportation
was led by carbon cycle; CO2 is the core of carbon cycle. Some
biomineralization reaction was led by calcium transportation. CA regulates

photosynthetic O2 evolution and calcification through the influence on the
concentration of inorganic carbon. The regulation of algal photosynthesis and
calcification by CA is, therefore, an important factor in the global carbon cycle
and biomineralization.
Acknowledgements
This work was supported by grant no. 40273038 of National Natural Science
Foundation of China, no. KZCX3-SW-140 of the Knowledge-Innovation
Program of Chinese Academy of Sciences, and no.2001-6-3 of Karst Dynamic
Laboratory, Ministry of Land and Resources, P.R.C.
References
[1] Schindler, D.W., Carbon cycling: The mysterious missing sink. Nature,
398, pp. 105–107, 1999.
[2] Arp, G., Reimer, A., Reitner, J., Photosynthesis-induced biofilm
calcification and calcium concentrations in Phanerozoic Oceans. Science,
292:pp1701-1704, 2001.
[3] Dong, L.F., Nimer, N.A., Okus, E. et al, Dissolved inorganic carbon
utilization in relation to calcite production in Emiliania huxleyi. New
Phytologist, 123:pp679-684, 1993.
[4] Skies, C.S., Roer, R.D., Wilbur, K.M. Photosynthesis and coccolith
formation: inorganic carbon sources and net inorganic reaction of
deposition. Limnology and Oceanography, 25:pp248-261, 1980.
[5] Haglund, K., Bjök, M., Ramazanov, Z. et al, Role of carbonic anhydrase
in photosynthesis and inorganic-carbon assimilation in the red alga
Gracilaria tenuistipitata. Planta, 187: pp275-281, 1992.
[6] Mercado, J.M., Niell, F.X., Figueroa, F.L., Regulation of the mechanism
for HCO3- use by the inorganic carbon level in Porphyra leucosticta Thur.
in Le Jolis (Rhodophyta). Planta, 201: pp319-325, 1996.
[7] Marshall, A.T. & Clode, P.L., Effect of increased calcium concentration
in sea water on calcification and photosynthesis in the scleractinian coral
Galaxea fascicularis. The Journal of Experimental Biology,
205:pp 2107-2113, 2002.
[8] Balch, W.M., Drapeau, D.T., Fritz, J.J., Monsoonal forcing of
calcification in the Arabian Sea. Deep-Sea Research II, 47:pp1301-1337,
2000.
[9] Sekino, K., Shiraiwa, Y., Accumulation and utilization of dissolved
inorganic carbon by a marine unicellular coccolithophorid Emiliania
huxleyi. Plant and Cell Physiology, 35:pp353–361, 1994.
[10] Wu, Y.Y., Li, P.P., Wang, B.L. et al, Significance of carbonic anhydrase
and its distribution in the Karst ecosystem. In: Tiezzi, E., Brebbia, C.A.,
Jørgensen, S.E. and Gomar, D.A. eds. Ecosystems and Sustainable
Development V. WIT Press, Southampton, pp.135-141, 2005.
Dynamics of suspended sediment concentration

and the impact on a lake-inhabiting bivalve
(Corbicula japonica) in the Abashiri River
basin, Hokkaido, northern Japan
S. Yanai1, Y. Nishihama2 & R. Tamura3
1
Hokkaido Institute of Technology, Japan
2
Tokyo University of Agriculture, Japan
3
Hokkaido Fish Hatchery, Japan
Abstract
The dynamics of suspended sediment and the impact on a lake-inhabiting bivalve
were studied in the Abashiri River basin, in eastern Hokkaido, northern Japan.
The Japanese bivalve: Corbicula japonica is the most important fishery resource
in Abashiri Lake. Torrential rain over 80 mm per day resulted in considerable
sediment concentration during the summer of 2003. Suspended sediment (SS)
concentrations >1600 mg/l were measured in the tributaries, where fragile rock
formations and faults had developed. The factors influencing SS concentration
were analyzed by multiple regression analysis. This led to the identification of
soft bedrock and fault densities as significant factors influencing sediment
concentration. Fine-sediment particles less than 0.063 mm in diameter accounted
for more than 80% of the samples at all stations. Since the lake bivalve prefers
coarse substrates, a negative relationship was observed between the percent of
fine sediment and the C. japonica distribution and fatness. The substrate around
the mouth of the river was also composed of fine sediment, which indicated that
the suspended sediment produced from the mountain tributaries has deteriorated
the lake environment, especially for C. japonica survival and growth.
Keywords: suspended sediment, Abashiri River, substrate, Corbicula japonica
and watershed management.
doi:10.2495/GEO060241
1 Introduction
Anthropogenic activities such as urbanization and agriculture alter water
infiltration rate and storage capacity and, hence, susceptibility to soil erosion.
These effects result in changes in run-off and sediment routing within watersheds
[16]. Hokkaido, the most northerly island in Japan, has been subject to increasing
human exploitation since the early 20th century. Land use has expanded
dramatically, with concomitantly increasing concern about environmental
degradation. The fine sediments produced in headwater areas flow downstream,
eventually reaching river mouths, which may damage numerous fishery-based
resources [11]. Despite increasing public awareness of the linkage between forest
and ocean, there have been few studies on the ecological impacts of agriculture
and deforestation on aquatic biota [10].
¯ Sapporo
Abashiri
!
Tokyo
!
0 75 150 300 450 600

Km
Figure 1: Location of the study watershed.
Here, we focus on a lake watershed (a relatively closed system in comparison

with coastal watersheds, in which the terrestrial impact is felt primarily by fish
populations). We first measured suspended sediment concentrations in rivers and
lakes over the whole watershed, and we discuss determinants of sediment
production. Subsequently, we observed the distribution and fatness of a lake-
inhabiting bivalve. We consider bivalves to be good indicators of environmental
impact, because they are largely sedentary and unable to escape the effects of
environmental degradation.
2 Study site
The Abashirti River basin is located in western Hokkaido, has an area of
1380 km2, a channel length of 115 km, and drains into Lake Abashiri (Fig.1).
This is a lagoon lake with a surface area of 32.3 km2, a maximum depth of 16.1
m and mean depth of 6.1 m. The fishery resource species in the lake are
Corbicula japonica, Hypomesus nipponensis and Salangichthys microdon. The
salinity in surface water ranges from 1.0 to 3.5 PSU between May and
December. A large portion of the lake is anoxic below a depth of 5 m. The
Abashiri River flows into the lake on the south shore, and out of the lake
northwards into the Sea of Okhotsk. The lower and middle reaches of the basin
are cultivated for potato, wheat and beet, while coniferous forests cover the
upper reaches.
The western portion of the watershed has fragile sedimentary rock, whereas
volcanic pyroclastic flow deposits are found in eastern regions, forming a plateau
along the middle reach of the river. Numerous landslide scars have been
observed in eastern parts of the watershed.
The annual mean precipitation is around 800 mm, with snow cover from
November to April. Snow melt usually occurs from early April to mid May.
Heavy rainfall sometimes occurs during summers, and daily precipitation may
exceed 100 mm.
¯ Abashiri L
m1
t1
m2
t2
m3
t3
t4
m4
m5
t5
t6
t8 m6 t7
t9
Tributary
t10 Main channel
m7
t11
0 2 4 8 12 16
Km
Figure 2: Water sample collection points.
3 Method
3.1 Sample collection
Seven water collection stations were established along the main river channel
(St.m1~m7), and 11 stations were established at the mouths of tributaries (St.
t1~t11) (Fig.2). Samples were collected by the Abashiri Watershed Environment
Protection Board during downpours in spring, summer and autumn from 2003 to
2005. Bivalves were collected with scooping net and bottom sediments were
collected with a Smith-McIntyre mud sampler (0.05 m2) at depths of 3.0 at 50

stations along the lake margins.
3.2 Sample analysis
Suspended sediment concentration in water samples was determined by

percolation through glass fiber filters (0.5 µm mesh). The percolated samples
were dried at 50oC for 48 h and weighed, then combusted at 550oC in a muffle
furnace for determination of organic dry weight (percent of ash free dry mass).
Sizes of particles included in water samples were analyzed using two methods:
1. particles larger than 0.063 mm were separated out in sieves with
different mesh sizes, then dried at 60oC for 24h;
2. smaller particles than 0.063 mm were analyzed with a particle analyzer
(CAPA-300 Horiba).
Bottom sediments were treated in the same way as suspended sediments.
The C. japonica specimens collected in the sampler were counted, and the
density per square meter was calculated. The width, height and length of each
shell were measured with vernier calipers (Mitutoya, CD-20GM). Each soft
tissue was dried and weighed. The condition of the animals was calculated as a
dry weight per volume index (x 1000).
3.3 Spatial analysis by Geographical Information Systems
Land use on each tributary was measured using digital land use maps supplied by
the national Ministry of Land Infrastructure and Transportation. Environmental
factors, such as forest coverage, agricultural land, lithological hardness, slope
gradient, and fault densities were also measured from these maps using GIS
software (TNTmips ver.6.8, ESRI Arc View 3.2a and Spatial analyst).
4 Results
4.1 Suspended sediment concentrations in spring and summer
During the study period, we collected water samples in August 2003, April 2004
and April 2005. There were no large discharges in the summer of 2004. The
highest concentrations of suspended sediment were observed in August 2003.
Figure 2 shows the discharge during downpours in that month. This discharge
was induced by torrential rainfall exceeding 79 mm on August 9-10. Sampling
was carried out during peak, regression and normal discharges (Fig.3).
Suspended sediment concentrations at all stations along the main channel,
except at St. m1 and m7, were higher than 1600 mg/l, and the tributaries draining
from the left bank transported sediment at even higher concentrations. These
values declined remarkably with time to <50 mg/l at all sampling stations after
ten days (Fig.4). Most of the sediment was inorganic [>80% except at St.m1
(uppermost headwater) and St.m7 (downstream from the lake)].
Peak dishcarge
4
Watr table (m)

3.5
3
One day after
2.5
9 days after
2
1.5
1
0.5
0
8/19
8/21
8/23
8/25
8/27
8/17
8/29
8/31
8/11
8/13
8/15
8/1
8/3
8/5
8/7
8/9
Date (2003)
Figure 3: Discharge at August 2003.
10th Aug. SS 11th Aug. SS 19th Aug. SS
L. Abashiri L. Abashiri L. Abashiri
Figure 4: SS concentrations observed in Aug. 2003.

4.2 Particle analysis
Particle distribution in the suspended sediments did not differ significantly
between the main channel and tributaries. Finer particles, less than 10 µm,
comprised 40% of all samples, followed in abundance by medium sized particles
between10 µm and 0.063 mm. Coarser particles usually comprised <10% of all
samples.
4.3 Bottom sediment and C. japonica in the lake
Bottom sediments were collected at three depths, at 50 stations around the lake.
Fig. 6 shows the percentage of fine particles <0.063 mm at each sampling
station. Fine particles were more abundant on the southern shore where the river
flowed in, and on the northern shore where the river flowed out. The percent of
finer sediments tended to increase with water depth.
<10μm 10μm～0.063mm 0.063mm>

100%
80%
60%
40%
20%
0%
St.m5
St.m6
St.m2
St.m3
St.m4
St.t1
St.t11
St.t2
St.t3
St.t4
St.t5
St.t6
St.t10
St.t7
St.t8
St.t9
Main channel Tributary
Figure 5: Particle distribution in suspended sediment at the all stations.
River flow out
Silt(%)
Silt(%) Density(N/m
Density(/m2) 2) Fatness
Fatness index
- 400 -15.0
-4
-16.5
- 14 - 1000
- 18.0
- 35 - 1800
- 19.5
River flow in - 68 - 2600
-
- 21.0
- 92 - 4000
0 1 2 4
Kilometers
Km
Figure 6: Percent of silty substrate (<0.063 mm) (left), density (N/m2)(center)

and fatness index of C. japonica (right) in the Abashiri Lake.
C. japonica specimens were collected at the same sampling stations.
Densities and fatness were higher on the western and eastern shores than on the
northern and southern shores (Fig.6). There was a negative relationship between
bivalve abundance and the percentage of finer particles. We also found a
negative relationship between the fatness of C. japonica and the percent of fine
particles.
4.4 Land-use and other factors affecting sediments discharge revealed by
GIS analysis
Factors affecting sediment yield were calculated for the eleven tributaries
(Table 1). The watershed area was 22.3 to 202.3 km2. The agricultural area
ranged from 0 to 56.5% of each watershed, while the forested land occupied
between 36.5 and 100%. The percentage of slopes steeper than 20 degrees was
smaller for downstream tributaries than for upstream tributaries. Soft rocks
dominated western parts of the watershed. Higher densities of faults developed
in these tributaries.
The relationships between suspended solid concentration and percentage
cover of agricultural land were unclear; the larger agricultural tributaries had
lower sediment concentrations in August 2003, whereas forested tributaries, such
as t9~t11, had the highest suspended solid concentrations. Slope steepness was
not correlated with the particle concentration. The t5 tributary had the steeper
slope, but not the highest suspended solid concentrations. Presence of soft rock
strongly influenced the suspended solid concentration (tributary t10 showed the
highest concentration).
Multiple regression analysis was performed to determine proportional effects
of environmental factors on variations in suspended solid concentration.
Lithological hardness and the densities of faults (medium) were positively
correlated with suspended solid concentrations (Table 2).
Table 1: Environmental variables in each tributary.
Area Environmental variable(%) Highest SS

Tributary 2 Agricultural Forest Higher density Steep slope concentration
(km ) Soft rock (%)
land (%) coverage (%) of faults (%) (>20°)(%) (mg/l)
t1 126.3 56.53 36.59 10.26 0 0.08 265.15
t2 39.7 35.57 59.46 12.39 0 3.7 377.2
t3 202.3 19.75 73.21 0 0 12.04 258.5
t4 54.6 9.7 89.49 11.59 11.63 48.63 335.05
t5 87.9 6.57 92.6 53.04 19.73 82.08 666.25
t6 152.2 8.25 86.81 14.87 0 59.49 208.05
t7 75.7 13.49 85.49 58.83 0 56.39 741.4
t8 52.5 5.79 91.02 55.07 35.37 82.35 405.25
t9 108 9.77 89.32 40.57 12.11 59.3 1499.4
t10 30 0.14 99.68 62.95 15.16 76 1872
t11 22.3 0 100 53.67 0 54.85 704.6
5 Discussion
Land use practice usually strongly influences sediment yield (suspended
sediment concentration correlates positively with the proportion of agricultural
land, and negatively with the proportion of forested land [15]). Our previous
study showed that sediment concentrations exceeding 2000 mg/l were produced
in highly agricultural watersheds; the sediment sources were landslide scars
resulting from soil compaction by agricultural machinery [11, 13]. However, in
the present study, we observed lower suspended yield in larger agricultural
tributaries. Instead, suspended sediment concentration was strongly correlated
with lithological conditions (Table 2). The western tributaries (where a large
portion of watershed is on soft rock) are assumed to be the major source of
suspended sediment for the main channels (Fig.4). Furthermore, higher
suspended solid discharge may affect sedimentation in Lake Abashiri (Fig.6).
The lithological characteristics of the watershed are such that land use must be
carefully managed, even for forested regions.
In these mountains, numerous landslide scars were observed along channels
(Sato unpublished data). These scars resulted from lateral scouring of stream
banks during large water discharges; soil wash out can be a serious consequence
of this. Other sources of stream sediments may be timber harvesting and logging
road construction. The higher densities of logging roads were near St.t9 and t10
tributaries. Roads can affect streams directly by accelerating erosion and
sediment loading, by altering channel morphology, and by changing the runoff
characteristic of the watershed [5]. In addition, cut branches and trunks are piled
along streams, and this debris is easily swept away during downpours. In order to
protect land cover, precautionary forest practices should be followed. For
instance, although there is no regulation on the cutting of riparian forest, a buffer
strip of more than 20 m wide should be protected. Logging roads should avoid
streams and inner valley gorges. The number of stream crossings should be
minimized [17]. The exposed soil of cuts and fills should be planted with grass
or rapidly growing shrubs. One effective method for decreasing surface water
runoff is the installation of ditches and cross-drain culverts along logging roads
[9]. Sediment ponds at the end of small catchments may be effective in trapping
and reducing suspended matter downstream.
Table 2: Results of multiple regression analysis.
Un standerdrized coefficient 2
Variable Probability R
B
Constant 26763.337 0.0573
Agricultural land (%) -279.891 0.0621
Forest coverage (%) -266.837 0.0522
0.801
Steeper slope <20°(%) -19.366 0.422
(P=0.132)
Soft rock (%) 33.933 0.0198
Fault density (medium) (%) 95.412 0.0475
Fault density (higher) (%) -103.864 0.0315
Suspended sediments have strong effects on aquatic biota [10, 17]. Inorganic
turbidity presents a wide range of problems, particularly for filter feeding
organisms. The harmful effects on suspension-feeding bivalves include gill
tissue fouling, increased pseudofaecal production [18], decline of ingestion rate
resulted from reduction of clearance rate [3], depressed growth rates through
overloading the gut and gills with inorganic solids [4, 12], reduced metabolic
rates and feeding efficiency [1]. Respiration rates are significantly affected by
acclimation temperature and turbidity [2, 14]). Based on these studies, it was
likely that fine sediment derived from upstream tributaries had a significant
influence on the lake inhabiting bivalve C. japonica. The higher suspended solid
concentration (>1600 mg/l) occurred at St. m2, which was the nearest station
upstream from the lake. Concentration declined drastically at St.m1, just
downstream from the lake (Fig.4). Hence, most of sediment transported from
upstream probably accumulates around the river mouth on the southern shore,
and the lower density and lower growth of C. japonica in that area may be
attributable to material produced upstream.
Salinity is a factor controlling distribution of the C. japonica [7]. In this
study, high salinity considered to be one of negative factor to influence the
density and fatness of C. japonica on the north shore (Fig.6) where seawater
flows in from the Sea of Okhotsk during high tides. Because the frequency of
open shells in C. japonica is generally lower in higher salinity conditions
Corbicula sp. contributes significantly to total benthic community respiration
[6]. In particular, C. japonica uses organic material derived from terrestrial
sources [8]. However, the proportion of organic matter was <15% in the
sediments deposited around the river mouth, and it is likely that the negative
effects of fine inorganic sediments outweigh the positive effects of any organic
food source in the sediment. Management of upstream forests to decrease fine
inorganic sediments and increase organic content may be important for
conservation of fishery resources in the lake represented by C. japonica.
Although concerns for river management have increased recently in this
region, a watershed network (a partnership of local residence involving various
stakeholders such as fisherman, farmers, and foresters aimed at preserving water
quality and regulating land use practice) has not been established in the province
of Hokkaido. More detailed data on the influence of land use practice on fish
resources will be required for better watershed management.
Acknowledgements
This study was performed with the cooperation of Marine science center of
Abashiri, and Abashiri fisheries cooperation. We would also like to be grateful to
Mr Kamada T., Kikuchi T.and Moriwaka H., who were students of Hokkaido
Institute of Technology and assisted a field investigation and laboratory analysis.
References
[1] Aldridge D. W., Payne B. S. & Miller A. C., The effects of intermittent
exposure to suspended solids and turbulence on three species of
freshwater mussels, Environ. Pollut. 45, pp. 17-28, 1987.
[2] Alxander J E., Thorp J. H. Jr., and Fell R. D., Turbidity and temperature
effects on oxygen consumption in the zebra mussel (Dreissena
polymorpha), Can. J. Fish. Aquat. Sci. 51, pp. 179-184, 1994.
[3] Bricelj, V. M. & Malouf R. E., Influence of algal and suspended sediment
concentrations on the feeding physiology of hard clam Mercenaria
mercenaria. Mar. Biol. 84, pp. 155-165, 1984.
[4] Bricelj, V. M., Malouf R. E., & de Quillfeldt C., Growth of juvenile
Mercenaria mercenairia and the effect of resuspeded bottom sediments,
Marl. Biol., 84, pp. 167-173, 1984.
[5] Furniss M. J., Roelofs T. D., & Yee C. S., Road construction and
maintenance. Influences of forest and rangeland management on salmonid
fishes and their habitats, ed. in W. R. Meehan, pp. 292-323, 1991.
[6] Hakenkamp C. C. & Palmer M. A., Introduced bivalves in firewater

ecosystems: the impact of Corbicula on organic matter dynamics in a
sandy stream, Oecologia, 119, pp. 445-451, 1999.
[7] Ishida O., & Ishii T., Salinity tolerance and regional morphological
differences for Corbicula japonica. Fishery culture 19(4), pp.167-182,
1971.
[8] Kasai, A. & Nakata A., Utilization of terrestrial organic matter by the
bivalve Corbicula japonica estimated from stable isotope analysis.
Fisheries Science, 71, 151-158, 2005.
[9] Moore G. D., Resource road rehabilitation handbook: Planning and
implementation guidelines. Watershed Restoration Technical Circular no.
3. Ministry of environment, lands and Parks and Ministry of Forest.
Province of British Columbia, Canada, pp.99, 1994.
[10] Nagasaka, A., Nakajima, M., Yanai, S. & Nagasaka, Y., Influences of
substrate composition on stream habitat and macroinvertebrate
communities: a comparative experiment in a forested and an agricultural
catchment. Ecology and Civil Engineering, 3, pp. 243-254, 2000 (in
Japanese with English abstract).
[11] Nagasaka A., Yanai S., Sato H. & Hasegawa S., Soil erosion and gully
growth associated with cultivation in south-western Hokkaido, Japan.
Ecological Engineering 24, pp. 503-508, 2005.
[12] Payne B. S., Lei J., Miller A. C. & Hubertz E. D., Adaptive variation in
palp and gill sizen of zebra mussel (Dreissena polymorpha) and Asian
clam (Corbicula fluminea). Can. J. Fish. Aquat. Sci. 52, pp. 1130-1134,
1995.
[13] Sato Hi., Yanai S., Nagasaka Y., Nagasaka A. & Sato Ha., Influences of
land-use on suspended sediment discharge from watersheds emptying into
Funka-Bay, south-western Hokkaido, Northern Japan, J. Japan Soc.
Hydrol. & Water Resour., 15, pp. 117-127, 2002 (in Japanese with
English abstract).
[14] Summers R. B., Thorp J H., Alexander J. E. & Fell R. D., Respiratory
adjustment of dreissenid mussels (Dreinnena polymorpha and Dereissena
bugensis) in response to chronic turbidity, Can. J. Fish. Aquat. Sci., 53,
pp. 1626-1631, 1996.
[15] Walling D. E. & Webb B. W., Water quality 1. Physical characteristics.
River Flow and Channel Forms, ed. G. Petts & P. Calow, Blackwell
Science Publishers, pp. 77-101, 1995.
[16] Walling, D. E., Suspended sediment yields in a changing environment.
Changing River Channels, ed. A. Gurnell & G., Petts, John Wiley & Sons
Ltd., 1995.
[17] Waters T. F., Sediment in streams, sources, biological effects and control,
American Fisheries Society Monoghraph 7, pp.251, American Fisheries
Society, Bethesda, Maryland, 1995.
[18] Way C. M., Hombach D. J., Miller-way C. A., Payne B. S. & Miller A.
C., Dynamics of filter feeding in Corbicula fluminea (Bivalvia:
Corbiculidae). Can. J. Zool. 68, pp. 115-120, 1990.
Potential of poplar plantation for enhancing

Polish farm sustainability
A. Czarnecki & A. Lewandowska-Czarnecka
Institute of Ecology and Environmental Protection,
Nicholas Copernicus University, Torun, Poland
Abstract
Farm sustainability depends on policy that is able to protect natural resources
while being open to the market. Traditional crop pressing on natural processes
creates an open input-output unstable system. Farming situated in a postglacial
landscape cannot counteract erosion and does not engage in all natural resources.
The poplar stand has an ability to renew the soil system and assimilate the
surplus of mobile substances. This paper presents a proposal to combine poplar
with a traditional crop to make the biosystem able to engage in all resources to
develop efficient production while making the system more stable. The concept
is related to the results of the investigation carried out on farms dealing with
typical problems and poplar stands in a similar soil condition.
Keywords: farming, poplar, crop, sustainability and policy.
1 Introduction
Until the 1980’s Polish agricultural production was controlled by the state,
whose aim was to produce maximum crops, the results of which strongly impact
on the environment, compromised natural resources and the functions of the
ecosystem on arable land. Poland’s past agriculture intensity reduced soil
structure [1] reaped the landscape of vital nutritional elements. Natural habitats,
hedgerows, the majority of wetlands and small surface water basins have all but
vanished [2], those which remain are (predominantly) highly polluted due to
neighboring agro-ecosystems. The 1990s saw the external conditions for farming
totally change. Production met the limited market, resulting in decreased prices
for primary agricultural products; forcing farms to begin careful management of
natural resources, and an initially expensive investment into farming methods, to
doi:10.2495/GEO060251
increase productivity. Land, which is more productive, is pushed to excessively

productive, whilst less productive land has been abandoned. The government is
proposing to re-forest, the so far, 2 million ha of abandoned arable land. For
national and international interests the reforestoration is a positive action,
however, the tree type requires careful consideration to achieve maximum
economic benefits for farmers of this unproductive arable land. Poplar has this
potential and timber as an initial crop. It grows fast, and reaches maturity within
25 years. Poplar requirements would combine not available for traditional crop
resources [3]. Timber and biomass can revive local economy on the basis of soil
productivity. Considering continuity of farm business the surplus of land and
other environmental and social resources ought to be engaged in agricultural
activity giving not only direct income but also combining into a process of
attaining sustainability [4]. Thus, poplar has potential as an extensive crop. The
aim of this work is to establish the total expected value from poplar cultivation
on a traditional farm in view of a policy driver for long-term sustainability.
2 Sustainability and agriculture

The farm manages ecosystem products and services. Basic human needs that are
expected to be provided by the farm are foodstuffs. However, economic viability
requires farmers to attentively manage ecosystems and social services provided
by the farm. Economic viability means reduction in manpower and
implementation of agricultural machinery, but it also means diversification and
looking for new markets. Biomass and timber are becoming important crops.
Presence of poplars can improve landscape for adaptation and inclusion of land
for leisure pursuits. Farmers should engage all resources proportional to the
markets, and create new jobs, though economic gain must be balanced with
natural processes of a cyclic renewal of resources, which dominate the location.
The position of farmers within local community means that they are responsible
for the state of resources [5] and the environment over a larger area than just the
farm, as farming processes interact with ecosystems [6]. Therefore, sustainable
management methods must respect natural resources, while retaining land
productivity. By using renewable resources produced under proposed poplar
plantation systems, such as biomass, farmers reduce productivity costs and
enhance efficiency. Long term changes in agricultural methods require
governmental and also farm policies to transform them into operational plans.
3 Methodology
Knowledge based decisions on farm tends to be sustainable and flexible, fully
combining resources and farm productivity with respect for ecosystems. When
considering the farm as a socio-economic unit, its state is important. An analysis
requires a collection of several data sets i.e. economic, socio-economic, and
ecological. This will produce a basis for establishing gaps between the actual and
sustainable farm status, highlighting obstacles preventing actualization of
sustainable methods. Problems encountered by farms in Polish lowlands were
used for this paper. The data were collected qualitatively over a 10-year period
by the authors. Each problem the study raised is shown in fig. 1.
Figure 1: Conceptual diagram presenting the approach to farm analysis.
The diagram (fig.1) enables farm needs to be quantified, and the potential
contribution to answer specific farm needs to before filled. Hydrological
processes and interaction between soil and trees productivity in time, the
retention of biodiversity and active ecosystems need to be considered when
establishing the cost both economically and in a sustainable way. The case
studies analyzed have enabled quantification of these problems, and the cost of
traditional farming methods vs. poplar plantations. The results of the farming
combining tradition and poplar crops will be discussed. The data are transformed
into economic–efficiency of input, socio-economic-employment, and ecological-
N circulation categories.
4 Site locations
All the farms are situated in postglacial terrain with differing relief within a
moraine complex and a river valley bottom on the moraine plateau, in a
depressing part of the ground within moraine, on the edge of a moraine and a
valley as well as in the river valley. All farms need reduced overheads but
increased efficient production by utilizing farmed derived biomass, to increase
employment opportunities for families and the local community. Farm 5 is
located in fragile areas; where surface erosion and loss of organic matter impair
production predominantly. Farms 1 and 3 situated in the river valley are affected
during periods of flood. Intense cattle farming in farm 2 and 4, within the river
catchments caused nitrogen contamination. The characteristics of each farm are
presented in table 2.
Table 1: Functional characteristics of farms.
Specifications Farm 1 Farm 2 Farm 3 Farm 4 Farm 5
farm areas (ha) 49,75 18,24 70,12 24,88 42,00

in that: forest 0,00 0,00 0,00 0,00 5,00
meadows and 28,06 0,00 32,12 10,28 4,00
pasture
set aside 2,94 0,61 0,00 1,10 0,00
arable fields 18,75 17,63 38,00 13,00 32,50
Main crops Fb, M, B, O, T, P, B, M, Sb, P, J, T, B, P, Z, M R, Po, Mz, Z,
W, T, P Z, B Po Bp
Livestock Dc 25 0 6 14 4
C 14 4 0 0 0
Bc 4 4 0 14 10
P 5 80 0 100 150
soil bonitation Al IV-VI, Al IV-VI Al II- IV Al III-IV, Al IV-VI
class P IV-VI, M IV M V-VI M V, VI P
N - mineral a 16,28 49,12 130,49 73,83 95,34
fertilizers b 43,20 50,82 240,79 138,46 107,08
N total input b 14,46 80,16 52,92 48,76 87,49
yield contents
Energy input: 18,89 35,78 33,94 30,3 28,46
(GJ/ha/year)
in that: renewal 8,38 13,3 8,99 9,39 10,75
Efficiency:
Money 1,41 3,54 2,56 5,79 1,95
(PZL/PZL)
cropping livestock Al – arable lands

W – wheat P – pigs M – meadows
T – triticale Dc – dairy cow P – pasture
B – barley Bc – beef cattle bonitation class I –VI
P – potatoes Sc – slaughter cattle a estimate on farm areas
Fb – fodder beet C – calves b estimate on crop
O – oats P – pigs
R – rye
M – maize
Poplar plantations situated on land exhibiting similar characteristics to that of

the five farms of the study were analyzed. These comparative sites consisted of
21 poplar plantations on medium to poor quality of soil. The plantations were
established by state forest services on abandoned arable land and meadow (tab.
2). Research into poplar sites was conducted under project PAMUCEAF
(FAIR6-CT98 4193).
Table 2: Poplar stands characteristics.
Characteristics Description Comments Comparing with arable

land
Suitable site, mineral Underlying Possible to form poplar Low agriculture value
soil, above 7% average coarse-textured plantations
clay content coarse- subsoils
textured
Very suitable Underlying Deep subsoil Also good for traditional
Medium-texture medium- Ground water supply crop
textured Not access to ground water
subsoils
Changes in soil N 22,14 g/kg Litter layer is forming, Two times higher
C 1,82 g/kg comparing with arable land
Bulk density soil 1,6 mg/m3 Soil is loosened by Lower then arable soils
subsoil 1,4 mg/m3 roots and macrofauna lower compaction
AWC (available water O,06 m3/m3 Lower bulk density and Slightly higher then arable
capacity) larger organic matter soils approximately 0,01-
retention 0,02 m3/m4
Contents in soil P 0,03 g/kg Soil is clean up In poplar stand three times
K 0,1 g/kg Soil matrix contents less
The same
Production per year/ha 10-20 m3 Depend on soil, Similar productivity
available water
N immobile per year To 200-300 kg Soil development, Accumulation and storing
biomass enlarging
5 Farmer’s needs vs. poplar contributions

Ecosystems confine the threshold of farm productivity within a sustainable
system. This needs to be considered when discussing a farming decision for an
economic output. This is represented by a group of variables: the state of the
farm, farmer’s requirements, socio-economic and environmental factors.
Poplar plantations are characterized by two main factors: natural species
composition and long duration without human intervention allowing ecosystem
development. From tables 2 and 3 it can be seen that poplar planted farms have
more positive net gains [7] when compared to similar arable lands. The ability to
store nutritional substances and water was higher due to soil developing an
accumulation of organic matter and soil loosening volume. The poplars enabled
soil to renew its profiles by accumulation and assimilation properties, and also
yielding timber and biomass, all positive outcomes of the poplar system (tab.3).

6.1 Quantifying problems on the investigated farm
On each farm one of the key problems was erosion and surplus of resources.
They were estimated quantitatively as losses and resources not fully engaged in
production. Other categories assessed were: reduction of annual income,
decrease in soil productivity (as natural capital), and an ability to cover
discounted rates of machinery and equipment. Some factors were considered
external factors; they included the effects on the local community and the
environment. Surplus nitrogen combined with eroded material deposited in the
valley bottom causes pollution and trophication. It can limit farm environment
and indirectly local development. The loss of nitrogen, due to geographic
location of the farm can be up to 200 kg N per ha/year (farm 5). All farms within
the study do not create adequate conditions for crops, hence they reduce yield by
approximately 40%. Farm 1 has very extensive production. Farm 3 with meadow
situated in a high-risk flood zone causes the farmer to limit productivity (see
table 1).
Table 3: Potential contribute of poplar to farm sustainability.
Problem/farmer needs Criteria Expectations from new crops
Better efficiency of input Fixed plant cover on fragile

Decreasing loss Fertility conservation land
Decreasing leaching of biogens Extensive crop
Loss of productivity caused by Mixed crop of different
erosion requirements
Biomass
Application for free Fragile land application Biomass from extensive long
resources term crop
Assimilation of N overdose Assimilation of organic
fertilizers
Application for surplus of human
and man-made resources New products/resources
Forming new resources/ carpentry, tourism, leisure
Prospect for future
Input decreasing
Better efficiency of applied
Requirements resources
Poplar stands Differentiation of activity
Improving landscape
Cheaper input
Enhancing income
Social acceptation Crop for abandoned land
Replace present crop
Capital Silvopasture system
Labor
Resources, which are not applied, reduce total farm productivity by 40-60%.
One of these is a sub-terrestrial hydrological system, which runs water from the
moraine to the valley floor, below the level of accessibility for traditional crop
roots. For all farms mechanical and human resources need to be applied more
effectively. The two farms (2 and 4), which conduct a secondary productivity
results in excess nitrogen being introduced into the valley catchments,
externalize effect of farm management.
Table 4: Losses in natural capital and not applied resources in investigated

farms.
factors criteria characteristics farm

1 2 3 4 5
Soil 3 2 3 1 3
Losses erosion Input 2 2 2
input Water 3 3 1 1
Yield 3 2 1 1
insensitivity Landscape 1 3 2 1
Biogens Surplus of N 3 3 3
Resources Productivity Meadows 3 3 1
not Land Field 3 3 2 1
applicable Labor Human resources 2 3 2 2
Man-made Machinery 3 1 3 1 2
capital
1 little 2 medium 3 substantial.
Table 5: Capacity of poplar to solve farm problems.
Process Farm 1 Farm 2 Farm 3 Farm 4 Farm 5

Apply all forms land 3 1 3 1 2
Low input 3 2 2 2 3
Flexibility 3 1 3 1 2
Development 3 1 3 1 3
Complement production 3 3 2
Biomass 2 3 2 2 2
Secondary production 2 2 1 3
Raw materials/workshops 3 1 3 1 2
Leaching and erosion 1 3 1 3 2
Assimilation 2 1 2 1 3
Extensive permanent crop 3 1 3 1 2
Landscape diversifying 1 3 2 3 2
Naturalness 1 2 2 1 1
Effect vs. problem 1 -3 little, medium, substantial.
6.2 Expected poplar contribution to farms
Farmers cannot afford to allow for further decrease in ecological processes and
ecological efficiency. However, the farmer requires mutual gain from protecting
these ecological systems, by means of free natural resources to improve
production. The solution for many of these environmental and ecological
problems is deep-rooted plant cover, which would stop and assimilate nutritional
substances. Poplar plantations with solve main problems so far raised but in
doing it with different efficiencies. Poplar crops yield raw material, which has a
possibility to be processed on the farm; trees as an extensive crop improve soil
development by creating better storage capacities for nitrogen and water. Poplars
causing soil to become looser, allowing it to be used as strips among fields or
part of crop rotation. It has been suggested compaction by heavy machinery
causes a dead pan that vanished under poplars. In soil under poplar, phosphorus
(P) quantities fall, likely to be caused by P’s immobility and assimilation into
biomass. Farms proposing such a dynamic system need to consider markets, the
environment, and ecosystem processes [8]. Management of these systems and
socio-economic systems require specific knowledge and skills to maximize
potential outputs. The movement and processing of lumber requires more
equipment, machinery, and man-power, a consideration within cost effectiveness
of poplar plantations, though the reduction in farm overheads by the use of
biomass, may go someway to meeting these costs. Poplar plantations are planned
for each farm not in view of above requirements but only due to the farm size,
the problem experienced by the farm, both economically and soil degradation,
and the proportion of resources free to be engaged within the project. In some
cases a part of traditional cropping would be converted into poplars, giving
benefits of increased income or meeting local community, or national legislation.
Knowing what poplar stand would change in it were estimated proposal in term
of stand characteristics considering free resources on farm. In relation to the
main problem on particular farm and free resources they dispose, poplar stands
differ substantially. Thus contribution to solve problems ranges between 20-
80%. The biggest potential contribution of poplar is on farm 2 and 5 no least in
farm 2 and 4. That’s mean that consequences to farmer needs differ between
farms.
7 Conclusion
Comparison between the factors that arise from transformed external conditions
for food production and the state of farms enabled to establish threats and gaps in
the state that prevents adaptation process. Results of the investigations proved
that poplar plantation would positively impact the state and money flow on farms
if only the whole farm is designed according to natural processes as surface,
interflow and base flow as well nutrients of water flow as well by fitting poplar
stands to overall money flow in farm. Farms situated in postglacial landscape in
Poland would use poplar as a crop to facilitate achieve the balance their
performance with nature and external conditioning for farming. In dependence of
main problem and quantity of free resources on farm, poplar stand would
contribute less or more to development of farm business as well as
internalization of matter turnover to satisfactory degree.
References
[1] Dziadowiec H., A. Czarnecki, J. Jonczak. Roczna i wieloletnia dynamika
wybranych właściwości uprawnych gleb płowych Glebowej Powierzchni
Testowej ZMŚP w Koniczynce. X Ogólnopolskie Sympozjum
Zintegrowanego Monitoringu Środowiska Przyrodniczego. Kampinoski
Park.
[2] Jaworowski P., Cz. Sobków, A. Czarnecki, T. Celmer, J. Szablowski.
Melioracje wodne, ich wpływ na środowisko przyrodnicze i gospodarkę
rolną. wyd. UMK, p. 210,. 1996.
[3] Cannell, M.G.R., Van Noordwijk, M., Ong, C.K., The central agroforestry
hypothesis: the trees must acquire resources that the crop would not
otherwise acquire. Agrofor. Syst. 34, pp, 27–31. 1996.
[4] Schoorl J.M., A. Veldkamp. Linking land use and landscape process
modeling: a case study for the Alora region (south Spain). Agriculture,
Ecosystems and Environment 85, pp, 281-292. 2001.
[5] Riley J. Multidisciplinary indicators of impact and change. Agriculture,
Ecosystems and Environment 87, pp, 245-259. 2001.
[6] Erickson J.D., J.M. Gowdy. Resource use, institutions and sustainability: a
tale of two Pacific island cultures. Land Economics, 76(3) pp, 345-354.
2000.
[7] Cacho O. An analysis of externalities in agroforestry systems in the
presence of land degradation. Ecological Economics 39, pp, 131–143.
2001.
[8] Torquebiau E. F. A renewed perspective on agroforestry concepts and
classification. C.R. Acad. Sci. Paris, Sciences de la vie / Life Sciences 323
pp, 1009–1017. 2000.
Section 7
Hydrological studies
Modified quantitative estimation model

of erosion and degradation in
three mountainous watersheds
D. Emmanouloudis1 & M. Kaikis2
1
Department of Forestry and Natural Environment Management
Technological Educational Institute of Kavala, Annex Drama, Drama,
Greece
2
Rhodes Municipal Environmental Organisation
Abstract
This paper presents a new version of a well-known stochastic model to
quantitatively assess drainage basin degradation in order to determine erosion in
three torrent basins. This new version, 3D-structured and using G.I.S. for data
processing, can provide a quantitative assessment of the degradation of any
torrent basin in a very short time. Furthermore, a system to prevent erosion and
degradation by means of agricultural and technical forest works is presented; the
system was originally designed for these three basins, but it is flexible and easy
to adapt to any torrent basin.
Keywords: stochastic model, GIS, degradation, 3D Digital Elevation Model.
1 Introduction
Drainage basin erosion and degradation is one of the most complex
environmental problems worldwide. Thus, it has naturally been the subject of
research for many decades in countries representative of all the main climatic
types on the planet. The problem is fairly aggravated in the Mediterranean
countries, due to their temperature range and other factors (irregularly distributed
rainfall, insufficient plant coverage, etc.). In our country, plagued by forest fires
in the past decades, it tends to acquire the dimensions of a real disaster for
mountainous areas, due to the degradation of the soil it implies.
doi:10.2495/GEO060261
The assessment of mountainous soil degradation due to the erosive action of

rainwater runoff is of two types:
a) Qualitative assessment (intensity, degree, speed, etc.)
b) Quantitative assessment, usually calculated in total m /year/basin, or in
m /year/Km, or finally in mm/year/basin.
Quantitative assessment is more difficult than the qualitative one, since it implies
an attempt at calculating magnitudes and volumes which are hardly measurable
in practice and are related to a diachronic effect. In the past few years several
methods have been developed for the quantitative assessment of degradation.
The more known methods are referred to be the USLE, WEPP, AGNPS1 [1],
CREAMS, ANSWERS [2], EPIC, the method of Rydgren [3] and Terry [4], and
the method of Mitas et al. [5], as well as the contemplative methods of Fournier,
Corbel and Gavrilovic. The most of them have been used mostly in United States
and less in Europe.
However, according to Kotoulas [6] the main feature of these methods is that
"they require the existence of basic data which however can not be often
designated precisely so that the precision and the potential of their application
are reduced". Therefore, according to the same author "the methods being
developed to date are far from the point to provide precise and reliable
forecasts". Moreover, in the chapter of erosion-degradation this is not the only
difficulty. There is questioning even on how to confront the phenomenon
particularly in situations during which we have intensive procedures (e.g. after a
fire).
In the present paperwork one from the above mentioned methods was
selected, the stochastic Gavrilovic [7] model, by the help of which effort was
made to estimate the degradation in the selected under research area. This
estimation was conducted by the help of specific G.I.S. software (GRASS),
which has helped out to simplify the model application on one hand and to
increase the precision of the results to an impressive way on the other. After the
estimation of degradation was implemented, the quotation of a protection and
how to handle a problem was attempted, by applying a system easy-to-use,
flexible and particularly effective, as proved by the experience gained to date.
2 Research area
As research area was selected the broader area of Kastoria Prefecture and
especially four small torrential watersheds from which two were directly
adjacent. As a reason of the above selection is quoted the fact that the aforesaid
area -because of its geomorphologic condition- has many, in direct or less direct
proximity, small runoff basins which particularly show a special vegetative and
geological diversity.
Of course, it is easily comprehensible why a model, in order to be elaborated,
requires watersheds of small size. Regarding the highest possible proximity, an
identification of the attack climatic factors in all parts of the sample is essential,
so that the results obtained by the research are directly comparative. Following
these thoughts we selected the above stated sample of four watersheds.
3 Research method
The morphometric characteristics of the selected watersheds have led us
necessarily (because the so called method provides particularly good results
when estimating the degradation in small basins of mountainous areas Gavrilovic
[7], Pintar [9], Kotoulas [6]), to the selection of the Gavrilovic method. This
method is uniquely appropriate to estimate the degradation in small mountainous
watersheds, as the ones included in our sample.
In Table 1 the morphometric characteristics of the sample watersheds are
quoted, which define the mountainous character of the specific torrents.
Table 1: Morphometric characteristics of the sample watersheds.
Fwat. (Km2) Hmax (m) JmW (%)

No Torrents Projective 3D" (max (mean watershed
elevation) slope)
1 Riganorema 9.43 10.03 1420 30.90
2 Fotini 4.23 4.46 1190 29.41
3 Triantafillia 8.42 8.98 1580 33.32
F3D stands for the watershed area, not as it results from the calculation on the
map (FPR0J) but for the actual area of the basin in three-dimensional format, as it
results from the calculation after a 3D construction of the watershed by the help
of GRASS. The F3D is slightly bigger than FPE0J and the deviation from each
other depends on the average surface slope of the watershed, Emmanouloudis
and Filippidis [10].
The above data was obtained from maps 1:50.000 of the G.A.S. (Sheets
KASTORIA, MESOPOTAMIA).
According to the Gavrilovic method eqn (1), the type providing the average
annual degradation in a watershed is:
W = T ⋅ h ⋅π ⋅ z 3 ⋅ F (1)
where T, eqn (2), is coefficient of temperature given by the type:
t0
T= + 0,1 (2)
10
with t0: average annual temperature in the watershed
h: average annual height of rain (mm)
π: 3.14
F: area of watershed (Km2) and
z: coefficient of erosion given by the ratio z = x ⋅ y ⋅ (φ + J )
with χ, y, φ being the partial coefficients that depend on the vegetation, the
geological base and the erosion degree of the watershed respectively whereas j is
the average slope of the watershed area expressed as angle tangent.
It is implied that the values of χ, y, and φ are given analytically in tables by
Gavrilovic. The selections of the suitable values assigned to each case are at the
aptitude of the researcher who applies the formula when needed. By applying the
method, each watershed under research was divided by the authors in spots
depending on:
- the vegetation species prevailing in each part
- the geological base it was consisted of
- the seats of producing debris material.
Therefore, for example, there could be 6 vegetation spots distributed as follows:
2 spots of forest, 1 spot with shrubby plants, 1 spot with pastures, 2 spots with
arid lands. There could also be 4 spots of geological base, 2 of them having
geological base of granite, one gneiss and the other schist and so on. It is implied
that the spots of vegetation, geology, seats of producing debris material are not
identical as far as their boundaries are concerned.
Given that often the spots of vegetation, geology, etc., were numerous, it
becomes clear that this method was knotty, time-consuming and not particularly
accurate, because the Gavrilovic method functions by the same slope for the
whole watershed.
To overcome this insufficiency, we stretched further the Gavrilovic model by
improving the ways of spot separation. The G.I.S. GRASS [11] served greatly
this effort. The following methodology was applied:
In sheets of the G.A.S., the research area was scanned on GIS layers and
converted to image with the use of image processing software.
Then, converted into an appropriate image file format readable by GRASS,
the scanned area was imported in GRASS in raster image format.
Image rectification or registration of the research area was the next processing
step. The rectification was originally carried out by digitization of 4 points, that
is 4 angles, of the area on the HATT grid, and then by conversion of the raster
co-ordinates imported in GRASS (which were in accordance to a local reference
system) into projectional co-ordinates, with the use of linear affine
transformation.
Then, using the scanned rectified map, the research area was digitized [12]
more specifically, digitization involved basin frontiers, peaks and contours (for
peaks and contours the relevant elevation data were provided). Following this,
the spots of vegetation, geology and erodibility were digitized, as suggested by
the Gavrilovic formula on the basis of its x,y and (p. These spots were numbered
and their respective values were provided according to the Gavrilovic tables.
Following the above processing, all the data were exported in vector format and,
after topology building, were converted in raster format, in view of producing all
necessary maps and proceed to the necessary calculations.
More specifically, from the 3 raster of the contours, the Digital Elevation
Models (DEM) were produced, 3D-views of which can be seen on figures 1,2
and 3.
Then, on the basis of these Digital Elevation Models, the slope maps were
created. In order for the DEMs to be accurate, the very same slopes were used,
which of course differed for every cell of the slope raster, i.e. an area of approx.
67 m (8.17m X 8.18m) for the torrents of Fotini and Riganorema and 36 m2 (6m
X 6m) for the torrent of Triantafillia.
Figure 1: 3D View of watersheds Riganorema and Fotini.
Figure 2: 3D View of Triantafilia.
Figure 3: Grey-toned map of potential degradation with legendary

classification (Triantafillia t.).
Then the GRASS software tool r.mapcalc [13] was used; this is a program
that allows users to manipulate, analyze, and create map data by performing
mathematical calculations on raster map layers.
The under process and adjustment model by this way shows an improvement in
the following points:
a. Precision of calculations: By the help of segmentation into cells we
have a complete splitting of the basin in very small areas of 36-67 m2
which are worked out one by one as to the erosion coefficient, which is
impossible without the G.I.S. process'. Besides, the most important
American models such as ANSWERS, AGNPS, etc. run with such sort
of splitting. In this adjustment model the watershed area slope is given
"pointly" thus it is very important for the estimation of erodibility.
b. Ease at entering unlimited parameters: By entering the parameters in
the form of raster maps in various layers, the participation process of
many variables into the same model becomes much easier.
c. Convenience in processing the variable values: By applying the
GRASS tool mapealc, any mathematical elaboration on the above
mentioned raster maps is feasible resulting that the most complex
combinations of parameter values become simple as to their calculation.
d. Appearance: 3D construction provided by the new model improves the
appearance of the "real" picture of watersheds, depicting them as they
are in nature and not as on the map.
4 Research results
The elaboration of the coefficients x, y, <p and the other factors of the method
come up the erosion z coefficients for each watershed. The final coefficient z of
each watershed was the weighed result of partial coefficients z, from which each
one represents the erosion coefficient of a group" of cells. It is evident that all
cells of the same group have the same z,. The groups for each watershed,
depending on the case, were ranging from tens to several hundreds. Meanwhile,
the size of the used analysis through the GRASS is extremely remarkable,
considering that each group may include some tens of cells. In Table 2 the partial
coefficients zx of the cells groups of Riganorema are quoted.
In addition, in Table 3 the values of the final z estimation for all the
watersheds of the sample but also the total respective annual sediment and debris
yield W, are quoted.
The analysis of the results gave the following findings:
The bigger annual degradation is displayed by the Riganorema torrent and the
smallest one by the Triantafillia.
The first one has a geological base of silt-mixed marls and sandstones
whereas the second one has pure granites. Also, the first one has relatively sparse
vegetation of degraded coniferous forests and grasslands while the second one
has extensive forest stands and shrubberies in good condition. Thus, the bigger
average annual production of debris material of the first one in relation to the
second is clearly explained. Finally, the other two watersheds, displaying
intermediate values of degradation, have respective intermediate values % and y

of the vegetation and geology coefficients as well.
Table 2: Partial coefficients z, of the cells groups of Riganorema.
Table 3: Final z values and W values.
Torrents Average
Average annual
Coefficient annual
sediment and debris
No Name z degradation
yield, W (m3/year)
(mm)
1 Riganorrema 0.78 15.82 1.57
2 Fotini 0.66 5.484 1.22
3 Triantafillia 0.75 6.823 0.75
Map and legend study indicates that the areas with light chromatic tones are
the ones with higher risk from erosion and leaching and which need immediate
protection. In other words, we notice that degradation from watershed to
watershed differs quite, and surely in all cases is bigger than the average
degradation given by Kotoulas for the Greek inland and which equals to 0.67
mm. Therefore, we realise that they are watersheds of mountainous character
with a significant discharge of debris material. Finally, in order to have an even
more integrated picture of the z distribution upon all watersheds of the sample,
we drew some maps as follows:
For the construction of maps we classified the values of Z; into 5 categories,
arising from the tables such table 2. This classification has applied a different
colour of grey tone for each category. Thus, surfaces with the same colour tone
corresponding to each particular category zi have arisen. In fact, by this way we

had one more (but much more synoptically) grouping of z, so that to have a final
result which is a coloured map of potential degradations with constant
categorised annotated chromatology (figure 4).
5 Protection system
The protection system recommended in the present paperwork for the erodible
areas, as mentioned earlier, is a mixed system of agro-plant-artificial works. The
agro-plant-artificial works is a relatively old protection technique against erosion
however specialised and with best results usually in the area of hillsides in
respect to area of gullies.
The selection of species and number of works depend on the slope of hillsides
or gullies as well as on the soil type and the aspect. So, the standardisation of
these works is possible (Emmanouloudis [14]) according to the case, when we
are aware of the above factors in an interference area. However, the elaboration
of the Gavrilovic model through the GRASS, in addition to its other mentioned
earlier advantages, provides higher possibility of having in each area the surface
slopes along with the aspect map. Thus, in conjunction to the aforesaid
standardisation, according to Emmanouloudis [14], we may suggest per z
category the following agro-plant-artificial works:
Area of hillsides
Z 0,6-0.8+ Construction of mosaic clusters

with very tolerant species
Z0,4-0,6 Construction of envelopments
Area of micro gullies
Z 0.8+ Construction of picket fence belt

Z0,6-0,8 Construction of envelopments belt
Z0,4-0,6 Construction of clusters belt
6 Conclusions
Summarising all the above mentioned, we drew up the following conclusions:
For the quantitative estimation of mountainous watersheds degradation a new
model was used. The contemplative model of Gavrilovic served as a starting-
point, which has been further stretched out by initiating appropriate adjustments
by the use of the G.I.S. GRASS In the case of the four runoff basins, the average
annual degradation has been calculated with a remarkably high precision due to
this improved model. Actually, it is a combination of the Gavrilovic and the most
known American definition model. Also, an attempt was made so that the
"Balkan" parameters of the first one (which resembles the Greek torrential
environment) and the American G.I.S. detailed techniques are blended. Four
maps of potentional erosion came out after calculating degradation; their study,
along with a standardised system of protection works, can provide an integrated
protection system planning. What's more important, this improved model of
estimation and the protection works system may be applied on any watershed.
References
[1] Line, D., Foster, M., "User's Manual for the watersheds GRASS - AGNPS
modelling tool", N. Carolina State University, U.S.A., 1996.
[2] Beasly, D., Huggins, L., "ANSWERS User's Manual", Purdue University,
U.S.A., 1991.
[3] Rydgren, B. ,"Soil erosion: its measurement, effects and prediction. Case
study from the southern Lesotho lowlands", z.f.Geom. N.F. 40, 429-445,
Dec. 1996, Uppsala., 1996.
[4] Terry, J.P., "Erosion pavement formation and slope process interactions in
commercial forest plantations", n. Portugal, z.f.Geom. N.F. 40, 107 - 115,
Marz 1996, Bunbury, 1996.
[5] Mitas, L., Mitasova, H., Brown, W., and Astley, M., "Interacting fields
approach for evolving spatial phenomena : application to erosion
simulation for optimized land use", National Center for Geog. Inf. and
Analysis, C.A., U.S.A., 1996.
[6] Kotoulas, D., Diefthetisis Himarikon Revmaton , A.U.T., Thessaloniki.,
1997.
[7] Gavrilovic, SI., "Inzenjcring o bujich im tokovima i eroziji", Beograd,
1972.
[8] Emmanouloudis, D., "Natural depositional landforms of the Greek
torrents", Ph.D. Thesis, Dpt. of Forestry and Nat. Environment, A.U.T.,
Thessaloniki., 1990.
[9] Pintar. J., "Grenzen und Moglichkeiten der vorbeugung vor
Unwetterkatastrophen im alpinen Raum", INTERPRAEVENT, Villach,
AUSTRIA., 1972.
[10] Emmanouloudis. D., Filippidis, E., "Torrents basins Morphometric
features calculation through 3-D models", Institute for education and
Technology of Drama. Special Edition, 2001.
[11] Shapiro, M., Westerveld, J, "GRASS user's manual", U.S. Army
Construction Engineering Research Laboratory, Champain, Illinois,
U.S.A. ., 1993.
[12] Neidig, C.A., Gerdes, D., Kos, Ch., "GRASS 4.0 Map Digitizing Manual :
v.digit", U.S. Army Construction Engineering Research Laboratory,
Champain, Illinois, U.S.A., 1991.
[13] Shapiro, M., Westerveld, J., "r.mapcalc. An algebra for G.I.S. and image
processing", U.S. Army Construction Engineering Research Laboratory,
Champain, Illinois, U.S.A. 1992.
[14] Emmanouloudis, D., "The contribution of watershed management to the
integrated flood protection of the Athens basins", 2nd International
Conference of C.N.W.R., Athens., 1994.
Bathymetric curve (75 years old) validation,

using the soil erosion transportation
at Cuitzeo Lake Watershed
J. Lafragua1, A. Gutiérrez1, A. Bahena1, G. Moriel2
& S. Férnández2
1
Mexican Institute of Water Technology, Mexico
2
Secretary of Communications and Transports, Michoacán Center,
Mexico
Abstract
A new main road called Copandaro-La Cinta was built in the state of Michoacán,
México; with a length of seventeen kilometers this road crosses Lake Cuitzeo
(the section under study). In order to preserve the hydrological regime of this
water area, the road structure was built with twenty sewers and four boat
passages. With a 4000 km2 basin area, Lake Cuitzeo’s watershed transports
1,296,461 tons of sediment per year. Due to this great amount of material, a
hydrological study was carried out. A hydraulic study was prepared including
elevation-storage and elevation-area relationship curves since 1930. Using a
longitudinal profile of the study section from April 2003, several bank levels are
used to make an adjustment between the bathymetric curve and longitudinal
profile, fixing some main structures points. In order to select the best alternative
(to verify hydraulic works dimensions), soil erosion in the watershed is
calculated using the Universal Soil Loss Equation (USLE); a sediment yield ratio
is also obtainable. The results show that only 7 percent of the soil loss computed
by the USLE appears as sediment yield in the watershed outlet. In the alternative
selected, this quantity represents approximately 19 cm of sediment near the study
section. With this alternative, the zero level in the longitudinal profile and in the
bathymetric curve is at an elevation of 1823.34 and 1818.00 masl, respectively.
Thus, all points of the bathymetric curve were adjusted at 5.34 m. Flow routing
through Lake Cuitzeo shows that the structures hydraulic capacity was enough to
allow free water flow.
Keywords: soil erosion, hydrologic routing, inflow hydrographs, bathymetric
curve, Lake Cuitzeo.
doi:10.2495/GEO060271
1 Introduction
The new Copandaro-La Cinta main road on the Morelia-Salamanca freeway
crosses the western area of Lake Cuitzeo. This structure has twenty sewers and
four boat passages, and a bridge called Dren La Cinta. In previous years, this
bridge allowed the passage of flows from Lake Cuitzeo to Lake Yuriria. To
check the dimensions of the existent hydraulic works and verify their proper
hydraulic operation it is necessary to consider the volumes of inflow for different
return periods. Lake Cuitzeo’s basin is divided into 25 sub-regions Lafragua et
al., [6]. To carry out the hydraulic flow analysis, a bathymetric curve
fundamental; however, the only one available was for the year 1930 (provided by
the National Water Commission, Conagua). The objective of this study is to
correct the bathymetric curve, taking into account the longitudinal profile of the
road (2003), provided by the Secretariat of Communications and Transport
(SCT). The amount of sediment deposited in the Lake is also required.
Figure 1: Location map of the study site.
2 Study area
Cuitzeo’s watershed is located in the central part of the state of Michoacan,
Mexico, between the coordinates 19°24’ to 20°05’ north latitude and 100°41’ to
101°33’ west longitude, and is part of hydrologic region number 12 called
Lerma-Santiago, figure 1. Total watershed surface is 4,000 km2, of which 409.82
km2 of the Lake include hydrophyte vegetation and open water. The main river is
Rio Grande of Morelia, with a drainage area of 2,043 square kilometers. The
climate is temperate and subhumid with summer rainfall. The annual mean
temperature ranges from 13°C to 20°C, and mean annual precipitation is 804.0
millimeters. The Copandaro-La Cinta new main road crosses Lake Cuitzeo along
seventeen kilometers (section in study) and has twenty sewers and four boat
passages, in addition to La Cinta drain, figure 2.
Figure 2: Hydraulic works location.
3 Procedure description
A bathymetric adjustment was conducted because there is no common level
between the bathymetric curve (Conagua curve) and the longitudinal profile
(SCT profile) of the study section. The procedure was as follows:
a) Drawing a longitudinal profile of the study section with the bathymetric
curve information (Conagua Profile).
b) Analyzing possible alternatives of elevation similarities between the
Conagua profile and the SCT profile.
c) Selecting the best alternative, by using the erosion value of the
watershed outlet.
In order to calculate the annual erosion of Lake Cuitzeo’s watershed, the
Universal Soil Loss Erosion (USLE) equation was used, Ponce [7], given by
A=R K LS C P (1)
where A is the annual mean soil loss in tons ha-1, R (MJ mm ha-1 h-1) is the
rainfall erosivity index, calculated according to Figueroa et al. [4], K (tons ha h
ha-1 MJ-1 mm-1) is the soil erodibility factor, calculated according to the FAO [3]
methodology, C is the cover and management factor, and P is the support
practice factor, assumed as 1.0, Izurieta et al. [5]. L is the slope length factor
and S is the slope gradient, both are known as topographic factors and were
calculated according to Izurieta et al. [5].
m
 λ  (2)
L= 
 22.1 
where λ (m) is the slope length. For this study area, according to Cortes [1],
value λ can be considered as 100 m and value m is the slope-dependent
dimension factor, considered as 0.4.
S = 10.8 sen ( sg ) + 0.3 with s < 9% (3a)
S = 16.8 sen ( sg ) − 0.5 with s > 9% (3b)

where sg and s are the slope of the land expressed in degrees and in percent,
respectively.
With eqn (1), the amount of solid material eroded due to rainfall in one year
is estimated; however, the total volume of material is not necessarily arriving at
the watershed outlet. Escalante [2] declares that there are differences between
soil loss in the watershed and the amount of sediment that the watershed outlet is
receiving, and suggests using the following equation:
AS = DR A (4)
-1
where AS (tons ha ) is sediment in the watershed outlet, A is the value
calculated by eqn (1), and DR is the sediment yield ratio, calculated by,
DR = 0.417662 Ac −0.134958 − 0.127097 (5)

2
where Ac (mi ) is the watershed area.
Table 1: Fixed points.

Alternative Point Characteristic Elevation difference (m)
number (SCT minus Conagua)
A-1 5 High point and located
outside the Lake. 5.83
A-2 41 High point. 4.43
A-3 47 Without major inflow. 5.70
A-4 27 High point and without 5.34
major inflow.
A-5 49 High point and without 5.27
major inflow.

Using the Conagua elevation curve, 55 points were selected, including the 25
hydraulic works, to draw the longitudinal profile of the new main road. Many
alternatives were analyzed by fixing points where erosion was not observed
through time. Table 1 shows five alternatives indicating the fixed point
considered, and table 2 shows the differences between the Conagua profile and
the SCT profile, the rest of the alternatives showed a similar behavior.
Table 2: Selected points and elevation differences.

No. Hydraulic Elevation Elevation
work (SCT) (Conagua) A-1 A-2 A-3 A-4 A-5
1 1827.18 1820.50 -0.85 -2.25 -0.98 -1.34 -1.41
2 1826.01 1820.00 -0.18 -1.58 -0.31 -0.67 -0.74
3 1825.63 1819.50 -0.30 -1.70 -0.43 -0.79 -0.86
4 1824.80 1819.00 0.03 -1.37 -0.10 -0.46 -0.53
5 1824.33 1818.50 0.00 -1.40 -0.13 -0.49 -0.56
6 1823.87 1818.00 -0.04 -1.44 -0.17 -0.53 -0.60
7 1823.86 1818.00 -0.03 -1.43 -0.16 -0.52 -0.59
8 1823.81 1818.00 0.02 -1.38 -0.11 -0.47 -0.54
9 1823.83 1818.00 0.00 -1.40 -0.13 -0.49 -0.56
10 1823.88 1818.50 0.45 -0.95 0.32 -0.04 -0.11
20 1824.38 1820.50 1.95 0.55 1.82 1.46 1.39
21 1824.48 1820.50 1.85 0.45 1.72 1.36 1.29
22 1824.76 1820.00 1.07 -0.33 0.94 0.58 0.51
23 1824.93 1819.50 0.40 -1.00 0.27 -0.09 -0.16
24 1824.56 1819.00 0.27 -1.13 0.14 -0.22 -0.29
25 1824.59 1819.00 0.24 -1.16 0.11 -0.25 -0.32
26 1824.66 1819.00 0.17 -1.23 0.04 -0.32 -0.39
27 1824.85 1819.51 0.49 -0.91 0.36 0.00 -0.07
28 1825.15 1819.86 0.54 -0.86 0.41 0.05 -0.02
29 1825.85 1820.00 -0.02 -1.42 -0.15 -0.51 -0.58
30 1825.59 1820.50 0.74 -0.66 0.61 0.25 0.18
40 VC10 1824.40 1819.62 1.05 -0.35 0.92 0.56 0.49
41 VC11 1824.55 1820.12 1.40 0.00 1.27 0.91 0.84
42 VC13 1824.85 1819.31 0.29 -1.11 0.16 -0.20 -0.27
43 VC14 1825.05 1819.17 -0.05 -1.45 -0.18 -0.54 -0.61
44 VC15 1824.45 1818.99 0.37 -1.03 0.24 -0.12 -0.19
45 VC16 1824.35 1819.00 0.48 -0.92 0.35 -0.01 -0.08
46 VC17 1824.60 1819.03 0.26 -1.14 0.13 -0.23 -0.30
47 VC18 1825.30 1819.60 0.13 -1.27 0.00 -0.36 -0.43
48 VC19 1824.80 1819.70 0.73 -0.67 0.60 0.24 0.17
49 VC21 1825.24 1819.97 0.56 -0.84 0.43 0.07 0.00
50 VC22 1825.24 1819.98 0.57 -0.83 0.44 0.08 0.01
51 PL1 1823.85 1818.02 0.00 -1.40 -0.13 -0.49 -0.56
52 PL2 1823.85 1818.02 0.00 -1.40 -0.13 -0.49 -0.56
53 PL3 1824.60 1819.00 0.23 -1.17 0.10 -0.26 -0.33
54 PL4 1824.85 1819.38 0.36 -1.04 0.23 -0.13 -0.20
55 PEMEX 1824.45 1819.10 0.48 -0.92 0.35 -0.01 -0.08
For each alternative, total sediment area was obtained by considering the
differences shown in table 2. Figure 3 shows the sediment areas of alternative 4,
of 34,800 to 45,504 km. Assuming a uniform distribution along the longitudinal
profile, an average sediment height is estimated, table 3.
In table 3, the negative values indicate that the fixed point was eroded;
therefore, these alternatives were eliminated. The rest of the alternatives have an
average (elevation) height of between 0.18 and 1.5 m. Consequently, we
compared water levels measured by Conagua and SCT during the construction of
the road, obtaining an average difference of 5.18 m. Otherwise, we considered
the incrustation value designated by the SCT (0.60 m), then the points with
differences between 0 and -0.60 m were counted. In alternatives 4 and 5, more

points, 30 and 32 points, respectively, were found. Therefore, alternatives 4 and
5 are good options. In order to select only one alternative, watershed outlet soil
erosion was considered. The variable values for each watershed from eqn (1) are
shown in table 4.
Figure 3: Sediment area along the new main road.
Table 3: Area and average sediment height.

Alternative Total sediment Average sediment
area, (m2) height, (m)
A-1 -5442.41 -0.31
A-2 18784.97 1.08
A-3 -3192.73 -0.18
A-4 3037.17 0.18
A-5 4248.54 0.24
The results of applying eqn (1) and eqn (2) are shown in the table 5. Soil loss
in all watersheds in Lake Cuitzeo is 17,672,146 tons year-1, but only 7 percent
appears as sediment yield at the watershed outlet. Assuming that 1,296,461 tons
of sediment arrived at the Lake per year in average, then in 75 years, the Lake
has received 97 hm3 of sediment. Due to unavailable information, we assume
that the sediment is first deposited into the deeper areas, resulting in 70 cm of
sediment in these areas and 19 cm approximately in the study area, figure 4.
Finally, alternative 4 was selected because it exhibits the value closest to 19 cm,
obtained by calculating the erosion.
In the selected alternative, the zero level in the SCT and Conagua profiles is
at an elevation of 1823.34 and 1818.00 masl, respectively, therefore, every point
of the bathymetric curve was adjusted at 5.34 m. Figure 5 shows the adjusted
bathymetric curve.
Table 4: USLE variables for each watershed.
Watershed Rainfall
number Area (ha) (mm) S (%) R K LS C P
1 1194.11 705.79 17.0 2445.68 0.02 3.91 0.04 1.00
2 2549.69 833.20 16.9 2883.03 0.02 4.09 0.13 1.00
3 1293.56 845.52 16.6 2925.26 0.02 4.22 0.16 1.00
4 2081.71 843.01 15.7 2916.70 0.02 3.81 0.20 1.00
5 17517.45 802.72 13.0 2778.56 0.02 3.03 0.17 1.00
14 4052.99 670.76 7.2 2325.29 0.02 1.41 0.34 1.00

15 4862.76 625.53 6.1 2168.99 0.03 1.11 0.29 1.00
16 5777.69 640.41 4.9 2220.35 0.03 0.83 0.23 1.00
17 7207.63 715.52 10.2 2478.98 0.02 2.01 0.15 1.00
18 2578.24 720.54 18.2 2496.25 0.02 5.09 0.24 1.00
19 2898.54 720.54 17.6 2496.25 0.02 5.02 0.28 1.00
20 6807.55 734.05 12.9 2542.42 0.03 3.16 0.29 1.00
21 10859.16 784.39 13.7 2714.75 0.03 3.25 0.16 1.00
22 707.42 742.85 16.4 2572.35 0.02 3.95 0.10 1.00
23 55985.36 767.67 16.9 2657.86 0.03 4.47 0.21 1.00
24 204294.24 860.61 12.2 2977.92 0.02 2.93 0.23 1.00
25 2513.50 714.65 7.3 2476.29 0.02 1.68 0.42 1.00
Table 5: Erosion results for each watershed.
Watershed Eqn. (1) DR Eqn. (4) Watershed Eqn. (1) DR Eqn. (4)
number (tons ha-1 yr-1) (tons yr-1) number (tons ha-1 yr-1) (tons yr-1)
1 7.39 0.21 1877.97 14 27.47 0.16 17932.76
2 28.19 0.18 12912.61 15 17.77 0.15 13312.72
3 36.75 0.21 9940.18 16 10.67 0.15 9100.33
4 39.67 0.19 15538.25 17 17.36 0.14 17454.16
5 26.89 0.11 51531.12 18 67.23 0.18 31059.16
6 18.06 0.17 9160.92 19 84.92 0.17 42924.66
7 24.06 0.20 7189.23 20 67.41 0.14 64968.17
8 33.96 0.13 39552.37 21 42.58 0.13 57874.83
9 22.46 0.16 16011.22 22 25.87 0.24 4347.56
10 128.75 0.23 25097.48 23 80.84 0.08 339784.28
11 51.41 0.19 17831.81 24 46.85 0.04 408418.37
12 101.81 0.20 33020.48 25 37.71 0.18 17084.97
13 52.54 0.16 32535.29 Total 1098.61 0.16 1296460.88
The adjusted bathymetric curve was used by routing floods with the storm
design of 50 years. We consider tree scenarios: a) inflow from the east and west
during the same time, figure 6, b) only inflow from the east, and c) only inflow
from west. The hydraulic results including the 25 hydraulic works located in the
study site and the sewer located in the freeway are shown in table 6. Maximum
elevation at the end inflow was 1824.53 masl from hydraulic work VC1 to VC5
and 1824.67 masl from VC10 to La Cinta drain. The maximum discharge
obtained in the hydraulic works was 4.94 m3 s-1 and maximum velocity was 1.33
m s-1. In the freeway sewer, the maximum discharge was 11.89 m3 s-1 and
maximum velocity was 2.8 m s-1 Lafragua et al. [6].
Figure 4: Scheme showing sediment deposit into Lake Cuitzeo.

Area, km2
50 100 150 200 250 300 350 400
1825.50
1825.00
Elevation, masl
e
m
lu
1824.50
Vo
1824.00
ea
Ar
1823.50
1823.00
100 200 300 400 500 600 700 800 900
Volume, Mm3
Figure 5: Adjusted bathymetric curve.
5 Conclusions
A bathymetric curve from 1930 was adjusted by using a longitudinal profile of
the study section from 2003, and one sediment yield ratio was used in order to
select the best alternative. All points of the bathymetric curve were adjusted at
5.34 m. Also considered were water levels measured by SCT and Conagua
during the building of the new main road, furthermore, 60 cm from incrustation
designed by SCT was considered.
Figure 6: Input floods for rainfall (50-years).
Table 6: Hydraulic results.
Hydraulic Footing Grade Initial End Initial End Area Discharge Velocity
Work Elevation Elevation Elevation Elevation Depth Depth
2
(masl) (masl) (m) (m) (m) (m) (m ) (m3 s-1) (m s-1)
VC1 1823.95 1826.60 1824.50 1824.53 0.55 0.58 1.10 0.42 0.38
PL1 1823.85 1829.75 1824.50 1824.53 0.65 0.68 20.38 4.71 0.23
VC2 1823.80 1826.60 1824.50 1824.53 0.70 0.73 1.40 0.58 0.42
PL2 1823.85 1829.65 1824.50 1824.53 0.65 0.68 20.38 4.94 0.24
VC3 1823.85 1826.60 1824.50 1824.53 0.65 0.68 1.31 0.41 0.32
VC4 1823.95 1826.60 1824.50 1824.53 0.55 0.58 1.11 0.44 0.40
VC9 1824.55 1826.30 1824.50 1824.71 -0.05 0.16 0.30 0.30 1.03
VC10 1824.40 1826.30 1824.50 1824.67 0.10 0.27 0.35 0.41 1.15
PEMEX 1824.45 1826.30 1824.50 1824.67 0.05 0.22 3.43 1.21 0.35
VC11 1824.55 1826.30 1824.50 1824.67 -0.05 0.12 0.07 0.03 0.41
VC13 1824.85 1826.30 - - - - - - -
VC14 1825.05 1826.30 - - - - - - -
PL3 1824.60 1830.90 1824.50 1824.67 -0.10 0.07 1.90 0.07 0.04
VC15 1824.45 1826.30 1824.50 1824.67 0.05 0.22 0.26 0.24 0.95
VC16 1824.35 1826.30 1824.50 1824.67 0.15 0.32 0.45 0.60 1.33
VC17 1824.60 1826.30 1824.50 1824.67 -0.10 0.07 0.13 0.01 0.06
PL4 1824.85 1830.90 - - - - - - -
LA
CINTA 1823.70 1829.40 1824.50 1824.67 0.80 0.987 24.35 0.28 0.01
Sewer
freeway
(D=5 m) 1823.34 1824.50 1824.64 1.16 1.30 4.13 11.89 2.88
The adjusted bathymetric curve was used by routing floods with the storm
design of 50 years. Flow routing through Lake Cuitzeo shows that hydraulic
structures were enough to permit a free flow through them.
References
[1] Cortés T.H., (2005). Personal communication, 10 October 2005.
Resarcher. Mexican Institute of Water Technology (IMTA).
[2] Escalante S.C., (2005). Capítulo 8. Efecto en la estimación del factor
erosivo de la lluvia en el aporte de sedimentos. En: Rivera-Trejo F.,
Gutiérrez-López A., Val-Segura R., Mejía-Zermeño R., Sánchez-Ruiz P.,
Aparicio-Mijares J, Díaz-Flores L., (Editores). “LA MEDICIÓN DE
SEDIMENTOS EN MÉXICO”. Ediciones IMTA-UJAT, México. 325 p.
ISBN-968-5536-53-8.
[3] FAO, (1980). Metodología provisional para la evaluación de la
degradación de los suelos. Roma. 86 p.
[4] Figueroa S.B, Amante O.A, Cortés T.H, Pimentel L.J, Osuna Ceja E.S,
Rodríguez O.J.M and Morales F.F.J, (1991). Manual de predicción de
pérdidas de suelo por erosión. Subdirección de Conservación del Suelo y
Agua, SARH.
[5] Izurieta J., Huerto D.R., Medina M.R., Cortés T.H., Spillecke W.K.W.,
Brena Z.J. y Castillo R.C., (2002). Estimación del impacto de las cargas
de contaminantes del Dren Zurumútaro en el lago de Pátzcuaro y
propuestas de tratamiento. SGC-UAPS-MICH-02-006-RF-CC, CNA-
IMTA.
[6] Lafragua C.J., Gutiérrez L.A., Báhena H.A., Leal B.G., and Peña P.T.,
(2005). Dimensionamiento de alcantarillas y pasos de lancha, en el tramo
carretero Copándaro-La Cinta, Morelia, Michoacán. IMTA-SCT.
Proyecto TH-0550.
[7] Ponce V.M., (1989). Engineering Hydrology: Principles and Practices.
Prentice Hall.
Efficient watershed modeling using a multi-site

weather generator for meteorological data
M. Khalili, R. Leconte & F. Brissette
Department of Construction Engineering,
École de Technologie Supérieure, Montréal, Québec, Canada
Abstract
The multi-site generation of precipitation data is developed using a Richardson
(1981) WGEN-type weather generator. This approach is based on spatial
autocorrelation to analyze patterns in space and investigate the dependence of
weather data at multiple locations. Reproducing the dependence between
meteorological data at several stations should make the hydrological model
results more realistic. The Chute du diable watershed and surrounding area
located in the province of Quebec, Canada was used to test the proposed
approach. Daily spatial autocorrelations between precipitation occurrences and
amounts were successfully reproduced as well as total monthly precipitation and
monthly numbers of rainy days. A hydrological model has been used to quantify
the natural inflow process. As envisaged, the multi-site generation of weather
data produced more practical natural inflow hydrographs, compared to those
obtained using a uni-site weather generator.
Keywords: weather generator, precipitation, Markov chain, spatial
autocorrelation, hydrological modeling.
1 Introduction
Weather generators can be used to generate climatic data (precipitation,
temperature, solar radiation...) with the same statistical properties as the
observed ones. Most weather generators operate for a single site, e.g. [15], [2]
and [7]. Therefore, they ignore the regional coherence and the spatial
dependence between the stations, which entail many problems in the
hydrological modeling results, obtained using the simulated time series of
meteorological data.
doi:10.2495/GEO060281
A few models were developed for multi-site simulation of weather variables,

in particular for daily precipitations such as space-time model of Bardossy and
Plate [1], the non-homogeneous hidden Markov model of Bellone et al. [3] and
Hughes et al. [11], the nearest-neighbor resampling approach developed by
Buishand and Brandsma [5], and a method based on serially independent but
spatially correlated random numbers developed by Wilks [16].
A regionalization approach based on spatial autocorrelation is proposed to
improve the watershed modeling. It is applied to the Chute du diable watershed
and surrounding area located in the province of Quebec, Canada. This technique
performed successfully in simulating meteorological data and runoff process.
The proposed methodology is presented in section 2 of this article. Section 3
describes the results obtained in the studied basin.
2 Methodology
2.1 Uni-site weather generator
Following the Richardson approach [13], a uni-site weather generator uses a

first-order two states Markov chain to simulate daily precipitation occurrence
X t (k ) at site k on day t. A uniform [0, 1] random number ut (k ) is compared
with a critical probability, which is equal to one of the transition probabilities
depending on the state of the previous day:
 p01 (k ), if X t −1 (k ) = 0
pc (k ) =  . (1)
 p11 (k ), if X t −1 (k ) = 1
A wet day is simulated if the random number is smaller than this critical
probability:
1, if ut (k ) ≤ pc (k )
X t (k ) =  . (2)
0, otherwise
Another uniform [0, 1] random number vt (k ) is used to simulate the synthetic

precipitation amounts by the inversion of the distribution function of amounts. In
the case of an exponential distribution [14], the precipitation amounts rt (k ) can
be computed as:
rt (k ) = − ln (1 − vt (k )) / λ ( k ). (3)
where λ is the parameter describing the exponential distribution function:
f [r (k )] = 1 − exp (− λ r (k )). (4)
2.2 Multi-site weather generator
The Multi-site weather generator used here is based on the concept of spatial
autocorrelation. This theory has been used in a wide array of applications in
which the spatial dependence has to be accounted for, such as social, economic
and physics sciences. Spatial autocorrelation is the correlation between values of
a single variable in geographic space. The analogous statistic to spatial
autocorrelation is serial autocorrelation which is the correlation between values
of a single variable at different time. Spatial autocorrelation can be measured by
statistical indicators such as Moran’s I [10, 12].
n n n n
∑ (xi − x )∑ wij (x j − x )/ ∑∑ wij

i =1 j =1 i =1 j =1
I= . (5)
n n
∑ (x
i =1
i − x) /n 2
∑ (x
i =1
i − x) /n
2
where xi denotes the observed value at location i, x is the average of the xi over
the n locations and wij is the spatial weight between two locations i and j. The
matrix form of Moran’s I contains a spatial weight matrix whose elements are
the weights wij . Generally, these weights are in a row-standardized form, which
means that all weights in a row sum to 1 and by convention wii = 0 . Moran’s I
takes values greater than zero if the geographically nearby observations are
similar, lower than zero if they are dissimilar and equal to zero if these
observations are independent.
The weather generator is modified to simulate precipitation data with daily
spatial autocorrelations, measured by Moran’s I, identical to those observed.
Indeed, the random numbers used in the weather generator, eqns. (2) and (3), are
transformed to spatially autocorrelated ones whose spatial autocorrelations will
reproduce the spatial autocorrelations computed between observed precipitation
series. To generate spatially autocorrelated random numbers, a spatial moving
average process [6, 8] is used:
V = γ × W × u + u. (6)
where V (n,1) is a vector of n spatially autocorrelated random numbers to be used

for n locations.
W (n, n ) is a weights matrix.
u (n,1) is a vector of n independent and uniformly distributed random variables.
γ is the moving average coefficient.
Different γ values provide random numbers with different spatial
autocorrelations and accordingly simulated precipitation processes exhibit
different daily spatial autocorrelations. Therefore, a relationship between the

coefficients γ and the spatial autocorrelations of precipitation occurrences and
amounts can be used to identify the particulars γ , which yield the set of random
numbers reproducing the observed daily spatial autocorrelations of precipitation
occurrences and amounts. These meteorological data will be then used as input
into the hydrological model.
2.3 Hydrological modeling
The HSAMI model [4] is used for hydrological modeling. HSAMI is a lumped
conceptual model consisting of three linear reservoirs in cascades currently used
by Hydro-Québec to forecast and to simulate natural inflows to reservoirs or
runoff at watershed outlets. A set of meteorological data consisting of rainfall,
snow, minimal and maximal temperature and insulation [9] is required to
simulate the hydrological processes. Fig. 1 gives a simplified diagram of HSAMI
[9].
Rainfall and snow evapotranspiration
Interception
According to
ground saturation
and frost
Horizontal flow Vertical flow
Inflow
Reservoir
Figure 1: Simplified diagram of HSAMI [9].

Seven weather stations were used in Chute du diable watershed and surrounding
area in the province Quebec of Canada (fig. 2): 1: Péribonca, 2: Normandin
CDA, 3: Hémon, 4: Bonnard, 5: Chute du diable, 6: Chute des passes, 7: St-
Leon-de-Labrecque. The uni-site and multi-site approaches described above

were used to simulate precipitation data at these stations to be incorporated in the
hydrological model. The uni-site approach simulates the data at each weather
station independently from the others. Thus, the spatial dependence in
precipitation data is ignored. The multi-site approach produces daily spatial
autocorrelations that are identical to those observed at the stations.
Figure 2: Chute du Diable watershed with locations of weather stations.

Fig. 3 gives an example of observed versus simulated monthly numbers of
rainy days at all stations for the month of March and fig. 4 gives the total
monthly precipitation at all stations for the month of September. Similar results
were obtained for all the other months. These curves show a good agreement
between the observed and simulated precipitation occurrences and amounts.
Furthermore, daily spatial autocorrelations of precipitation occurrences and
amounts were well reproduced, as shown by fig. 5 and 6, which illustrate these
results for March and September respectively. Again, similar results were
obtained for the remaining months.
Monthly numbers of rainy days

14
12
10
6 Observed monthly numbers of rainy days

Simulated monthly numbers of rainy days
4
1 2 3 4 5 6 7
Stations used
Figure 3: Observed versus simulated monthly numbers of rainy days at the

seven stations for March.
130
Total monthly precipitations
120
110
100
90
80
Observed total monthly precipitations (mm)
70
Simulated total monthly precipitations (mm)
60
1 2 3 4 5 6 7
Stations used
Figure 4: Observed and simulated Total monthly precipitations at the seven

stations for September.
HSAMI hydrological model requires the mean precipitation over the studied
area. Five weather stations were then selected within and around Chute du diable
watershed according to their proximity to the watershed (stations 1, 4, 5, 6 and
7). An interpolation of precipitation values over this area is required to compute
the mean precipitation using the Thiessen polygons. This operation was done
using the precipitation data simulated by the uni-site and multi-site approaches.
However, the temperature data were simulated only by the uni-site approach, as a
multi-site model of temperature data is currently under development. The
meteorological data were then incorporated in the HSAMI hydrological model.
Five years of meteorological and natural inflow data were used for calibrating
the HSAMI model. Natural inflows were then simulated using the two types of
precipitation data. The simulation results from the input precipitation data
generated using the multi-site approach were found to be more realistic than
using the input precipitation data generated from the uni-site approach.
Typically, the simulated inflow hydrograph from the multi-site approach
displays the spring as well as the late summer and early fall floods (August-
October), while the simulated hydrograph from the uni-site approach indicates
only the spring floods, see fig. 7.
Observed daily spatial autocorrelations

Simulated daily spatial autocorrelations
Daily spatial autocorrelations
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31
March days
Figure 5: Observed and simulated daily spatial autocorrelations of

precipitation occurrences for March.
Observed daily spatial autocorrelations

0.6
Simulated daily spatial autocorrelations
Daily spatial autocorrelations
0.5
0.4
0.3
0.2
0.1
0
-0.1 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
-0.2 September days
Figure 6: Observed and simulated daily spatial autocorrelations of

precipitation amounts for September.
The hydrological model fed with multi-site generated precipitation data was
able to better mimic the observed process because of spatial dependence of
rainfall and snowfall, which was reproduced between the multiple sites in spite
of the lumped nature of the HSAMI. In reality, the climate responsible of the
precipitation process extends at the regional scale rather than at the station
location and constrains the observations in a given station to be correlated to
those in nearby area. Table 1 shows how the multi-site approach improves the
simulation of extreme events. Indeed, the extreme natural inflows from the
multi-site approach (Multi) were more consistent with the observed ones (Obs)
while those from the uni-site approach (Uni) were systematically
underestimated.
1800
1600
1400 sim (multi) sim (uni)
Natural inflows
1200
1000
800
600
400
200
0
1 25 49 73 97 121 145 169 193 217 241 265 289 313 337 361
Days
Figure 7: Example for Multisite and unisite simulated natural inflows.
Table 1: Observed and simulated extreme events by the two approaches.
Natural inflows (m3/s)

Return periods Obs Multi Uni
2 1243.5 1321.45 1022.45
5 1419 1315.1 1211.2
10 1729 1813.55 1339.3
4 Conclusion
The Chute du diable watershed and surrounding area located in the province of
Quebec, Canada is used to investigate the behaviour of hydrological modeling
using uni-site and multi-site generation approaches for precipitation data. The
multi-site approach performed successfully in simulating both precipitation
occurrences and amounts. This result impacts on the hydrological modeling, as
demonstrated using the HSAMI model, which better displayed the late summer-
early fall flood when multi-site approach was used. Other performance criteria
are under investigation to further demonstrate the potential and capabilities of
coupling a multi-site precipitation model to a hydrological model.
References
[1] Bardossy, A. & Plate, E.J., Space-time model for daily rainfall using
atmospheric circulation patterns. Water resources research, 28, pp. 1247-
1259, 1992.
[2] Bardossy, A. & Plate, E.J., Modeling daily rainfall using a semi-Markov
representation of circulation pattern occurrence. Journal of Hydrology,
122, pp. 33-47, 1991.
[3] Bellone, E., Hughes, J.P. & Guttorp, P., A hidden Markov model for
downscaling synoptic atmospheric patterns to precipitation amounts.
Climate Research, 15, pp. 1-12, 2000.
[4] Bisson, J.L. & Roberge, F., Prévision des apports naturels: Expérience
d’Hydro-Québec. Compte-rendu de l’Atelier sur la prévision du débit,
Toronto, novembre 1983.
[5] Buishand, T.A. & Brandsma, T., Multisite simulation of daily
precipitation and temperature in the Rhine basin by nearest-neighbour
resampling. Water resources research, 37(11), pp. 2761-2776, 2001.
[6] Cliff, A.D. & Ord, J.K., Spatial processes: Models and applications.
London: Pion, 1981.
[7] Corte-Real, J., Xu, H. & Qian, B. A weather generator for obtaining daily
precipitation scenarios based on circulation patterns. Climate Research,
13, pp. 61-75, 1999.
[8] Cressie, N.A.C., Statistics for spatial data. Wiley series in probability and
mathematical statistics, John Wiley & Sons, 900 p, 1993.
[9] Fortin, V., Le modèle météo-apport HSAMI: historique, théorie et
application. Rapport de recherche, révision 1,5. Institut de recherche
d’Hydro-Québec (IREQ), Varennes, Qué. 68 p, 2000.
[10] Griffith, D.A., Spatial autocorrelation and spatial filtering: Gaining
understanding through theory and scientific visualization. Advances in
spatial science, Springer, 247 p, 2003.
[11] Hughes, J.P., Guttorp, P. & Charles, S., A nonhomogeneous hidden
Markov model for precipitation occurrence. Applied statistics, 48, pp. 15-
30, 1999.
[12] Odland. J., Spatial autocorrelation. Sage Publications: Newbury Park, 87
p, 1988.
[13] Richardson, C.W., Stochastic simulation of daily precipitation,
temperature, and solar radiation. Water resources research, 17(1), pp.
182-190, 1981.
[14] Todorovic, P. & Woolhiser, D.A., Stochastic model of daily rainfall. Proc.
the symposium on statistical hydrology. Misc. Publ. 1275, U. S. D. A.
Washington, Dc., pp. 232-246, 1974.
[15] Wilks, D.S. Conditioning stochastic daily precipitation models on total
monthly precipitation. Water resources research, 25, pp. 1429-1439,
1989.
[16] Wilks, D.S., Multisite generalization of a daily stochastic precipitation
generation model. Journal of Hydrology, 210, pp. 178-191, 1998.
Hydrological modelling for river basin

management in a highly hydro-geological
conditioned environment
D. Guida, A. Longobardi & P. Villani
Department of Civil Engineering, University of Salerno, Italy
Abstract
Water resources management represents a present key issue in hydrology, and
hydrological models generating streamflow time series are useful tools in this
field. It is possible to refer, in the extreme, to lumped or fully distributed
approaches, but when river basins with particular features have to be modeled it
is possible to take advantage of a semi-distributed formulation. In this study we
propose a semi-distributed conceptually based modeling approach, supported by
field measurements collected within several seasonal campaigns, that has been
set up for the Bussento river basin, located in southern Italy, characterized by
soils and rocks with highly different hydraulic permeability and above all a
highly hydro-geological conditioning. The proposed approach, which joins
together all hydraulic, hydrological and geological data, is able to reproduce the
river discharge mean characteristic.
Keywords: rainfall–runoff model, water resources management, hydro-
geological conditioning.
1 Introduction
Water resources management, more and more limited and poor in quality,
represents a present key issue in hydrology. The development of a community is
highly related to the management of the water resources available for the
community itself and there is a need, for this reason, to rationalize the existing
resources, to plan water resources use, to preserve water quality and, on the other
had, to prevent flood risk.
From this point of view, hydrological models, generating streamflow time
series which are statistically equivalent to the observed streamflow time series,
doi:10.2495/GEO060291
become useful tools. When river basins with particular features have to be
modeled, both traditionally conceptually based models and more recent
sophisticated distributed models appear to give not very reliable results. In those
cases it is possible to take advantage of a semi-distributed formulation, where
every sub-catchment is modeled to account for its features and information
coming from all the sub-catchments are related to each other in order to improve
the system description.
In this study we propose a semi-distributed conceptually based modeling
approach, supported by field measurements collected within several seasonal
campaigns, spanning over two years, that has been set up for the Bussento river
basin, located in southern Italy, well know to hydrogeology and geomorphology
scientists for its karst features, characterized by soils and rocks with highly
different hydraulic permeability and above all an highly hydrogeological
conditioning. The groundwater circulation is very complex, as it will be later
discussed and frequently groundwater inflows from the outside of the
hydrological watershed and groundwater outflows toward surrounding drainage
systems occur. Even though the proposed approach has some similarity with a
few well known conceptually schemes, based on the existence of linear
reservoirs and liner channel to describe the different components the streamflow
can be decomposed in, it is valuable because of the possibility, which is in this
case the necessity, to join all together hydraulic, hydrological and geological data
to achieve reliable results.
2 The Bussento river basin geomorphological and hydro-

geological features
The Bussento River drainage basin, located in southern Italy, Campania Region,
within the Cilento and Vallo di Diano National Park, is well know to
geomorphologist and hydro-geologist for its widely and deeply karst features, as
summit highland with dolines and poljes, lowland with blind valleys,
disappearing streams into sinkholes, cave systems, and karst-induced
groundwater aquifers. The main stream originates from Mount Cervati springs
(1888 m), one of the highest mountain ridge in the Southern Apennine, then it
flows downstream carving steep gorges and rapids, where further springs, along
the streambed, increase the river discharge. The upper right area is characterize
by marly-arenaceous rocks outcrop (M.nt Marchese hilly ridge), while the left
upper area is characterized by limestone sequences (M.nt Rotondo highland and
Serra Forcella). Further down, the Bussento river flows into “La Rupe “sinkhole
channelling the surface flow into karst cave system and emerging four kilometers
downstream, close to Morigerati town, at “Grotta Inferiore del Bussento”. A few
hundred meters downstream, the Bussento river merges with the Bussentino
creek, originating from eastern sectors of the drainage basin, flowing along
canyons and deep gorges carved into Meso-cenozoic limestone sequences.
In the western and southern sectors of the basin (Sciarapotamo creek sub-
basin), marly-argillaceous successions of the Liguride and “affinità sicilide”
Complex (Bonardi G. [5]) dominate the hilly landscape, whereas they underlie
the arenaceous-conglomerate sequences (Guida et al. [6]) at M.nt Centaurino

(1511 m).
Figure 1: Hydro-geological scheme of the region (Celico [2]).
Based on the previous comments and given the presence of a complex

hydroelectric system, the Bussento river basin cannot be considered a simple
drainage basin, but a very complex Hydro-geological System (BHS). A number
of homogeneous hydro-geological sectors can be recognized and outlined within
it (figure 1).
3 Conceptual hydro-geological modelling

Because of the hydrogeological complexity of the BHS previously presented, a
water deep circulation conceptual model has to be outlined before any
hydrological computer-aided model can be built. The conceptual model is a
preliminary physical-based model, accounting for an interconnected sequence of
geologic substrates, permeability distribution, recharge areas and discharge
points, that collectively provide a physical scheme of the recharge system, the
storage system and the routing system.
Karst aquifers modelling is not an easy task. Anderson and Woessner [3]
indicate “karst” as one of the advanced topics in the groundwater researches, and
summarize a few attempted models, none of which produces reliable results. One
of the earliest attempt of karst aquifers conceptualization (White [9]) focused on

the variety of geologic settings and their controlling influence on groundwater
flow patterns. This scheme was later expanded (White [11]) to take into account
the overall area of the groundwater system. Based only on the type of
permeability, Shuster and White [8] divided aquifers into “conduit flow”
aquifers, which contain well developed conduit systems, and “diffuse flow”
aquifers which do not. Following White [10], the basic components of the
generic karst aquifer flow system can be sketched as in figure 3. Clearly, not all
of these components are present in all aquifers, and their presence and relative
importance is a fundamental point of distinguishing one aquifer from another.
With reference to Iaccarino et al. [7] and White [10], this general conceptual
model has been applied to the Bussento Hydro-geological System (BHS),
recognizing the following recharge-discharge components (figure 2).
Allogenic Recharge Diffuse

infiltration
Internal
runoff
Surficial infiltration
sinking stream
infiltration
Epikarst
Depression Vadose
Deep sinking Zones Quick returns
stream infiltration overflow springs
Delayed returns
Conduit System
Aquiclude underflow springs
Bedrock
Deep
stream Deep fracture
groundwater
infiltration system
flow
Aquiclude
Deep
groundwater
flow losses
toward the sea
Figure 2: Specific conceptual model of the karst aquifers in the Bussento

Hydrological System (BHS).
Four sources of recharge for karst aquifers can be recognized: i) allogenic

recharge: as surface water collected on outside aquifer basin and injected into the
aquifer via sinking streams developed on surrounding aquiclude; ii) internal
runoff recharge: as overland flow comes into closed depressions (dolines and
poljes) where it enters the aquifer through sinkhole drainage; iii) diffuse
infiltration recharge: as precipitation on the land surface, where it infiltrates
through the soil and rock, remaining for days or weeks in the vadose epikarst
zone, before it migrates downward, through the rock fractures, into the
percolation zone, finally reaching the water table into the saturation zone; iv)
recharge from fractured bedrock streams: as perched groundwater systems above
carbonate aquifers, in which water reaches the main aquifer by means of vadose
shafts and open fracture systems along the margins of the perched aquifers and
carbonate fractured bedrock streams. A distinguishing feature of Bussento karst
aquifers is that most of the groundwater is discharged through a small number of

large springs.
Figure 3 illustrate a simplified scheme of the Bussento river network
indicating, on a planimetric point of view, the complex interaction between the
diffuse springs system, which generally determine an increase in river discharge,
and the hydroelectric system, i.e. the anthropic impact, which generally
determine a decrease in river discharge, retained and diverted within dams,
artificial lake and weirs for human water uses.
+10
-10
(-30)
1
+50
(-60)
2
+100
+30 +200
(-20)
3
+30 +200
+2500
+20
(-20)
9
(-3000)
4
[+600]
5
-3000 -50 -10
+100
+50
+200 (-30)
8
(-600)
5
+200
+10
(-500)
6
-100 (-600)
+100 (-50)
7
+5
[+9000] +1000
10
+100
+50
+10
+100
+2000
Policastro Gulf
Figure 3: Bussento river network and interactions with the springs system
and the hydroelectric system.
4 Semi-distributed coupled hydro-geological and hydrological

modelling
4.1 The monitoring campaigns
Currently no working river flow discharge measurement stations are present over
the catchment. Two stations were actually working, for a short period, over the
decade 1960-1970, thus very short streamflow time series are indeed available.
For this reason, on January 2003 the Regional Water Basin Authority, Sinistra
Sele, started a monitoring campaign with the aim of measure in many different
sections and on a monthly time scale the Bussento river discharge. Based on the
Bussento catchment (313 km2) geomorphological and hydro-geological features
described in the previous paragraphs, 13 stations were indicated as significant to
individuate the river regime (figure 4). The monitoring campaigns is currently in
progress managed by CUGRI, but so far only a two years periods has been
analyzed. In the following paragraphs some results relative to collected data

analysis, particularly referring to modelling applications, are given.
4.2 Data analysis and modelling results
Given the Bussento catchment geomorphological and hydro-geological features

described in the previous paragraphs, a lumped model cannot guarantee reliable
results. For this reason and taking advantage of the dense monitoring campaign,
a semi-distributed formulation, accounting for each sub-basin particular
characteristics, seems to be more appropriate.
B_AS01
###
B_AS06 B07
B06
#
B_AS04
#
B05
#
B_AS02
#
B04 # B_AS03
#
B_AS05
B03 #
B02 #
#
B01
#
Figure 4: The Bussento river basin monitoring network.
When dealing with the monthly time scale each sub-basin can be described
(figure 5) as two linear reservoirs in parallel, representing the groundwater flow
and the deep subsurface flow, whereas the rainfall contributes which are
characterized by delay times smaller then a month are supposed to reach the
outlet through a linear channel (Claps et al. [1]). The scheme is also supported by
the conceptual hydro-geological model described in the previous paragraph. In
this case coupling the linear reservoirs balance equations with the whole system
balance equation, total streamflow D at each time step is related to the net input
by means of an ARMA (2,2) model, which stochastic formulations corresponds
to:
D( t ) − Φ 1 D( t − 1 ) − Φ 2 D( t − 2 ) = ε ( t ) − Θ 1ε ( t − 1 ) + Θ 2ε ( t − 2 ) (1)
where ε is the model residual, related to the net input I, that is then a periodic
independent random process, and Φ1, Φ2, Θ1, and Θ2 are the model stochastic
parameters, related to the model conceptual parameters K1, K2 (reservoirs
response times), a and b (recharge coefficients).
In its original formulation the model algorithm, starting from an observed
streamflow time series, estimates the model parameters and, because of the
univariate approach, with an inverse procedure, the net rainfall input.
Collected streamflow data consist of mean monthly values, measured at each
section, over two years. These are not enough to run the ARMA(2,2) model with
its inverse procedure.
It
a It b It (1-b) It
GROUNDWATER DEEP
FLOW SUBSURFACE
FLOW
Vt
Wt
a It fk Ck Vt-1 b It f q Cq W t-1
Dt
Figure 5: Linear system of monthly streamflow time series.
To set up a modelling approach able to reproduce observed discharge values,

we estimated a priori model parameters and net rainfall input. Equation (1) as
been then used to generate 1000 years monthly streamflow time series at each
section, comparing thus the discharge probability distributions, at each section
and for each month, with the occurred values.
The response times have been evaluated from streamflow collected data,
applying the base flow recession equation:
Q( t ) = Q0 e − t / K (2)
whereas the recharge coefficients corresponding to each section, i.e. the outlet of
a sub-basin, have been initially assigned on the basis of the relative sub-basin
hydrological and hydro-geological features and successively modified, with an

iterative procedure, to achieve the better model results (table 1).
Table 1: Model parameters, response times and recharge coefficients.
Section K1 (days) K2 (days) a 1-a

B_AS 05 - 11.07 0.00 1.00
B 04** 149.71 7.37 0.70 0.30
B 04 153.66 119.75 0.70 0.30
B_AS 03 211.12 10.16 0.70 0.30
B_AS 02 81.21 25.42 0.30 0.70
B 01 280.94 31.12 0.60 0.40
B 02 280.00 32.61 0.70 0.30
B 03 293.69 46.62 0.80 0.20
B_AS 04 108.36 30.51 0.70 0.30
B 06 86.12 35.77 0.70 0.30
B 07 75.11 18.69 0.20 0.80
B_AS 01 - 30.07 0.00 1.00
B_AS 06 301.36 - 1.00 0.00
With regard to the model rainfall net input, the procedure we pursued was to
generate it from its probability density distribution, with given parameters. The
I(t) probabilistic representation is the Bessel distribution, which is the sum of a
Poissonian number of events with exponentially distributed intensity:
P [ I = 0 ] = e −υ I =0
(3)
−λI −υ
fI ( I ) = e ( υλ / I )ℑ1 [ 2( υλI ) 1/ 2
] I >0
where λ=1/β is the exponential parameter, υ is the Poisson parameter and

ℑ1 ( x ) is the modified Bessel function of order 1. The rationale for such
probabilistic representation is given by the positive values and finite probability
at zero that I(t) has to present.
Parameters β and υ are estimated from the existing two streamflow time
series. The temporal patterns found for the two series are rather similar, thus we
assumed β and υ spatially invariant over the catchment (table 2).
Table 2: Net rainfall input distribution β and υ parameters.
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
υ 1.05 1.19 1.21 1.39 1.48 1.82 2.07 1.66 1.28 1.09 1.04 1.03
β 5.29 2.65 1.52 1.43 6.33 0.73 0.30 0.86 2.85 2.69 1.77 3.63
Figure 6 shows, as an example, the generated streamflow probability

distribution, at section B07 for august. Measured values during august 2003 and
august 2004 correspond respectively to the 55 percentile and to the 98 percentile.
Model performance in 2004 is generally poor because of particular climatic
conditions occurred in that year.
0.999
0.99
0.96
0.90
0.75
0.50
0.25
0.10
Probability
August 2003
August 2004
0.05
0.02
0.01
0.003
0.001
1
10
river discharge l/sec
Figure 6: Generated streamflow probability distribution compared with

occurred values over two years.
Similar results have been found at the remaining sections, with model
adequacy being affected by the degree of anthropic impact and hydro-geological
conditioning over the corresponding sub-basin.
5 Conclusion
In this paper we highlight the modeling difficulties that have to be faced when
basin with particular features have to be investigated. The case study is the
Bussento river basin, located in southern Italy, which is well know to
hydrogeology and geomorphology scientists for its karst features, characterized
by soils and rocks with highly different hydraulic permeability and above all an
highly hydrogeological conditioning. In this case, a lumped methodology cannot
guarantee satisfactory results but a semi-distributed formulation can be more
appropriate. The methodology we have presented has some similarity with a few
well known conceptually schemes, based on the existence of linear reservoirs
and liner channel to describe the different components the streamflow can be
decomposed in, but it join all together hydraulic, hydrological and geological
data. In particular, the hydro-geological catchment’s features have been used to
drive the identification of the monitoring sections, i.e. the extension of each sub-
basin, and to assign a priori model parameters. Overall, the proposed modeling
approach has a good performance, even though it appears in some sections
inadequate, acknowledging the poor available database and the complexity of the
system we have tried to model.
Acknowledgements
The authors wish to thank the Regional Water Basin Authority, Sinistra Sele, and
the CUGRI, Centro Universitario per la Previsione e Prevenzione dei Grandi
Rischi, for their support. The research was partially supported by MIUR and
MURST grant.
References
[1] Claps, P., Rossi, F. & Vitale, C., Conceptual- stochastic modelling of
seasonal runoff using autoregressive moving average models at different
scales of aggregation. Water Resources Research, 29(8), pp. 2545–2559,
1993.
[2] Celico, P. B., Schema idrogeologico dell’Appennino Meridionale; Mem e
Note Ist. Geol. Appl., 19 Napoli, 1978.
[3] Anderson, M.P. & Woessner, W.W., Applied Ground Water Modelling.
Academic Press, San Diego, 1992.
[4] Bonardi, G., Ciampo, G. & Perrone, V., La formazione di Albidona
nell’Appennino calabro-lucano: ulteriori dati stratigrafici e relazioni con
le unità esterne appenniniche. Boll. Soc. Geologica Italiana, 104, Roma ,
1985.
[5] Bonardi, G., Ciampo, G. & Perrone, V., La formazione di Albidona
nell'Appennino calabro-lucano: ulteriori dati stratigrafici e relazioni con le
unità esterne appenniniche. Boll. Soc. Geologica Italiana, 104, Roma ,
1988.
[6] Guida, D., Iaccarino, G. & Perrone, V., Nuovi dati sulla successione del
Flysch del Cilento nell’area di M.te Centaurino: relazioni fra Unità
Litostratigrafiche, Unità Litotecniche e principali Sistemi Franosi”. Mem.
Soc. Geol., 41, 1988.
[7] Iaccarino, G., Guida, D. & Basso, C., Caratteristiche idrogeologiche della
struttura carbonatica di Morigerati. Mem. Soc. Geologica Italiana, 41,
1065-1077, Roma, 1988.
[8] Shuster , E., T. & White, W., B., Seasonal fluctuations in the chemistry of
limestone springs. A possible means for characterizing carbonate
acquifers. J. of Hydrology, 14, 93-128, 1971.
[9] White, W. B., Conceptual models for limestone acquifers. Groundwater, 7
(3), 15-21, 1969.
[10] White, W. B., Conceptual model for karstic acquifers. Speleogenesis and
Karstic Aquifers - The virtual Scientifical Journal, 1, 1-6, 2002.
[11] White, W. B., Conceptual models for carbonatee acquifers: revised. In
Dilamarter, R. R. and Casallany, S. C. (Eds): Hydrologic Problems in
Karst Terrain. Western Kentucky University, Bowling Green, KY, 176-
187, 1977.
[12] White, W. B., Conceptual model for karstic acquifers. Speleogenesis and
Karstic Aquifers - The virtual Scientifical Journal, 1, 1-6, 2003.
Hydrological modelling of snow cover in

the large upper Po river basin:
winter 2004 results and validation with
snow cover estimation from satellite
D. Rabuffetti, A. Salandin & R. Cremonini
Arpa Piemonte, Italy
Abstract
The study of the hydrological budget of mountainous river basins requires the
understanding of the annual cycle of snow accumulation and melting. In fact a
deep knowledge of snow cover distribution and dynamic offers several
possibilities to improve water resource management and exploitation. The
implemented model reproduces an energetic budget of the snowy mantle at small
scales using DEM resolution. The use of a distributed model accounts for the
high time-space variability of meteorological factors, such as precipitation, air
temperature and solar radiation, whose fields at soil level are reconstructed from
spot measurements through interpolation procedures based on the topography of
the river basin. The modelled variables are the snow water equivalent (SWE) and
the discharge generated by the snowmelt while the mechanics of the snowy
mantle like thickness and density are not considered. At the same time the
simplified degree day method, that uses only air temperature as an index to melt
is also implemented and compared. The propagation of melting water inside the
snowy mantle is modelled through the conceptual linear reservoir approach; the
outflow from snowy mantle is propagated as superficial run-off to the closing
section of the river basin through the Muskingum-Cunge hydrological model.
The basin of the river Po closed at the section of Ponte Becca is studied in the
simulation; it covers a surface of approximately 38000 km2. The simulation
ranges from winter 2004 to spring 2005. The modelled snow cover over the
catchment is compared with estimation from satellite, derived using the
Normalized Difference Snow Index (NDSI), concurring to the validation of the
model.
Keywords: snow, energy budget, degree day, satellite images, model validation.
doi:10.2495/GEO060301
1 Introduction
In hydrology, the study of accumulated water as snow and its successive release
during melting season plays particular importance either in the long period
budget and in short flood simulations. In the first case, the knowledge of SWE
dynamics allows one to understand the availability of water resource stored in
snow, and contributes to improve its management policy. For example, with
regard to engineering applications, one can point out the useful information that
snow models give for the management of hydroelectric or irrigation reservoirs.
In the second case stored water in snowy mantle reduces the available volumes
for floods and their peak discharge; flood wave estimation relative to meteoric
phenomena in melting season are also improved.
This work aims to study the dynamics of the snowy mantle exclusively from a
hydrological point of view, with the hypothesis that the parameters of the
constituent equations stay constant and that wind doesn’t affect significantly the
distribution of snowy mantle in space.
In particular a numerical tool for the calculation of SWE and the relative
processes of melting and propagation of the outflow is implemented. The
validation of the model is obtained through local verifications (qualitative
comparison between the measure snowy height and SWE modelled) as well as
through the comparison with MODIS satellite images of snow cover.
An important feature of this work is linked to the space scale of interest. The
large Po catchment and its varied characteristics with the high Alps surrounding
the Padana plain represent a significant test case for the hydrological snow
model.
2 The hydrological snow model

The application of distributed models is based on the solution of the equations
describing the physical phenomena at local level in particular for each
elementary unit in which the river basin is subdivided; this allows to describe the
hydrological answer to meteorological variables either at local and at catchment
scale. The model of the snowy mantle essentially concerns the following
processes: accumulation of the snow on the ground, it’s melting and the
propagation of the melting water inside of the snowy mantle. The snow model
described has been then included into the hydrological distributed model FEST,
Mancini et al. [1], for the application at catchment scale.
2.1 Accumulation model
Precipitation measured by rain-gauges can be distinguished by the use of the

atmospheric temperature into rain and snow, Tarboton et al. [2].
Pl = α P P
(1)
Ps = ( 1 − α P )P
where P is the total precipitation volume (millimeter), Pl and Ps are respectively

liquid and solid precipitation, αP the fraction of liquid precipitation. αP value is
calculated with the successive equations
α P = 0 →
IF
Ta ≤ Tinf
Ta − Tinf
αP = →
IF
Tinf ≤ Ta ≤ Tsup (2)
Tsup − Tinf
α P = 1 →
IF
Ta ≥ Tsup
where Ta represents air temperature, Tinf and Tsup are two threshold temperatures
to be determined during model calibration.
2.2 Energy budget
The sources of energy that drive snowmelt include both shortwave and longwave
net radiation, convection from the air (sensible energy), vapour condensation
(latent energy), conduction from the ground, and energy income with rainfall;
these fluxes are labelled Qsn , Qln , Qh , Qg , Qp respectively. The energy budget
equation that describes the energy available for snowmelt is given in eqn (3); the
total energy available for snowmelt is Qm (kJ/m2), ∆Qi is the increment of
internal energy stored in the snow per unit area of snowpack
Qm = Qsn + Qln + Qh + Qe + Qg + Q p − ∆Qi (3)
In this work the snowpack is considered single-layer with constant medium

temperature equal to 0°C. This is equivalent to consider always null variations of
the snowy mantle internal energy; this hypothesis is generally considered
acceptable in the alpine environment where the thickness of the snowy mantle is
elevated for long periods. Sensible and latent energy can be often ignored
because quantitatively less important than radiative exchange; moreover this also
avoids to estimate wind speed and humidity fields which are affected by high
approximation at river basin scale. The amount of snowmelt M (mm of water
equivalent) may be expressed by
Qm
M = (4)
334.9 ρ w B
B is the thermal quality of the snow (ratio of heat required to melt a unit weight
of snow to that of ice at 0°C), 334.9 (kJ/kg) is latent heat of ice melting, and ρw
is the density of water.
2.3 Degree day method
In order to carry out a correct estimation of the snowmelt with a full physical
approach, as seen in 2.2, it is necessary to know the time and space variability of
the heat flows as well as of the characteristics of the snow. For large scale
applications, the possible lack of accuracy in the necessary information has
carried to the development of models that operate with only an index that is a
variable used in order to explain a physical phenomenon in integral way. In the
calculation of the energetic budget of the snow, the temperature carries out a
predominant role and it’s often used like index of snowmelt.
The use of the degree-day in the valuation of the energetic contribution to the
snowmelt is originally due to Martinec [3]. The melting rate (m/sec) is
proportional to the measured temperature described by the introduction of the
coefficient Cm
M s = C m (Ta − Tb ) (5)
where Tb and Ta are respectively base and air temperature. In this work, the value
of Cm is kept constant during all the day and the entire simulation. Base
temperature and Cm values are obtained from calibration. Generally coefficient
Cm ranges from 4.8*10-8 to 6.9*10-8 m/(°C s) and base temperature is 0°C.
2.4 Propagation model
To avoid the complex solution of the equations that describe the motion of the
water inside the snowy mantle one can decide to use a conceptual approach, i.e.
the linear reservoir cascade. The propagation of liquid fraction of SWE in every
cell is modelled as a linear tank; the outgoing flow is proportional, through the
use of a constant of time, τ, to the volume of the liquid member in the snowy
mantle. The value of the constant τ is determined from the value, commonly
found in literature, of the average speed of the water inside of the snowy mantle,
equal to 6 (m/h).
The propagation of the water inside of the mantle is based on the solution of
the mass budget
dhl
= Qin + Qs − Qu (6)
dt
where hl the height of the liquid fraction of the tank, Qin is the flow coming from
the upstream cells,
Qu = τhl (7)
it’s the flow directed to the downstream cell and
Qs = Pl + SM (8)
it’s the source term that is the sum of the water produced from the snowmelt and
the liquid part of the precipitation in that cell. When the propagated flow gets up
to the limit of the snowy mantle, it is added to overland flow.
3 The Po river basin

3.1 Geographical data
In this study the Po river basin closed at Becca bridge covers a total area of
38000 km2 including the Aosta Valley and some areas in Liguria, Lombardia and
Switzerland. It is situated on the Padana plain and bounded on three sides by
mountain chains. Hydrologically speaking, the Upper and Mid Po catchment is
varied. The mountainous areas of the Alps provide a complex hydrological
regime. In the winter, the Alps are covered in snow and most of the precipitation
is stored as snow and glacial depth. In the spring the snow melts, aided by
rainfall, which can result in the spring floods.
The adopted DEM mesh of 1 km subdivide the river basin in square cells with
time-invariant characteristics of the territory such as elevation, aspect and
inclination; moreover every cell is characterized with a value of precipitation,
temperature and solar radiation obtained from the interpolation of site
measurements.
3.2 Meteorological data
There are over 300 rainfall-temperature stations providing measure depths at 30

minute intervals. The model creates hourly rainfall depth surfaces which are
constructed from the rainfall data interpolated using the I.W.D. (inverse weight
distance) method. To avoid the well known problem of precipitation
underestimation due to very low temperatures only gauges with elevation lower
than 2000 m asl are considered.
The temperature surface maps are constructed from the spot data with
adjustment to consider its dependency on elevation. In particular spot
measurements are all referred to the same level (1000 m asl) by a constant lapse
rate (6.5°C/km) and then interpolated. Finally the temperature of the generic cell
is calculated on its real altitude again through the use of the same lapse rate.
The calculation of the short waves incident radiation on the single cell
happens through a simplified model of the phenomena of transmission,
absorption and atmospheric spread also consider the geographical characteristics
of the territory, D. Rabuffetti et al. [4].
Over 60 snow-gauge stations provide a measure of height at 30 minute
intervals and they are used for validation at local scale.
3.3 Satellite data
The MODerate resolution Imaging Spectroradiometer (MODIS) instrument

provides high radiometric sensitivity (12 bit) in 36 discrete spectral bands
ranging in wavelength from 0.4 µm to 14.4 µm, Salomonson and Toll [5]. A
±55-degree scanning pattern at the EOS orbit of 705 km achieves a 2330-km
swath and provides global coverage every one to two days. Snow has strong
visible reflectance and strong short-wave IR absorbing characteristics. The
Normalized Difference Snow Index (NDSI), defined as
( TMBand 2 − TMBand 5 )
NDSI = (9)
( TMBand 2 + TMBand 5 )
is an effective way to distinguish snow from many other surface features. Pixels
that are approximately 50% or greater covered by snow have been found to have
NDSI values about 0.4,[6]. Snow cover over eight days is mapped as maximum
snow extent in one SDS and as a chronology of observations in the other SDS
stored in HDF-EOS format. Eight-day periods begin on the first day of the year
and extend into the next year. The product can be produced with two to eight
days of input. Using HDFlook-Modis utility, the MODIS/Terra snow products
have been extracted and it have been remapped in UTM European Datum 50, to
compare with ground observation. The images obtained have a resolution of
approximately 400 m, different from the snow model resolution. For the
comparison we report these images to model resolution of 1 km considering that
each cell is covered by snow if it is so for more than 30% of its surface.
4 Results
The simulation starts from 1-10-2004 until 31-03-2005. Since the interest is to
evaluate performance of the model for the annual cycle of snow accumulation
and melting, it is assumed that, at the starting time, the snowy mantle is present
nowhere. Two validation approaches are used: first the comparison with snow
gauges, second with snow cover images obtained from MODIS satellite. It must
be underlined that in both cases only a qualitative verification of the model is
possible because either snow gauges and satellite provide just an indirect
estimation of SWE. Anyway it is important to highlight the very different scales
in which these two approaches allow investigating.
Figure 1 shows the qualitative comparison in the station of Formazza, taken
as an example, with both models. Simulation look fairly good picking the main
periods of the snowy mantle accumulation and melting.
Figure 1: Comparison between local measure of snowy height and modelled

SWE with the different melting models.
For a general overview, the correlation coefficients (ρ) between observed

snowy height and predicted SWE at gauge site have been calculated for some
representative stations all around the region.
Table 1: Local validation.
Energy budget model Degree day model

Station ρ ρ
Alpe Veglia 0.91 0.48
Capanne Marcarolo 0.71 0.93
Colle Lombarda 0.80 0.60
Formazza 0.84 0.94
Formazza Bruggi 0.92 0.88
Lago Agnel 0.49 0.70
Larecchio 0.89 0.82
Monviso 0.04 0.78
Passo del Moro 0.89 0.93
In general correlation values show a good performance of both models, even

if for few sites (Alpe Veglia, Lago Agnel, Monviso) important differences
among the models are present, probably due to gauges problems.
As far as satellite images are concerned the periods used for the validation are
chosen on available data; in particular clear sky periods are considered. Data
used covers autumn with high snow precipitation in the mountains, winter with
very low precipitation and spring with focus on heavy snowy day in the
lowlands. As an example, Figure 2 shows the result of the comparison between
model and observation is expressed by means of a contingency matrix.
Some classical indices based on this contingency matrix are calculated to
assess model performances: threat score (TS) is the statistical measure of
accuracy taking into account the numbers of missed and false predictions (the
model accuracy improves as the threat score comes closer to 1); bias (B)
measures prediction over/under-estimates of snow cover (overestimation for bias
scores major than 1 and vice versa); hit rate (HR) represents model capability to
predict snow covered areas (best for hit rat equals to 1), Murphy and Winkler
[7].
Even though the energy budget model obtains local results as good as the
degree day model, it proves to be less satisfactory at distributed level. Table 2
gathers the result for the different analysed periods and the energetic model
budget model is better only for the period from 6/3/2005 to 13/3/2005, where a
heavy snow fall was registered in the low lands.
To allow a deeper understanding of model behaviour, statistics were
calculated focusing on the mountain areas (i.e. elevation greater than
1000 m asl).
Results resumed in table 3 show a good performance of both models with a
better behaviour of the degree day model. This can be addressed to a high
melting rate showed by the energy budget model mainly in the north-eastern
areas.
Period: 2005-01-25 – 2005-02-01
Energy budget Degree Day
Period: 2005-03-06 – 2005-03-13

Energy budget Degree Day
Figure 2: Contingency matrix obtained with two models.
Finally, comparing tables 2 and 3, one can clearly notice that model
performance in mountain is definitely enhanced than in the whole catchment.
This highlights a poor capability to represent snow dynamics in the lowlands,

where a general quick melting is observed in the results.
Table 2: Contingency tables statistics.

Period (8 days) TS B HR TS B HR
01/01/2005 0.59 1.49 0.93 0.65 1.00 0.79
25/01/2005 0.63 1.53 0.97 0.70 1.27 0.93
02/02/2005 0.58 1.55 0.93 0.67 1.04 0.82
26/02/2005 0.65 1.32 0.92 0.65 1.36 0.93
06/03/2005 0.57 1.06 0.75 0.55 0.92 0.68
14/03/2005 0.56 0.96 0.71 0.67 1.02 0.81
Table 3: Contingency tables statistics for mountain.

Period (8 days) TS B HR TS B HR
01/01/2005 0.78 1.24 0.98 0.74 1.07 0.88
25/01/2005 0.68 1.41 0.98 0.72 1.25 0.94
02/02/2005 0.62 1.47 0.95 0.69 1.04 0.83
26/02/2005 0.76 1.20 0.95 0.78 1.27 0.99
06/03/2005 0.72 1.13 0.89 0.76 1.29 0.99
14/03/2005 0.57 0.95 0.71 0.67 1.02 0.81
5 Conclusion
The application of hydrological snow model at distributed level on the large Po
catchment give satisfactory results; nevertheless both the melting models
adopted need improving. Probably the characteristics of the snowy mantle, i.e.
the albedo dynamics for the energy budget model and coefficient Cm for the
degree day model, change during the winter season and require a better
description.
Local comparison with snow gauges allow a sound validation of the model
and shows that where meteorological forcing is known one can reproduce the
snowy mantle dynamics with precision either with a physical description of the
processes involved either with the degree day conceptual approach.
The use of snow cover derived from satellite images offers a further and
important means for validation and for understanding model behaviour.
Looking at the entire catchment scale, the simplified degree day model picks in a
slightly better way the complex dynamics of the snowy mantle in spite of the
energetic budget model that needs much more parameters and data.
Finally it is important to emphasize that the model results can be used itself
for a better exploitation of satellite image. In cloud covered areas satellite cannot
give any information: a jointly usage of model and images can certainly produce
better estimates of snow cover.
Future developments will be addressed to the extension of the analysis to a
longer period in order to obtain more significant statistics. An on-line prototype
application will also provide a sound evaluation of practical usefulness of this
kind of products.
References
[1] Mancini, M., Montaldo, N. & Rosso, R., La modellazione distribuita nella
valutazione degli effetti di laminazione di un sistema d’invasi artificiali
nel bacino del fiume Toce, L’acqua, pp. 31-42, 2000.
[2] Tarboton, D. G., Chowdhury, T. G. & Jackson, T. H., A spatially
distributed energy balance snowmelt model, Utah Water Research
Laboratory, 1994.
[3] Martinec, J., The degree-day factor for snowmelt runoff forecasting. Proc.
of general assembly of Helsinki commission on surface waters, IASH
publication n° 51, 1960.
[4] Rabuffetti, D., Salandin, A., Volontè, G. & Mancini, M., Modellazione
idrologica del manto nevoso. Il caso del lago epiglaciale del ghiacciaio del
Belvedere sul Monte Rosa. Atti del XXIX Convegno di Idraulica e
costruzioni idrauliche, vol. 3, pp. 867-874, 2004.
[5] Salomonson, V.V., & Toll, D.L., The moderate resolution imaging
spectrometer-radar (MODIS-N) facility instrument, Advances in Space
Research, 11, pp 231-236, 1991.
[6] MODIS Snow Products, http://modis-snow-ice.gsfc.nasa.gov/sugkc2.html
[7] Murphy, A. H. & Winkler, R. L., A general framework for forecast
verification, Monthly Weather Review, 115, pp. 1330-1338, 1997.
Section 8
Landscape analysis
Spatial correlograms and landscape metrics

as indicators of land use changes
R. Aunap, E. Uuemaa, J. Roosaare & Ü. Mander
Abstract
Land use changes over time can be analysed in several ways. We studied the
spatial autocorrelation (Moran’s I) of raster format land use maps from three
different time periods (1900, 1940, and 2000) in 13 study areas representing
most of the landscape regions in Estonia. Human influence was taken into
consideration in compiling a scale of the contrast between 10 land use groups.
We introduce a simple characteristic based on spatial correlograms: a half-value
distance lag, hI=0.5 – a distance where Moran’s I drops below 0.5. No significant
change was detected in values of hI=0.5 over time. In addition, we did not detect a
difference between lowlands and heights. In analysis of landscape metrics Edge
Density (ED), Patch Density (PD), Contrast Weighted Edge Density (CWED),
Mean Patch Area Distribution (AREA_MN), and Percentage of Like
Adjacencies (PLADJ) showed significant changes comparing the year 2000 with
1900 and 1940. However, the results showed no significant change in landscape
metrics between 1900 and 1940. ED, PD and CWED had higher values in 2000
than in 1900 and 1940. Therefore landscape heterogeneity has increased in recent
decades. ED, PD, CWED, AREA_MN and PLADJ metrics also indicated a
significant difference between lowlands and heights. It appeared that heights
have a more heterogeneous landscape structure than lowlands. Generally, the
heterogeneity of Estonia’s landscapes overall has changed within recent decades.
Keywords: FRAGSTATS metrics, landscape pattern, landscape regions, land use
change, Moran’s I, spatial autocorrelation.
1 Introduction
Studies on land use changes is the basic area in landscape research [1, 2] being
one of the key issues in global environmental change [3]. Both natural and socio-
doi:10.2495/GEO060311
economic factors have been used in the analysis of land use changes [4, 5]. In the
majority of them, the main problem is to correctly characterize the spatial pattern
using various landscape metrics [6, 7].
Widely used means to describe landscape texture metrics can be calculated
with the help of FRAGSTATS [8]. It has been shown [9, 10] that these metrics
are scale-dependent, and not all indexes demonstrate a regular behaviour in
relation to scale changes.
Despite the great number of indexes, FRAGSTATS does not include
measures of variography, e.g. different spatial structure functions such as
correlograms and variograms, which are popular in geostatistics [11], and
describe the dependence of variability on distance. In order to study landscape
heterogeneity and spatial autocorrelation, correlograms are preferred to
semivariograms, since – according to Legendre and Fortin [6] – they are
standardized and make it possible to compare different landscapes. This method
is used by Radeloff et al. [12] to study artificial landscapes with a regular
pattern, in order to detect periodicity in their correlograms. The behaviour of a
correlogram’s wavelength and amplitude within a specific range of spatial orders
can be used as an indicator of spatial pattern [13].
A classical estimator of spatial dependence is Moran’s I, “associated with
statistician P.A. Moran (1948)” and proposed as the spatial analogy of
autocorrelation used in time series analysis [14]. Since the introduction of the
autocorrelation index by Moran [15], spatial correlograms have been used for the
spatial analysis of several natural and social phenomena. At present, various
studies use Moran’s I correlograms to avoid systematic mistakes due to spatial
autocorrelation in spatial analyses [6, 16, 17, 18], or using the correlograms for
ecological and landscape analyses at different spatial scales. Koenig [19] showed
the importance of the Moran effect and spatial autocorrelation (environmental
synchrony) for the analysis of patterns of animal populations at continental and
global scale. Likewise, Diniz-Filho et al. [20] calculated the Moran’s
autocorrelation value when analysing avian populations at continental scale.
Large-scale (100-1000 km) analysis of Moran’s I correlograms was carried out
for the investigation of vegetation pattern dynamics in the Great Lakes region
during the Holocene [21], for the prediction of deforestations in Saskatchewan
[22], for the analysis of human land transformations in South African avian
diversity [23], for the identification of operational units for conservation in
continuous populations [24], and for the analysis of alien plant invasion in
Catalonia [25]. At the local and landscape scale (10-1000 m), Moran’s I was
used for analyses of the distribution pattern of several bird and mammal species
[26], carabid beetles [27], Neotropical migrant songbirds [17], the impacts of
logging in Amazonian forests [28], urban spatial features [29], the variability of
soil properties in wetlands [30], and the spatial patterns of greenhouse gas
emissions in tropical rainforests [31].
Surprisingly, we were able to find only two papers in which Moran’s I
statistic has been used in connection with changes in land use/cover [32] or
vegetation cover [33]. Read and Lam [32] have found that Moran’s I is effective
at detecting changes in land cover types in Costa Rica’s lowlands, whereas

FRAGSTATS indexes were not sufficient in this sense.
The main objectives of this study were: (1) to analyse three map series (from
approximately 1900, 1940, and 2000) of selected landscape areas in Estonia
concerning their differences in spatial autocorrelation and FRAGSTATS
indexes; (2) to find out whether the Moran’s I characteristic and landscape
indexes respond to the land cover changes.
2 Material and methods
2.1 Study areas
Thirteen study areas were selected on the basis of Estonian landscape regions so
that they represent most Estonian landscape types (Fig. 1). The selection of study
sites was based on Uuemaa et al [34].
Figure 1: Study areas and landscape regions of Estonia.
Of 13 study areas, there were sites dominated by agricultural land use, forests,
bogs or urban areas. Study areas were formed on the basis of Estonian Basic
Map Sheets. Each study area was 5*5 km.
2.2 Land use data
Land use data was derived from three maps from different time periods: 1:42,000
(from Russian topographic map sheets dating from 1886-1917; later referred to
as “1900”), 1: 50,000 (topographic map sheets published by the Estonian
Military Topo-Hydrographic Department, in 1935-1939; later referred to as

“1940”), and 1:20,000 (Estonian Basic Map sheets from 1998-2004; generalised
to 1:50,000 scale; later referred to as “2000”) and converted into raster format
using 10 m pixel size.
1 2 3 4 5 6 7 8 9 10
Bogs, fens
forests
Paludified
Forests
Bushes
Grasslands
Arable lands
forests
Clear-cut
Mining areas
Urban areas
Water
Figure 2: Scale of the contrast of main land use types.
For the generalisation of the 1:20,000 Estonian Basic Map sheets into the
1:50,000 scale, we used the MapInfo tools Polygon Area Thinning and Gap
Removal. The minimum recognisable area was set to 0.4 ha.
The maps of the test sites were scanned (except for the Estonian Basic Map,
which is already in digital form), digitised, and rasterized. Ten land use types
distinguished from all map series were reclassified so that new type numbers
could be used as contrast indexes (Fig. 2). In the case of land use types Mi and
Mj, their difference (|i-j|) shows the contrast between these types.
2.3 Moran’s I
In our analysis we applied the Idrisi Kilimanjaro software [35]. A module named
AUTOCORR calculates the first-lag autocorrelation coefficient of an image. The
following equation, which is similar to the usage in other software, is applied:
n n
∑∑ w ( y
i =1 j =1
ij i − µ )( y j − µ )
I = n⋅ (1)
(  n
∑i≠ j ∑ ij  ∑
w ) 
( yi − µ ) 2 
i =1 
where n – number of values to be taken into account (in the case of a raster
image, pixels); w – spatial weights: 1 in the directions up/down/left/right,
0.70711 (square root of 2) as a weight for the diagonal neighbouring pixels; yi/j –
value of pixel i resp. j; µ – mean of values y [36].
In addition to Moran’s I, Idrisi also calculates several statistics including tests
of significance under two null hypothesis assumptions.
For raster images, the autocorrelation has been calculated with all appropriate
pixels using so-called King’s case analysis [35].
Using auxiliary images (the CONTRACT module with so-called pixel

thinning), we computed the 1st, 2nd, 3rd etc. lags of Moran’s I value. Accordingly,
we found all necessary Ih, changing h as a multiple of pixel resolution. In our
investigation, the size of the pixel side was 10 m, and for the test sites we
calculated series of Ih, h=10, 20, 30, …, 100, 120, …,200, 300, 400, 500 and
1000 m. We used the results to construct the graphs of I(h), called correlograms,
which ideally are monotonically decreasing curves. Since n is very large (tens
and hundreds of thousands), all I(h), except for some I(1000), are statistically
significant.
We investigated the correlograms of the test areas based on 3 map series from
different periods, and found these to be quite regular. In order to compare
Moran’s I correlograms from different test areas and different map series and
also with other landscape metrics derived from FRAGSTATS, we introduced a
simple characteristic of half-value distances: hI=0.5 – the distance lag where
Moran’s I drops below 0.5 [34].
2.4 Landscape metrics
Using FRAGSTATS 3.3, landscape metrics were calculated for all study sites.
We calculated the following landscape metrics: 1) Edge Density (ED); Patch
Density (PD); Mean Patch Area Distribution (AREA_MN); Mean Shape Index
(SHAPE_MN), Contrast Weighted Edge Density (CWED); Percentage of Like
Adjacencies (PLADJ); Contagion (CONTAG) and Shannon’s Diversity Index
(SHDI). For details and metrics formulae see McGarigal and Marks [8].
2.5 Statistical analysis
According to the Kolmogorov-Smirnov test for normality, all of the variables

under consideration were normally distributed. The homogeneity of variances
was verified using the Cochran C and Levene tests. In correlogram analysis,
when comparing different groups we used a one-way ANOVA (Tukey’s HSD
test). In landscape metric analysis, most of the variables did not meet the analysis
of variance assumptions (Levene and Cochran C tests). Therefore, the
significance of differences was analysed using the non-parametric Kruskal-
Wallis test. For the statistical analysis of all data, the computer program
STATISTICA 7.1 was used. The level of significance of α = 0.05 was accepted in
all cases.
3.1 Moran’s I correlograms
We did not detect statistically significant land use changes during the years
1886-2004. The overall significance of the model was >0.3, and also Tukey’s
HSD test did not show significant differences between group means. Therefore
we could say that landscapes overall have not changed significantly over the past
100 years in Estonia. However, Figure 3 shows that the mean of the hI=0.5
decreased from 250 in 1900 to 160 in 2000 (Fig. 3). Also, the variation is smaller
in 2000. Therefore we could say that human influence has made landscapes
slightly more heterogeneous. We could even claim that land use in some areas
did change dramatically. In Alutaguse paludified lowland, for example, the
forests and bogs were turned into mining areas during the last 50 years (Fig. 4).
However, the heterogeneity of the landscape has decreased (Fig. 5). In 2000 the
spatial autocorrelation is highest, and in 1940 lowest. In the case of West Estonia
it is vice versa (Fig. 5). In recent decades the spatial autocorrelation has
decreased, i.e. human influence has increased the heterogeneity of landscape. In
the case of all heights except Vooremaa, the spatial autocorrelation had
decreased in recent decades. Landscapes were more heterogeneous in 2000 than
they were a hundred years ago.
450 450
400 400
Mean Mean±SD Mean Mean±SD
350
350
300
300
250
I h=0,5
I h=0,5
250
200
200
150
100 150
50 100
0 50
1990 1940 2000 lowlands heights
Figure 3: Average and standard deviation values of Moran’s Ih50 of 3 map

series (1900, 1940, 2000), and heights and lowlands over all 13 test
sites.
1900 1940 2000
Figure 4: Land use change in the test area of the Alutaguse 2 paludified
lowlands. For numbers of land use types, see Table 1.
We also found that there is no significant difference between Ih50 in heights

and lowlands as could be expected based on the results obtained by Uuemaa et
al. [34]. The Tukey HSD test did not show a statistically significant difference
between heights and lowlands. Nevertheless, it can be seen from Figure 3 that
the mean of the Ih50 is 155 in the case of heights and 250 in the case of lowlands.
The variation of Ih50 is also smaller. This shows that the spatial autocorrelation of
heights is lower than the autocorrelation of lowlands, i.e. lowlands are more
spatially homogenous. This can also be seen from Fig. 5, where the decrease of
correlograms is more abrupt in the case of heights (Vooremaa and Otepää) than
in the case of lowlands (Alutaguse 2 and West-Estonia).
Alutaguse 2 Vooremaa
1 1
0.8 1900 0.8 1900
1940 1940
Moran's I
Moran's I
0.6 0.6
2000 2000
0.4 0.4
0.2 0.2
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000
lag(m) lag(m)
West -Estonia Otepää

1 1
0.8 0.8 1900

1940
Moran's I
Moran's I
0.6 0.6
2000
1900
0.4 0.4
1940
0.2 2000 0.2
0 0
0 200 400 600 800 1000 0 200 400 600 800 1000
lag(m) lag(m)
Figure 5: Spatial correlograms of 3 map series (1900, 1940, 2000) from 4 test
areas.
Interestingly, we found that near 180m and 400m of lag distance, a “jump”
appeared on almost each correlogram (Fig. 5). This phenomenon is probably
related to the periodicity of Moran’s I correlograms, which was, however, only
detected at larger scales [12].
3.2 Landscape metrics
Although no change in Ih50 values appeared in the analysis of correlograms,

statistically significant changes were detected in several values of landscape
metrics.
The average values and standard deviations in the years 1900 and 1940 are
very stable (Table 1). In 2000, remarkable changes can be detected in values of
PD, ED, AREA_MN, CWED and PLADJ. These metrics also gave statistically
significant changes in the Kruskal-Wallis test (Table 2). No change was detected
between 1900 and 1940, which also confirms the results of Palang et al., [37].
ED, PD and CWED had increased in 2000 compared to earlier dates (Fig. 6).
This shows that heterogeneity has increased in recent decades. AREA_MN and
PLADJ had decreased in recent decades, which also indicates the increase in
landscape heterogeneity.
We also tried to identify differences between heights and uplands. The results
showed that PD, ED and CWED values are significantly lower in the lowlands
(Table 2 and Fig. 6), i.e. the landscape structure is more homogenous than in the
heights. Palang et al. [37] obtained similar results: in southern Estonia, where all
the heights are located, test sites had higher heterogeneity. AREA_MN and
PLADJ had lower values in heights (Table 2), i.e. landscape patches are smaller,
and therefore the landscape has a more complex structure than in the lowlands.
Table 1: Average values and standard deviations of measured landscape

metrics. *significant difference with 2000 (p<0,01).
1900 1940 2000
PD 2.66±1.78* 3.26±1.75* 15.03±7.15

ED 46.467±20.73* 51.94±19.75* 100.83±28.85
AREA_MN 56.97±42.56* 38.58±18.63* 8.20±3.81
SHAPE_MN 1.88±0.28 1.76±0.12 1.81±0.24
CWED 19.43±7.57* 22.47±7.65* 44.01±12.96
CONTAG 63.384±7.86 60.14±6.02 61.82±6.44
PLADJ 97.48±1.04* 97.20±0.99* 94.76±1.44
SHDI 1.3±0.30 1.39±0.22 1.28±0.22
Table 2: Changes in landscape metrics agewise and difference between

lowlands and heights. Estimated by Kruskal-Wallis test
(significance levels ***p< 0.001; **p<0.01; *p<0.05).
Agewise Lowlands and heights
PD 1900 and 1940<2000*** lowlands<heights*

ED 1900 and 1940<2000** lowlands<heights**
AREA_MN 1900 and 1940>2000*** lowlands>heights*
SHAPE_MN no difference no difference
CWED 1900 and 1940<2000** lowlands<heights*
CONTAG no difference no difference
PLADJ 1900 and 1940>2000** lowlands>heights**
SHDI no difference no difference
Summarizing the results of this study, the heterogeneity of Estonian

landscapes has changed over the last 50 years, according to landscape metrics
analysis. However, the spatial autocorrelation of landscapes has not changed
significantly, but it also showed decreasing trends in recent decades. The results
of this analysis were not unequivocal, because in some study areas the
heterogeneity had increased, and some study areas had become more
homogeneous over time. Thus the overall change in Estonian landscapes is not
so remarkable.
160 160
Median 25%-75% Median 25%-75%
140 140
Min-Max Min-Max
120 120
100 100
ED (m/ha)
ED(m/ha)
80 80
60 60
40 40
20 20
0 0
1900 1940 2000 lowlands heights
Figure 6: Median, quartiles, maximum and minimum values of ED of 3 map

series (1900, 1940, 2000), and heights and lowlands over all 13 test
sites.
4 Conclusions
The results of the study demonstrated that the average value of spatial
autocorrelation in Estonian landscapes has not significantly changed over time.
We were also unable to find a significant difference between spatial
autocorrelation in heights and lowlands. We propose the distance (lag) of spatial
correlograms at which the Moran’s I value reaches 50% of the maximal value
(hI=0.5) as a new landscape metric for the characterization of landscape pattern. Its
benefit is its simple interpretability and the independence of the scale. Thus this
characteristic can effectively be used as an indicator in landscape planning and
management.
Although the analysis of correlograms did not show significant change over
time, several landscape metrics indicated that landscapes are more heterogeneous
in 2000 than they were in 1900 or 1940. There was also a statistically significant
difference between heights and lowlands, latter being more homogeneous.
Acknowledgements
Science of Estonia. We acknowledge Sten Mander for his help in digitizing
maps.
References
[1] Forman, R.T.T., Godron, M., Landscape Ecology, John Wiley and Sons,
New York, 619 p., 1986.
[2] Zonneveld, I., Land Ecology, SPB Academic Publishing, Amsterdam, 199
pp., 1995.
[3] Alcamo, J., Leemans, R., Kreileman. E., Global Change Scenarios of the
21st Century, Results from the IMAGE 2.1 Model, Elsevier Science
Publishers, London, UK, 287 pp., 1998.
[4] Wear, D.N., Bolstad, P., Land-use changes in southern Appalachian

landscapes: Spatial analysis and forest evaluation, Ecosystems, 1(6), 575-
594, 1998.
[5] McDonald, R. I., Urban, D.L., Spatially varying rules of landscape
change: lessons from a case study, Landscape Urban Plan. 74(1), 7-20,
2006.
[6] Legendre, P., Fortin, M.-J., Spatial pattern and ecological analysis,
Vegetatio, 80(2):107–138, 1989.
[7] Gustafson, E.J., Quantifying landscape spatial pattern: What is the state of
the art? Ecosystems 1(1), 143-156, 1998.
[8] McGarigal, K., Marks, B.J., FRAGSTATS: spatial pattern analysis
program for quantifying landscape structure, USDA Forest Serv. Gen.
Tech. Rep. PNW-351, 1995.
[9] Wu, J., Shen, W., Sun, W., Tueller, P. T., Empirical patterns of the effects
of changing scale on landscape metrics, Landscape Ecol. 17(8), 761-782,
2002.
[10] Uuemaa, E., Roosaare, J., Mander, Ü., Scale dependence of landscape
metrics and their indicatory value for nutrient and organic matter losses
from catchments, Ecol. Indic. 5(4), 350-369, 2005.
[11] Cressie, N.A.C., Statistics for Spatial Data, Revised Edition. Wiley, New
York, 1993.
[12] Radeloff, V.C., Miller, T.F., He, H.S., Mladenoff, D.J., Periodicity in
spatial data and geostatistical models: autocorrelation between patches,
Ecography 23(1), 81-91 2000.
[13] Chou, Y.H., Spatial pattern and spatial autocorrelation, Lect. Notes
Comput. Sc. 988, 365-376, 1995.
[14] Taylor, P.J., Quantitative methods in geography, Houston Mifflin
Company, Boston 1977.
[15] Moran, P.A., The interpretation of statistical maps, J. Royal Stat. Soc. B
10, 243–251 1948.
[16] Legendre, P., Spatial autocorrelation: Trouble or new paradigm? Ecology,
74(6): 1659-1673, 1993.
[17] Lichstein, J.W., Simons, T.R., Shriner, S.A., Franzreb, K.E., Spatial
autocorrelation and autoregressive models in ecology, Ecol. Monogr.
72(3), 445-463, 2002.
[18] Overmars, K.P., de Koning, G.H.J., Veldkamp, A., Spatial autocorrelation
in multi-scale land use models, Ecol. Model. 164(2-3), 257-270 2003.
[19] Koenig, W.D., Global patterns of environmental synchrony and the Moran
effect, Ecography 25(3), 283-288, 2002.
[20] Diniz-Filho, J.A.F., Bini, L.M., Hawkins, B.A., Spatial autocorrelation
and red herrings in geographical ecology, Global Ecol. Biogeogr. 12, 53-
64, 2003.
[21] Graumlich, L.J., Davis, M.B., Holocene variation in spatial scales of
vegetation pattern in the upper Great Lakes, Ecology 74(3), 826-839,
1993.
[22] Hobson, K.A., Bayne, E.M., van Wilgenburg, S.L., Large-scale

conversion of forest to agrculture in the boreal plains of Saskatchewan,
Conserv. Biol. 16(6), 1530-1541, 2002.
[23] Fairbanks, D.H.K., Kshatriya, M., van Jaarsveld, A.S., Underhill, L.G.,
Scales and consequences of human land transformation on South African
avian diversity and structure, Anim. Conserv. 5, 61-73, 2002.
[24] Diniz, J.A.F., Telles, M.P.D., Spatial autocorrelation analysis and the
identification of operational units for conservation in continuous
populations, Conserv. Biol. 16(4), 924-935, 2002.
[25] Pino, J., Font, X., Carbo, J., Jove, M., Pallares, L., Large-scale correlates
of alien plant invasion in Catalonia (NE of Spain), Biol. Conserv. 122(2),
339-350, 2005.
[26] Koenig, W.D., Knops, J.M.H., Testing for spatial autocorrelation in
ecological studies, Ecography 21(4), 423-429, 1998.
[27] Judas, M., Dornieden, K., Strothmann, U., Distribution patterns of carabid
beetle species at the landscape-level, J. Biogeogr. 29(4), 491-508, 2002.
[28] Read, J.M., Spatial analyses of logging impacts in Amazonia using
remotely sensed data, Photogramm. Eng. Rem. S. 69(3), 275-282, 2003.
[29] Myint, S.W., Kam, N.S.-N., Tyler, J.M., Wavelets for urban spatial
feature discrimination: Comparisons with fractal, spatial autocorrelation,
and spatial co-occurrence approaches, Photogramm. Eng. Rem. S. 70(7),
803-812, 2004.
[30] Bruland, G.L., Richardson, C.J., Spatial variability of soil properties in
created, restored, and paired natural wetlands, Soil Sci. Soc. Am. J. 69,
273-284, 2005.
[31] Ishizuka, S., Iswandi, A., Nakajima, Y., Yonemura, S., Sudo, S., Tsuruta,
H., Muriyarso, D., Daniel Spatial patterns of greenhouse gas emission in a
tropical rainforest in Indonesia, Nutr. Cycl. Agroecosys. 71(1), 55-62,
2005.
[32] Read, J.M., Lam, N.S.-N., Spatial methods for characterising land cover
and detecting land-cover changes for the tropics, Int. J. Rem. S. 23(12),
2457-2474, 2002.
[33] Landis, A.G., Bailey, J.D., Reconstruction of age structure and spatial
arrangement of piñon-juniper woodlands and savannas of Anderson Mesa,
northern Arizona, Forest Ecol. Manag. 204(2-3), 221-236, 2005.
[34] Uuemaa, E., Roosaare, J., Kanal, A., Mander, Ü., Spatial correlograms of
soil cover as an indicator of landscape heterogeneity, Ecol. Indic.
Submitted, 200X.
[35] Eastman, R., IDRISI Kilimanjaro: Guide to GIS and Image Processing,
Clark University, 2003.
[36] Single Raster Analysis Tools VII,
http://www.sbg.ac.at/geo/idrisi/wwwtutor/s_tools7.htm [24.11.05], 1999.
[37] Palang, H., Mander, Ü., Luud, A., Landscape diversity changes in Estonia,
Landscape Urban Plan. 41(3-4), 163-170, 1998.
The role of geosciences and landscape in the

management of Natural Parks of Guadalajara
(Central Spain): in search of suitable
applications
A. García-Quintana, M. P. Abad, M. Aguilar, L. Alcalá, I. Barrera,
M. Cebrián, M. C. Fernández de Villalta, J. F. García-Hidalgo,
R. Giménez, A. E. Godfrey, J. A. González-Martín, A. Lucía,
J. F. Martín-Duque, M. Martín-Loeches, E. Quijada,
J. M. Rodríguez-Borreguero, R. Ruiz López de la Cova & A. Solís
PANAGU Project, Complutense University, Spain
Abstract
University professors, researchers and professionals from different governmental
bodies in Madrid and Guadalajara have formed a research group to investigate
how to use geo-environmental information in the management of Spanish
Natural Parks. Funding has been obtained from the Spanish Ministry of Science
and Technology. The pilot area for the project is the Natural Parks of the
province of Guadalajara in the Castilla – La Mancha Autonomous Region.
This article describes two proposals to analyse the physical environment and
the landscape of the natural parks of Guadalajara: analytical-parametric
inventory, and synthetic-physiographic inventory. The objective is to provide
useful information on how to manage these areas. Additionally, the article
includes results of the study of active karstic processes in two of these natural
parks, Río Dulce and Alto Tajo.
Keywords: natural parks, landscape, environmental geology, karstic processes,
Guadalajara.
1 Introduction
Landscape and geology are often the most important factors leading to the
designation of an area as a national or natural park. This has been the case since
the beginning of the national park movement in North America (Yellowstone
doi:10.2495/GEO060321
1872, Waterton Lakes 1885) and in Spain (Picos de Europa, Ordesa and Monte
Perdido, 1918). However, once a park is established, geological and landscape
value tends to be relegated to a secondary role in the daily management of such
places. Growing interest in the preservation of fauna and flora during the second
half of the 20th century, the ambiguity of the term “landscape”, and a lack of
general knowledge about geology on the part of Mediterranean societies have all
been obstacles to parks’ maximizing and managing these resources.
Faced with this reality, studies on geology and landscape appear to be
undervalued and underutilised in the management of natural places in many
countries, among them Spain. This limits the chances that societies will protect,
enjoy, and value these areas. Nevertheless, the information from such studies can
provide: knowledge about active processes, conceptual tools, classifications, and
territorial compositions, predictions, evaluations, etc.
This perceived deficit is the starting point for the PANAGU project.
Landscape is the “meeting point” of the project’s component parts, as it is the
most conspicuous and highly valued element in most natural parks. We believe
that analysing geology and geomorphology can provide a well-informed
understanding of landscape. Landforms (bedrock, topography and processes)
affect microclimate conditions, soils, vegetation, land use and the visual
composition of many of the natural parks of the Iberian Peninsula. For this
reason, although the research team is led by geologists and geomorphologists,
other specialists on landscape are also involved.
2 Geological and geographical setting

The province of Guadalajara is situated on the Meseta, a region composed of
wide barren plains and mountain chains. This morphological unit occupies the
greater part of the Iberian Peninsula; it has an average altitude of 600 m.a.s.l. and
constitutes the Peninsula’s dominant geographic element. Three types of
landscape stand out: mountains, open plains, and gorges cut into the plains. The
gorges are the result of the big difference in altitude between the central zone of
the Meseta and sea level. The three natural parks of Guadalajara offer fine
examples of these landscapes. The province of Guadalajara is shaped by the
three large regional geological units that intersect there: the Central System, the
Iberian Chain and the Tajo Basin.
The Central System is a reactivation and uplift of the ancient Variscan
Massif, composed predominantly of metamorphic and igneous rocks. The Tejera
Negra is located in this Central System. It is quite small (1,600 ha) and was
created to protect a relic beech tree (Fagus selvatica) forest, although its name is
due to the presence of stands of yews (Taxus baccata, in Spanish tejeras). The
park has relatively simple topography shape and a fairly uniform geological
composition. It is characterised by sharp peaks and crests that rise above 2000
m.a.s.l., sculpted from quartzite and black slate of Ordovician age. It is poorly
developed, far from large population centres, and has low human impact. At
present, there is a proposal to considerably expand the park’s size due to the
beauty and untouched nature of much of the surrounding area.
The Iberian Range is a foreland mountain chain, with a very thick

sedimentary layer predominantly of carbonate rocks. Due to intense erosion
during the Neogene, today it appears as a set of barren plains from which some
residual mountain chains jut out. These mountains control the shape of the
stream network. The Río Dulce and the Alto Tajo are situated in the Iberian
Range. Both parks are characterised by narrow, deep canyons and river gorges
which in places can drop to 500 m below the plains and plateaus, located at
1,100 m above sea level.
The Río Dulce Natural Park (literally, “Sweetwater River”, because it does
not have a high percentage of salts, unlike other rivers in this region which have
been exploited for their salts since the days of the Roman Empire because they
drain salty Triassic formations) is small in size (8,300 ha) and, although its main
landscape types are gorges and plains, it has a more complex topographic
structure. The plains have different shapes and land uses as they are developed
on substrates of different lithologies and structures. The same can be said of the
Rio Dulce and its tributaries. This has led to a diversity of landscapes. There has
also been greater human impact, thanks to its location near the historic city of
Sigüenza, today a magnet for cultural and gastronomic tourism. Since the Middle
Ages, the area has been well travelled, first by two drovers’ routes and royal
roads that passed close by. Today it is flanked by a national highway, a regional
highway, and it can be traversed by a local road.
The Alto Tajo natural park (literally, “High Valley” of the “Cut River”) is
larger (106,000 ha). It constitutes one of the most unique natural areas on the
Iberian Peninsula so its designation as a national park is under consideration. The
park is famous for its deeply incised fluvial network. There are more than 100
km of spectacular canyons, deep gorges, and narrow valleys with steep slopes,
some covered by thick forests. To this scientifically and aesthetically interesting
geo-morphologic setting one must add geological, plant and ethnographic
formations that make the park a valuable national heritage site: unique large
sedimentary structures, juniper groves (Juniperus thurifera), and “pallozas”
(primitive dwellings of pre-Roman technology).
Figure 1: Location of the three natural parks within the regional geological
units of Guadalajara and the Iberian Peninsula. DB, Duero Basin; IR,
Iberian Range; CS, Central System; TB, Tajo Basin; EB, Ebro Basin;
BR, Betic Range; GB, Guadalquivir Basin; HM, Hercynian
(Variscan) Massif, CPR, Cantabrian Pyrenees Range.
3 The landscape study of the natural parks of Guadalajara

The primary and operative objective of the PANAGU Project is to carry out an
analysis of the landscape of the protected natural areas of the province of
Guadalajara (Castilla-La Mancha). This analysis is first and foremost
cartographic and digital and is being performed following two concepts.
3.1 Analytical or parametric method
The analytical or parametric method examines a series of maps and thematic
reports, whose partial results are integrated to establish the true relationship
between different elements of the environment.
3.1.1 Lithology and tectonic structure

In this research, rocks are studied not so much for their own sake but rather as to
how they condition ecosystems, human uses and landscape. Therefore, using a
standard 1/50,000 geological map, lithological maps were compiled that grouped
rock units with a high degree of physical and chemical homogeneity, and whose
ecological and visual response were also homogeneous. A subsequent evaluation
of these units allowed the capabilities and limitations of use to be established as
well as their relationships with other environmental factors.
For each scenic landscape site, the size, orientation and form of each
topographic element, as well as the surface area, intensity and type of land uses
and the placement of human settlements and developments can be explained in
the litho-structural configuration of the different plots in the territory. Each body
of rock provides at least one distinguishing property to the landscape: colour,
alteration, texture, permeability, forms of erosion, etc. The tectonic structure
affects the park’s landscape on a larger scale, such as: surface forms, folded
areas with irregular relief, valleys oriented along faults, etc.
Detailed analysis of the parks’ territory by a large group of experienced
professionals revealed geological formations that should be of aesthetic interest
to the public, and of scientific interest as national treasures, but which at the
moment are underappreciated. These include ancient valleys with hanging
formations, karst systems cut by modern valleys, prehistoric landslides, etc.
3.1.2 Relief and topography

The study of landforms plays a fundamental role in the study of landscape. From
a strictly visual standpoint, landforms mould both the viewpoints and the view
shed. Within the framework of the study of “natural scenarios”, it seems
necessary to make an objective classification of the landforms and their
characteristics if we want to be able to measure, correlate and study a landscape.
In so doing we facilitate comparison of the results of the different studies.
The way to accomplish this objective is through a mathematical manipulation
of numerical data that can be computer processed. Therefore, we use the Digital
Elevation Models (DEM), as an approximation of reality.
In general, the DEM creates a representation of physical phenomena, in this
case the ground surface, from a spatial, geometric point of view. This digital
representation is, by definition, only an approximation to reality through a finite

sample of values and a set of functions.
In order to achieve the best adaptation possible to the phenomenon it
represents, the project used the geodesic reference system (ED50) and the UTM
projection (huso 30), the official system of Spanish cartography. The DEM has a
25 m grid spacing, and the data are referenced to the Alicante sea level datum.
Accordingly, the DEM fulfils two functions. On the one hand, it computes
relief and its derivatives (slopes, orientations, curvatures, etc.). These
characteristics constitute a Digital Terrain Models (DTM). On the other hand, the
DEM serves as a basis for geo-referencing other thematic maps.
The project will attempt, on the basis of this modelling of the landscape
features of the three natural parks of Guadalajara, to acquire a fuller knowledge
of the physical condition and construct a tool to improve management and
planning of these protected natural places. All this information is being
processed with the help of appropriate software, a Geographical Information
System, whose purpose is to optimise and facilitate analysis of the data.
3.1.3 Hydrogeology
Understanding groundwater is very important to the management of the Río
Dulce and Alto Tajo natural parks. Most of both parks are underlain by karstic
aquifers, highly susceptible to contamination. Hydrogeological research is
focused on the precise localisation of the groundwater recharge areas, as these
are the points where potentially contaminating agricultural or livestock activities
can be regulated.
In some cases recharge occurs outside the park boundaries, so that the
mapping of these areas is necessary to justify their protection. The precision of
this research includes carrying out classic tracer tests. The protection limits
should also take the discharge points into account, some of which appear to be
independent of the river basins.
Also important in their implications for the management of these two parks
are the hydro-geochemical processes in the Keuper facies that result in
mineralized springs. The deposited minerals are essential to plant growth and
attract certain animals in search of water and dietary salt.
The metamorphic bedrock and mountainous relief which predominate in the
Tejera Negra natural park prevent any significant infiltration of precipitation.
However, we intend to monitor the permeability of the numerous debris flows,
scree slopes and areas of surface alteration to determine their contribution to base
flow and response times.
3.1.4 Soils
Soils are one of the more dynamic components of any landscape. They are
constantly being formed by physical, chemical, and biologic processes. They are
essential to the maintenance of a healthy ecosystem, especially in these natural
parks.
Land management without adequate knowledge of soils and their correct
handling have in many cases destroyed what they intended to conserve. Such
actions have damaged the entire natural environment and increased erosive
activity on the remaining resources, with the consequent modification of the

landscape. This is especially true in these areas of the Iberian Peninsula where
the severity of current climatic favours intense soil destruction. Conserving the
soil resources depends on a complex combination of economic factors, on the
knowledge of conservation means and techniques, and on the possibilities within
the social environment which favour it.
For this reason, the soil surveys of the three natural parks in Guadalajara
consist of a soils mapping and the evaluation of how vulnerable they are to
physical and chemical degradation.
3.1.5. Vegetation
The plant cover often hinders the observation of underlying geological elements.
However, the specialised observation of the plant cover allows the geological
characteristics hidden from sight to be deduced.
The relationship and correspondences between geology and botany have long
been dealt with and written about from very different viewpoints: from those
who consider geology as a secondary factor [1] to those who put it first when
justifying the discontinuities and even the vegetation typology [2]. The most
obvious approximations, such as those in cases of azonal vegetation [3], or in the
ridge crest-slope-valley floor models, are relatively easy, but more complex
approaches can lead to difficulties. The discrepancies and difficulties are due
above all to the web of complex relationships within the natural environment and
also to the different rhythms of geological and vegetation processes.
When geology is cited as the vegetation controller there are many factors
involved, for example topography, lithology, chemical composition, pH, joints,
soil development... Further, each factor can vary in importance depending on the
case. Some medium-scale (1/50,000) studies carried out by drawing up
independent thematic mappings have produced only partly satisfactory results
[4]. The destruction of the plant cover and its spontaneous recovery tendency, are
the origin of the mosaics of different plant communities existing within
geologically uniform areas because of differing regeneration requirements. But
in most cases, the botanical component of the landscape mirrors the history of
land uses. To overcome these problems requires the use of mapping units which
combine arrangement and dynamism [5], grouping complexes of plant
communities linked to each other through dynamic relationships or
“eucomplexes”, and separating them from the mosaics of communities without
these relationships (the so-called “geo-complexes”).
Within this framework, one of the aims of the PANAGU project is to discover
the spatial and mapping coincidences between these units, eucomplexes, catenas
and mosaics, and the synthetic units defined by the geo-edaphic elements which
model the landscape, lithology, topography and processes. In this way the
analysis of the “visual” landscape, where the plant cover is an important element,
can be completed.
3.1.6 Visual landscape

The visual elements of a landscape are probably the first and most noticed by the
general public in the three natural parks of Guadalajara. Since it is the result of
all past geologic, climatic, and vegetative history, it is truly integrative. Although
the landscape constitutes the integrating element of the whole project, its
“specific” study is approached with four premises: a) to prioritise the correlation
between the compositional and visual properties of the territory; b) to build a
typological classification for the territory of the parks and their surroundings; c)
to carry out a specific mapping of landscape types; and d) to analyze the visual
scenarios and their relationship with geology (rocks, structure, morphogenetic
processes). Given that the three natural parks have different characteristics,
different aims have been set and different methodologies followed for each area.
For the Río Dulce natural park a detailed mapping of landscape scenarios has
already been carried out (Fig. 2), along with a catalogue and a genetic
classification. Three main scenario groupings have been differentiated: (a) the
more-or-less flat uplands; (b) the narrow valleys of the Río Dulce and some of its
tributaries; and (c) slopes of transition territory between groups A and B. The
geological substratum of the uplands conditions its partial erosion and the
agricultural use. Thirty upland scenarios have been distinguished, belonging to
ten landscape types. Nineteen valley scenarios have been distinguished, grouped
in six landscape types. Thirty five transition area scenarios have been
differentiated, belonging to six landscape types.
Figure 2: Example of landscape mapping in the Río Dulce natural park.
3.1.7 The artistic contribution

Science, at its core, really means the scientific method, a way of observing or
accessing nature from the senses and reasoning, from observation and the
making of hypotheses, from verification and raising them to the category of
scientific theories or relative interpretation models. It is never absolute, but
rather subject to changes in time, in line with advances in other fields such as
technology. Art, however, can be defined historically: we cannot confine it to a
single conception. We can only say that art is the way artists interpret nature, so
that there will always be as many “arts” as there are artists, although it is true
that there are certain categories which repeat themselves throughout the history
of art. They are driven by the search for expression, for harmony, for beauty, etc.
Every artist sees the landscape according to his or her own sensitivity, and for
this reason it is an open-ended way of looking, a poetic reading of nature,
receptive to suggestions and to the recognition of signs and symbols of every

sort. The scientific reading of the landscape, on the other hand, although not
without sensitivity, gives a reading in prose, demanding in its words, in the
concepts and in the rigorous and systematic description of the medium. In this
context, the artistic contribution to this project attempts to: (1) represent the
different landscape typologies obtained from their classification; and (2) make
abstractions from the different elements which make up the landscape (Fig. 3).
Figure 3: The artistic representation allows the abstraction of the landscape

components; left, geological skeleton; right, a delicate tapestry of
vegetation provides the visible scenario (M. Cebrián).
3.2 Synthetic or physiographic method
The integrated study of the landscape, represented in maps of homogeneous

territorial units, is being carried out on the basis of geologic and geomorphologic
information, which are given an integrated content through the progressive and
comprehensive inclusion of existing ecological relationships, and not simply by
addition by the use of overlays of other elements, such as soils or vegetation.
3.2.1 The mapping of landscape elements in the Tejera Negra natural park
The bedrock of this natural park is mostly a series of highly fractured Silurian
and Ordovician black slates. These slates are exposed in numerous surface
outcrops. In the higher areas there is discontinuous debris overburden while in
the lower areas colluvium overburden can be found. Finally, the main rivers, the
Lillas and the Sorbe, flow over an alluvial deposit of some depth.
A detailed geomorphological mapping (scale 1:25,000) of the whole natural
park is being carried out because of the importance it has for understanding soil
development, and subsequently for vegetation. This mapping is of the synthetic
or physiographic type, not the morphogenetic as it will look at and describe
bedrock plus landform. This mapping of the landforms will then be used to
analyse the relationship of the soils and finally the relationship of these soils with
the existing vegetation types.
The aim is to obtain a map of the homogeneous territorial units in detail, or
“landscape elements”. We intend that the characterisation and evaluation of these
units, both ecological and visual, will provide useful information for the
management of this natural area.
4 Active karstic processes in the Guadalajara natural parks

The methodology proposed in the previous paragraph attempts to approach the
study of the physical environment and the landscape of the natural parks in
Guadalajara from an integral and methodological point of view. Another
essential aim of the PANAGU project is the characterisation and quantification
of the geologic and geomorphologic processes active in these karstic parks, in so
far as these processes form part of the natural dynamic, with repercussions for
ecological and human activity.
4.1 Tufa deposits in the Alto Tajo Natural Park
In the Alto Tajo Natural Park there are numerous karstic features, which are
undoubtedly of landscape, environmental and scientific interest [6, 7, 8]).
Among these, there are numerous outstanding tufa deposits. They range in age
from Neogene to late Holocene times with some still currently active, although
their growth is increasingly slow. From the paleoclimatological point of view,
the tufa deposits are contemporary with the last karstification periods which
occurred in the Iberian Chain throughout the Quaternary Age. They developed
during the Marine Oxygen Isotopic Stages (OIS), especially the odd numbers 7,
5 and 1, in Mediterranean climates characterised by high humidity and by
development of the plant covers [9, 10]). Within the Alto Tajo natural park, the
following tufa typologies can be distinguished:
a) Perched springline tufas developed from important karstic springs. They
are the most numerous and spectacular tufa build-ups in the canyon of the Alto
Tajo [10] due to the sizes. They reach volumes larger than a million m3. Their
outcrop geometry is composed of tiers with subhorizontal tops leading to
waterfalls of considerable height containing fossil or functional moss curtains.
Their stratigraphy contains tufas with different petrological facies. There are also
included fluvial deposits, periglacial lithoclast microbreccias, and paleosols
which result from a varied paleoclimatic record [10].
b) Fluvial tufas. The narrow, deeply cut gorges carved out by the Tajo river
controlled the genesis of the numerous phytoherm barrages of different
palaeoclimatic stages. These barrages are frequently located close to knick-
points associated with differential erosion of the underlying Mesozoic strata.
The oldest generation of these barrages, lower Pleistocene, is placed some 200 m
above the river bed of the Alto Tajo, while the most recent is Holocene and can
be found at 5 to 10 m above the stream (Fig. 4) [10]. In the ponded areas
upstream of the dams there are accumulations of abundant fine materials of
tufaceous origin. In some non-contaminated stretches of the river in the Alto
Tajo there are certain stromatolite groups currently growing laterally in the
channel margins [11]. These are deposits of microbial origin and have not yet
been investigated in spite of their rarity in Mediterranean river beds.
The tufaceous deposits of the area key elements in establishing the most
recent evolution of karstic landscapes. In fact, during the Holocene, numerous
tufa barrages formed an authentic fluviolacustrine landscape (Fig. 4) in the Alto
Tajo and in some of its tributaries [10]. After the Bronze Age, these systems
became degraded by human impact. As a result, in the Iron Age certain valleys
record a change in their sedimentary dynamic caused by the removal of the plant
cover for agriculture, cattle rearing or metal working activities. The numerous
terrigenous deposits, which reached the river beds from the slopes, put an end to
the precipitation stage of the carbonates and to the construction of the tufa
barrages. The increasingly numerous detrital deposits gradually destroyed these
vulnerable natural structures. At the present time, only one grouping of
functional barrage survives in the Alto Tajo, which dams the Taravilla Lake.
Figure 4: Sketch showing one hypothetical Holocene barrage tufa in the Alto
Tajo Valley (Bronze Age, 3,500 B.P.) (in [12] p. 160.)
4.2 Karstic erosion of the Río Dulce natural park
One of the most active geomorphologic processes in the Río Dulce natural park
is karstic erosion. The Río Dulce runs along the floor of a narrow canyon cut in a
Jurassic and Cretaceous limestone and dolostone plateau, where it is fed by
various karstic springs. The recharge of these springs is from the plateau, which
hardly shows any drainage. The precipitation filters through the soil and into the
numerous fractures in the rock below. To determine the chemical dissolution
rates within the carbonates, a protocol has been set up to periodically sample the
waters of springs, for chemical analysis of dissolved ions (Figure 5, Table 1).
Given the river’s mainly subterranean contribution, these values may be
considered as representative of the Río Dulce. This means that if the quantities of
dissolved calcite and dolomite are multiplied by the annual mean flow of the Río
Dulce (27.5 hm3 at the stream gage at Aragosa, five km downstream from the La
Cabrera spring) the following dissolution volumes are obtained:
1) 225.25 mg/L of calcite x 27.5 x 109 litres / year= 6,194.375 x 109 mg of
calcite / 109 = 6,194,375 tons / 2.71 ton/m3 = 2285.74 m3 of CaCO3 dissolved per
year.
2) 70.07 mg/L of dolomite x 27.5 x 109 litres / year= 1,926.925 x 109 mg of
dolomite / 109 = 1,926.925 tons / 3 ton/m3 = 642.30 m3 of MgCO3 dissolved per
year.
To sum up, although these results are preliminary and perhaps only valid to
give a general idea of the magnitude of the phenomenon, it can be estimated that
a total of 2,928.04 m3 (c.3,000 m3) of rock is dissolved annually in the karstic
system drained by the Río Dulce within the natural park.
Figure 5: The Río Dulce natural park and its hydrological setting: Dulce river
watershed, springs, stream gage and bedrock lithology.
Table 1: The mean content of calcite and dolomite dissolved in the water from
the Río Dulce springs, for the analyses carried out to date (HCO3 total is
296 mg/l).
Ions Mg / L Ions mg / L
Ca 90.2 Mg 20.2
HCO3 dependent (Ca) 135.05 HCO3 dependent (Mg) 49.86
CaCO3 225.25 MgCO3 70.07
5 Discussion and conclusions

By carrying out territorial inventories of the natural parks in Guadalajara using
both analytic and synthetic procedures, an attempt is being made to compare the
advantages and disadvantages of each of these procedures, as determined by the
usefulness of the information obtained. Given that these inventories have not yet
been concluded, it is still too early to know which kind of inventory is better
adapted to the characteristics of each of the parks.
As far as the tufaceous geo-systems are concerned, it is important to highlight
the distinctly high value which their related landscapes offer in the Alto Tajo
natural park. This is due to the exceptional dimensions of specific tufaceous
deposits, to their “capricious shapes” and to their spectacular waterfalls and the
humidity which impregnates the atmosphere of the systems which are still
functional, almost all of them sited within the domain of the slopes where they
are fed by the karstic springs. These geo-systems, however, are in grave danger:
easily degraded and vulnerable to environmental changes. In addition their
current growth rates are much slower than several centuries ago. Are recent
climatic changes responsible for this? Is human activity the essential reason for
this loss of functionality? To delimit the natural and human causes responsible
for this deterioration is one of the aims of this project. Fortunately, this fragility
is well known and reflected in legislation, thanks to our earlier research.
Evidence of this is the protection and conservation framework which covers the
tufa outcrops in the recent Law on Nature Conservation and Environmental
Impact Assessment passed by the Castilla – La Mancha Government.
With respect to the volume of rock dissolved in the Río Dulce, 3,000 m3 may
seem at first a significant volume, but in proportion to the area and volume
occupied by the carbonate massif susceptible to karstification, this dissolution is
tiny. Comparison between the annual dissolved volume and the volume of the
carbonate rocks of the massif gives that the percentage of this total which is
dissolved annually (2 x 10-5 %) is not much. This, in turn, allows an approximate
calculation of how many years it will take to dissolve the total of the park down
to river level (!) or, on the other hand, to carry out evolutionary interpretations of
the formation of the canyon and of the park itself.
References
[1] Cain, S.A., Foundations of Plant Geography, Harper: New York, 1944.
[2] Kruckeberg, A.R. Geology and Plant Life, University of Washington
Press: Seattle, 2002.
[3] Walter, E. Vegetation of the Earth and Ecological Systems of the
Biosphere, Springer-Verlag: Berlin, 1979.
[4] Forteza, J., Cruz, R., Goy, J., Barrera, I., et al., Soil representation in
landscape and geomorphology for the regulation of the Candelario Natural
Reserve (Salamanca, Spain), GIS use and limitations. Proc. of the III Int.
Congress of European Society for Soil Conservation, Valencia, 2000.
[5] Vigo, J. Some reflections on geobotany and vegetation mapping. Acta Bot.
Barc, 45, pp. 535-556.
[6] González Amuchastegui, M.J., Parameras de Molina y el Cañón del Alto
Tajo. Guía de los Espacios Naturales de Castilla – La Mancha, pp. 201-202,
1991.
[7] García Quintana, A., García Hidalgo, J.F., Martín Duque, J.F. et al.,
Geological factors of the Guadalajara Landscapes (Central Spain) and
their relevance to landscape studies. Landscape & Urban Planning, 69,
417-435.
[8] García Quintana, A., Martín Duque, J.F., González Martín, J.F., et al.
Geology and rural landscapes in central Spain (Guadalajara, Castilla – La
Mancha). Environmental Geology, 47, pp. 782-794.
[9] Ordóñez, S. González, J.A., and García del Cura, M.A., Datación
radiogénica (U-234/U-238 y Th-230/U-234) de sistemas travertínicos del
Alto Tajo (Guadalajara), Geogaceta, 8, pp.53-56.
[10] González Amuchastegui, M.J. y González, J.A., Significado
geomorfológico de las acumulaciones tobáceas del alto valle del río Tajo
(sector Peñalen-Huertapelayo). El Cuaternario en España y Portugal,
Vol. 1, ITGE, Madrid, pp. 99-109, 1993.
[11] Guerrero, I. y González, J.A., Características geomorfológicos del modelo
de construcción tobáceo del Alto Tajo en su fondo de valle (Peralejos de
las Truchas, Guadalajara), Geotemas, 1(3), pp. 375-378.
[12] González, J.A. y Rubio, V., Las transformaciones antrópicas del paisaje
de los sistemas fluviales tobáceos del Centro de España. Bol. R. Soc. Esp.
Hist. Nat. (Geología), 96, pp. 155-186.
Exploring the effect of demographic elements

on the evaluation of the scenic beauty of
various landforms – preliminary results
A. Tsouchlaraki
Hellenic Open University, Greece
Abstract
This paper reports on a piece of research which attempts to take 32 different
images of landforms and to show how different classes of people perceive or
“value” the scenic beauty of these landforms. This is an interesting question and
one which bears research. In this case the term “scenic beauty” refers to the
public preference of various forms of the earth's relief. It is a figure which, even
though it depends on various subjective factors, aims to quantify the general
preference of the public for various landforms. A questionnaire survey takes
place in order to investigate the public preference, using a sample of 221 persons
in the area of Athens and Piraeus, and the city suburbs. This area concentrates a
very large proportion of Greece's population and can ensure variety in terms of
social and demographic status of the sampled population. The means used to
demonstrate the various forms of relief are 32 digital relief images, created with
the use of an algorithm developed for this purpose. The representative selection
of the sample of digital images took place after the classification of the forms of
relief in Lefka Ori mountain range in Crete. The questionnaire included 11
questions describing the person questioned in relation to the environment he/she
has experienced or knows and his/her contact with the countryside. Each
question corresponded to a factor (e.g. age, sex, education, income etc.) that had
been generalised in categories. Following this, a primary statistical analysis of
variance was carried out for each of the factors examined and some preliminary
results are reported. Little research in this specific area has been done and it is
interesting to explore further the way people with different social backgrounds
react or perceive the various landforms.
Keywords: scenic beauty, landscape aesthetic, perception, landforms,
demographic elements.
doi:10.2495/GEO060331
1 Introduction
Landscape is the combined result of physiography, geological formations,
vegetation, waters and the various cultural interventions that occur in a given
area. This combination attributes shape, line, colour and texture to a landscape,
while the aesthetic result is considered on the basis of the variety or the
uniqueness offered and is usually classified into three main classes: 1)
indistinctive, 2) common, 3) distinctive (USDA, [1]; USDA, [2]). This
classification determines those landscapes which are most important and those
which are of lesser value from the standpoint of scenic beauty. The classification
is based on the premise that all landscapes have some scenic value, but those
with the most variety or diversity have the greatest potential for high scenic
beauty. The various approaches developed for determining landscape visual
quality refer to the general output of the synthesis of all physical variables of the
environment, physiography, soil, vegetation, hydrological elements, and not
individually to each element. In these approaches, the visual quality of a
landscape is evaluated either indirectly by thematic maps according to
standardised criteria based on the experience of the scientists, or directly on the
basis of psychometrical methods that quantify directly the public’s preference by
demonstrating through a certain means the landscape under evaluation (Kaplan et
al [3]; USDA [1]; Kaplan [4]; Daniel and Boster [5]; U.S. BLM [6]; Palmer [7];
Smardon et al [8], Hunziker and Kienast [9]). Relief constitutes, however, a part
of these approaches and not the main objective.
There are cases where it would be desirable to isolate the relief from the other
elements which are making up the visual environment and to investigate its
scenic beauty separately. Such cases include the technical works that cause major
and permanent alterations to the relief, as well as the works, of which the spatial
arrangement depends on the morphology of the ground. Slopes, distances,
hypsometric difference values, viewsheds are elements of the relief that affect
the visual quality of a landscape, but also the ability of a particular landscape to
accept and absorb new activities.
The experience of the scientists show that the more mountainous the form of
the relief, the more distinct the landscape category offered (USDA [1]; U.S.
BLM [6]). It is interesting in this case to explore the way various persons react,
based on different demographic elements, to the different forms of the relief. The
way each observer evaluates any given landscape is a very complicated issue, a
matter difficult to predict. Many factors can possibly influence this evaluation,
ranging from factors that can be registered, such as the usual demographic
elements, to imponderable factors, such as the mood of the observers at the time
of the evaluation. Besides, what “one likes or dislikes” does not remain constant
with time. As a person matures, his/her attitude towards many things in life
changes. Therefore it would be utopia to try to predict with precision the
preference of an observer for a given landscape, but this is not the aim of this
research. This paper presents some preliminary results of a research work that
aims to investigate the way in which the various demographic elements of the
persons questioned influence the preference in the case of the evaluation of the
scenic beauty of a landform. Whatever conclusion can be drawn, even based on

the few factors that can be registered, would be useful in the interpretation of the
phenomenon.
2 Investigation methodology
2.1 Digital relief visualisation
The investigation of the public preference to the various landforms can be

achieved with the use of questionnaires, by utilising some means for presenting
the different forms of relief. In this case, it was not possible to use actual
photographs as a means to demonstrate these different forms, because they
would provide simultaneously information on the vegetation, the soil and the
land uses of a given area, something that could influence the preference of the
persons questioned. Therefore, the survey was effected using digital visual
representations as a means to demonstrate the various forms of relief. These are
images created with the use of a Digital Terrain Model and the respective
shading image. The creation of these images relied on an algorithm that had been
developed for this purpose, producing the perspective image of the relief, as this
would look like if it had been photographed from a known shooting point in
relation to a given target, with the use of a photo camera of known geometry
(Tsouchlaraki [10]). All shootings are strictly horizontal, considering the
geometry of a 35 mm camera with a normal lens (f=50 mm) and a predetermined
data analysis scheme.
The study area was that of the Lefka Ori mountain range in Crete, owning to
the variety of the forms of relief it includes. The IDRISI GIS package was used
for the processing of information, as well as the creation of a DTM and of other
derivative elements.
The representative selection of the sample of the digital representations that
would be included in the questionnaire took place following a classification of
the relief forms, based on a method developed for this purpose, which is a
modification of Hammond’s classification method (Tsouchlaraki [11,12]).
The classification resulted in 32 relief categories. A horizontal position was
selected for each category, for creating a digital representation. The 32 images
created are representative of the different forms of relief in the study area and are
depicted in Figure 1. As far as the scenic beauty is concerned, the persons
questioned were asked to rate it on a 1-10 scale (1 representing a very small
preference concerning the scenic beauty of the landform and 10 representing a
high preference). This scale was chosen because the studies of landscape
aesthetics which use photographic imagery (both actual and simulated) and 1-10
scales for response are common in use.
2.2 Questionnaire design and execution
There are many rules for the design and execution of a questionnaire and also
many decisions that have to be taken (Damianou [13], Koutsopoulos [14]). The
1 2 3 4
5 6 7 8
9 10 11 12
13 14 15 16
17 18 19 20
21 22 23 24
25 26 27 28
29 30 31 32
Figure 1: The 32 images of landforms included in the questionnaire.
size of the sample, the sampling methodology, the form and type of
questionnaire, the duration of a questionnaire are some of the matters that have
to be examined. This subsection is devoted to these matters.
The questionnaire was used in the area of Athens, Piraeus and the city
suburbs, using a sample of 221 persons. This area, comprised of 52
municipalities and 5 communities, offered the following advantages:
A) It is an area of Greece’s capital that concentrates a very large proportion of
the population and is suitable for ensuring variety in terms of the different
social and demographic characteristics of the persons questioned.
B) The population of Greece’s capital comes mostly from different areas of
Greece, and allows the inclusion of people from the provinces, with different
experiences and mental representations of the various forms of the relief.
The total population of the study area is 3,020,562. The area includes both
rich and poor quarters, both densely and scarcely populated, all of them being
within the city plan limits so that they are mapped and facilitate the organisation
and implementation of the questionnaire.
A stratification of the municipalities of Athens was effected before the
selection of the persons to be questioned. In order to ensure the representation of
the entire population, two individual layers were used prior to adopting the final
sampling. Upon the first layering, the municipalities were classified in three
categories, according to their population: small (X<30,000 inhabitants), average
(30,000<X<70,000) and big (X>70,000 inhabitants). Upon the second layering,
we used the existing classification of the prefectures: Athens, Eastern Attica,
Western Attica and Piraeus. We thus created 3x4=12 layers.
In a simple random sampling, the sample size is calculated as follows
(Damianou [13]):
P [ | y - Y | < d ] > 1-a (1)

where:
Y : the population average of the requisite characteristic, which in this
case is the visual value;
y : the population average estimate derived from the sample;
d: the error margin or the desired measurement accuracy of the average;
1-a: the confidence coefficient.
It is clear from equation (1) that the size of the “n” sample is determined by
the following formula (Damianou [13]):
n0 (2)
n =
n
1+ 0
N
in which:
sz a/ 2 2
n0 = ( ) (3)
d
where:
Ν: the total population;
s2: the dispersion of the characteristic;
za/2: the upper a/2-point of the normal distribution Ν (0,1).
The term n0/N, when the value of N is very big, tends towards zero and is
ignored. In the layered sampling analysis, the same reasoning is followed, with
the exception being that the above formulae (2,3) are applied in each layer, using
the respective dispersion of the characteristic which each layer exhibits. In case
the dispersion is unknown, we can use its value from previous studies related to
the subject matter (Damianou [13]).
Because of the fact that there are no previous studies related to the certain
subject matter, we used the results of an experimental implementation in 55
students of the 2nd and 8th semesters of the Department of Rural and Surveying
Engineering of the National Technical University of Athens. This
implementation was a simple questionnaire including the 32 images and asking
students to evaluate them on a 1-10 scale, according to their preference. In this
pilot study the previous experience of the scientists came true as the more
mountainous the form of the relief, the higher the values that resulted from the
students’ preference. The results of this implementation showed that the standard
deviation differs among images and ranges from 1.64 to 2.75 units. In order to
determine the final sample, we assumed that all layers present the same
dispersion and took the worst case to be the standard deviation: s=2.75 units. The
desired measurement accuracy of the mean was determined to be d=0.40, a value
lower than the unit half. Therefore, for a 95% probability, the application of the
above formulae shows that the size of the sample is:
2.75 x1.96 2
n =( ) =182 (4)
0.40
This figure was increased, for the sake of safety, by about 20 percent and the
final sample was thus determined to be 221 persons. This sample was selected
for two main advantages: 1) it was small in relation to the size and therefore
useful for carrying out a further research; 2) it could assure the required
reliability in order to draw some initial conclusions. The students’ responses
were not used in this final sample or in the further analysis. The sample was
distributed among the individual layers, based on the percentage of the
population in each of the 12 layers in relation to the total population of the city.
Following this, we proceeded to random sampling for the selection of the sample
decided for each layer, using the files of the National Statistical Service of
Greece.
The questionnaire included 11 questions and was accompanied with the 32
coloured images. The questions had a multiple choice form and were selected in
such a way as to describe the person questioned in relation to the environment
he/she has lived in or is familiar with, during his/her contact with the
countryside. Also basic demographic elements, such as sex, age, education,
profession, income are included in the questionnaire. Each question corresponds

to a single factor, generalised in a maximum of three categories. Besides, the
relatively small size of the sample would not have benefited the larger in number
categories. The questionnaire is presented in the following Figure 2. The eighth
question was devoted to the scenic beauty of the various landforms presented in
the coloured images. The responder was asked to rate the landforms on a 1-10
scale according to his/her preference. With regard to the ninth question about the
profession, the third category entitled “other” refers to unemployed, students,
housewives and in general people who do not work.
The technique used for the questionnaire was that of the interviews. The main
advantage of this technique is the direct contact with the person questioned and
therefore presents the highest participation rates. The time available for the
interview depends mainly on the way it is conducted. In this case, we decided to
visit the persons questioned at home, in order to allow for a 10-15 minutes
interview.
2.3 Questionnaire results
Table 1 shows the descriptive elements of the results’ distribution, as derived for
the scenic beauty of the various landforms. For the other questions the results are
shown in Table 2. For the sake of brevity, all references hereinbelow will use the
symbolism of factors shown in Table 2.
It is clear from the frequency values that there is sufficient number of
observations in each category. The sample includes individuals for all the
categories of age, education, income, profession and also individuals that come
from different places of Greece or abroad.
As pertains to sex, women are the majority, however, the percentages are
close to the respective percentages of the official census, in which men
correspond to about 49% of the population and women to 51%; therefore these
percentages are considered to be satisfactory for the balance between the two
sexes.
As pertains to the scenic beauty, all images were rated with values from 1-10.
The average responses for each image range within 5.02-7.72, with standard
deviations from 1.91 to 2.62. Therefore, the standard deviations are within the
range of the values that had been observed from the experimental questionnaire
to the students.
When observing the images in a descending order as to the mean scenic
beauty, we can easily conclude that low values correspond to more plane forms,
while high values to mountainous forms. As it has been already mentioned, this
is one of the criteria used in landscape analysis and evaluation, because
mountainous forms present greater variety in the relief elements (slopes,
curvatures, crest etc.) in relation to plane forms.
An interesting observation may also derive, when we also look at the mean
values and the respective deviations of the answers. Let’s consider that the
standard deviation is the measurement of disagreement among the respondents,
then we observe a greater disagreement of opinions in image 1, and a greater
agreement in image 21. By observing the remaining images, it seems that
disagreement tends to increase as the relief’s visual value decreases. The

respondents therefore seem to agree more on what is good rather than bad. This
observation is not an object of this study, but we present it as an interesting issue
for further investigation.
Scenic Beauty of Various Landforms
Please give a single answer to the following questions.
1. Sex:
1. [ ] Male 2. [ ] Female
2. Age:
1. [ ] 18-35 years 2. [ ] 35-50 years 3. [ ] more than 50 years
3. You come from:

1. [ ] Athens 2. [ ] Outside Athens 3. [ ] Outside Greece
4. Where have you spent most part of your life?

1. [ ] Big city ( ………name of the city) 2. [ ] Small town (……… name of the town)
5. How often do you visit the countryside?

1. [ ] Very often, almost every month 2. [ ] Not very often, two or three times every
year
6. Usually for what purposes do you visit the countryside?

1. [ ] Leisure 2. [ ] Other
7. Using a scale of 1 (low preference) to 10 (high preference) how much would you evaluate
the scenic beauty of the landform that is represented in each image?
Image 1: ___ Image 2:___ Image 3:___ Image 4:___ Image 5:___
Image 31: ___ Image 32:___
8. Education:
1. [ ] Primary 2. [ ] Secondary 3. [ ] Higher
9. Profession
1. [ ] Employees 2. [ ] Self-employed 3. [ ] Other
10. Income yearly

1. [ ] Less than 5 millions GRD 2. [ ] More than 5 millions GRD
11. Usually how long do your vacation last?

1. [ ] Less than 15 days 2. [ ] 15-30 days 3. [ ] More than 30 days
Figure 2: The questionnaire.
Table 1: Descriptive elements of the distribution of results for the visual

value of the landform of each image.
Image Mean Standard error Standard

deviation s
x s( x )
1 5.02 0.18 2.62
2 6.81 0.16 2.42
3 5.71 0.13 1.97
4 5.49 0.14 2.06
5 5.98 0.15 2.19
6 6.28 0.13 1.95
7 6.43 0.15 2.30
8 5.67 0.17 2.48
9 6.34 0.13 1.96
10 7.05 0.15 2.28
11 5.70 0.13 1.95
12 6.02 0.15 2.21
13 5.76 0.14 2.13
14 5.56 0.16 2.38
15 7.38 0.14 2.01
16 6.51 0.14 2.09
17 5.44 0.17 2.49
18 7.41 0.14 2.14
19 5.61 0.16 2.36
20 7.49 0.14 2.06
21 6.04 0.13 1.91
22 6.59 0.13 2.00
23 5.86 0.14 2.01
24 6.10 0.14 2.14
25 6.76 0.13 1.98
26 5.86 0.17 2.50
27 6.10 0.13 2.00
28 5.86 0.14 2.06
29 5.89 0.13 1.95
30 5.68 0.15 2.28
31 7.14 0.14 2.05
32 7.72 0.14 2.01
3 Analysis of variance – preliminary results

The results were proceeded with the analysis of variance for each individual
factor (Bora-Senta and Moysiadis [15]). The dependent variable was the scenic
beauty of landforms. For any one interviewee there is only one set of factors
which are repeated against all 32 images. Repetition of characteristics in
multivariate analysis has unpredictable results. In order to avoid this problem
and get some preliminary results, instead of using all the ratings given by each
interviewee we used the sum of ratings of all the images for each interviewee. In
this manner the data matrix included 221 observations.
Table 2: Categories of factors.
Question/ Factor General Categories Occurrence Frequencies

- male 93
Sex (SEX)
- female 128
- 18 - 35 94
Age (AGE) - 35 - 50 76
- > 50 51
- Athens 90
Origin (FROM) - outside Athens 115
- outside Greece 16
Where have you spent most part - big city 177
of your life? (LP) - small town 44
How often do you visit the - > 1 / month 65
countryside? (CVI) - < 2-3 / year 156
- leisure 161
For what purposes; (RCVI)
- other 44
- primary 61
Education (EDU) - secondary 94
- higher 66
- Employees - Self- 85
Profession (PROF) employed - Other 61
75
- < 15,000 EURO 151
Income (FIN)
- > 15,000 EURO 70
- < 15 days 59
Time of vacation (HOL) - 15 - 30 days 122
- >30 days 40
From the results gathered from this preliminary factor analysis of variance,
we can draw some conclusions. In general, and in almost all of the factors, the
mean values of the individual categories differ from one another with a high
degree of reliability. The F-ratio level of significance ranges from 0.00 to
0.0022. With a probability of almost 100%, this means that the average values of
the individual categories are not equal. The sum of the squares of errors between
the groups, which the higher it is, the better the factor classifies the dependent
variable, can become also a benchmark for the factors. Factor analysis extracts
factors which maximise the variance explained in order of the most important
first and so on.
Considering the results of the analysis, the variation of the mean values in the
categories of each factor leads to the following conclusions:
1. The persons who have spent most part of their lives in small towns or villages
give higher ratings than those who have lived in cities or in city suburbs.
2. The persons who originate from areas outside Greece give ratings higher than
those who come from areas outside Athens, and the latter give ratings higher
than those who come from Athens.
3. Women give higher ratings than men.
4. The persons who spend more than 30 days annually for holidays give higher
ratings compared to persons who spend 15-30 days, and the latter give higher
ratings than persons who spend less than 15 days.
5. The persons with a family income below EURO 15,000 give higher ratings
compared to those with a family income of more than EURO 15,000.
6. The persons aged 50 years or more give higher ratings compared to those in
the 35-50 age group and the latter give higher ratings compared to persons in
the 18-35 age group.
7. The primary education graduates give higher ratings compared to the
secondary education graduates, and the latter give higher ratings compared to
university graduates.
8. Persons visiting the countryside for leisure purposes only give higher ratings
than persons visiting the countryside for other reasons.
9. Unemployed persons give higher ratings than self-employed and self-
employed give higher ratings than employees.
10. The persons who visit the countryside more often than once a month give
higher ratings compared to the persons who visit the countryside from time to
time, 2-3 times a year or do not visit it at all.
The way the respondents used the same scale of values is different from
person to person, however it seems feasible to group and generalize their
behaviour. This is a first conclusion and perhaps one we would expect. What is
that makes certain groups of people use higher values in relation to other groups?
An important generalisation drawn from the observation is the following: the
persons who have or had in the past more chances to come to contact with a
physical environment use higher ratings. This probably explains why the persons
who have lived in or come from the countryside, together with the persons who
visit the countryside more often, the persons with more freedom in their work
time, or the more aged persons who had more chances to visit the countryside,
are the ones who know the physical environment better and thus give higher
ratings. It is through the high degrees of freedom that they express their
preference for nature.
Up to here, the previous observations derive from the whole sample of the
respondents. However, the fact that women gave higher ratings than men is a
matter of concern and shows that we should perform a further investigation for
both sexes separately. This will be the main concern of the following phase of
this research, in order to find out whether the factors that affect each sex remain
the same or not.
4 Discussion
This study addresses the issue of the influence of the demographic data on the
evaluation of the landforms scenic beauty. It is clear from the results that the
factors examined might influence to a certain extent the scenic beauty of the
landforms. Maybe there are many other factors, perhaps even more important
than the ones examined. No relative research had been conducted in the past, so
as to allow for a comparison. For example, perhaps the morphology of the place
of origin of the respondent or the place where he/she has lived most of his/her
life plays an important role in his/hers preferences and his evaluation, since each
respondent is familiar, due to his experiences, to certain relief forms. The
investigation of such factors is not part of this study, however it is a very

interesting issue for further research, in order to better understand and interpret
the public’s preferences.
References
[1] USDA Forest Service, National Forest Landscape Management,
Government Printing Office, Ag. Handbook 434, Washington, 1974.
[2] USDA Forest Service, Landscape Aesthetics, Government Printing Office,
Ag. Handbook 701, Washington, 1995.
[3] Kaplan R., Kaplan S., Wendt J.S. Rated preference and complexity for
natural and urban visual material, Perception and Psychophysics,
12(4):354-356, 1972.
[4] Kaplan S., Some methods and strategies in the prediction of preference, In
Landscape Assessment – Values, Perceptions and Resources, edited by
E.H. Zube, R.O. Brush, and J.A. Fabos, Stroudsburg, PA.:Dowden,
Hutchinson and Ross, pp.118-119, 1975.
[5] Daniel C.T., Boster S.R., Measuring Landscape Aesthetics: The Scenic
Beauty Estimation Method, USDA Forest Service, Research Paper RM-
167, 1976.
[6] U.S. BLM, Visual resource management: Visual resource management
program, U.S. Government Printing Office, USA, 1980.
[7] Palmer J., A visual character approach to the classification of backcountry
trail environments, Landscape Journal, 2(1), USA, 1983.
[8] Smardon R., Palmer J., Felleman J., Foundations for Visual Project
Analysis, John Willey & Sons, New York, 1986.
[9] Hunziker M., Kienast F., Potential impacts of changing agricultural
activities on scenic beauty – A prototypical technique for automated rapid
assessment, Landscape Ecology, 14(2), pp.161-176, 1999.
[10] Tsouchlaraki A., 1996. Digital Relief Visualisation in Landscape
Analysis, Technika Chronika, Scientific edition of the Technical Chamber
of Greece: I, 33, Athens, pp. 27-37, 1996 [in Greek with English extended
summary].
[11] Tsouchlaraki A., A landform classification method with GIS for landscape
visual analysis purposes, Proceedings of the Second International
Conference on Sustainable Planning, 12-14 September 2005, Bologna,
Italy.
[12] Tsouchlaraki A., A Methodology for the evaluation of the visual value of
natural relief. Ph.D., Department of Rural and Surveying Engineering,
NTUA, Athens, 1997.
[13] Damianou, Ch., Sampling methodology – Techniques and applications,
Aithra, Athens 1992 [in Greek]
[14] Koutsopoulos K., Geography: Methodology and Spatial Analysis
Methods, Symmetria editions, Athens, pp. 367-422, 1990 [in Greek].
[15] Bora-Senta E., Moysiadis X., Applied Statistics, ZITI editions,
Thessaloniki, 1992 [in Greek].
An approach to the landscape analysis

B. Badiani
Researcher of Urban and Regional Planning,
University of Brescia, Italy
Abstract
The starting-point is the new conception of landscape proposed by the European
Convention on Landscape (Florence, 2000): the landscape is a combination of
elements, natural and not, that encircle us.
In light of this new concept, the work proposes a reflection on the actual
modalities of the description and interpretation of the landscape used as the first
step in environmental planning.
In 2002, the Lombardia region approved a law concerning the guide lines for
the examination of the projects in relation with the landscape value. So it is very
important to make clear what a landscape means. But the elements suggested in
this law to describe the landscape do not seem to be useful in estimating its
value.
Taking into account the proposed methodologies of this regional law, an
attempt has been made to devise one practical and flexible procedure for the
evaluation of the characters of the landscape.
The aim is to reach a fundamental knowledge of what a landscape is, taking
into account its concrete and sensitive aspects and their possible descriptions.
The methodology is based on a check-list of natural and human elements,
implemented on a GIS. A database has been associated with each element. The
information put into the database is organized in a way to facilitate the
evaluation of the landscape, the communication of the result of the analysis and
the participation of the citizens in the decisional phase.
The methodology has been applied to a territory of a municipality in Northern
Italy to test the good flexibility, the simplicity and the clarity of the description
of the landscape image and the improvement of the quality and completeness of
the check-lists.
Keywords: landscape analysis, environmental planning, landscape planning,
GIS.
doi:10.2495/GEO060341
1 Introduction
Safeguarding the landscape mainly entails governing its transformations due to
the action of man or natural events. This translates into the fact that each
landscape and territorial policy initiative must therefore originate from tools
useful to govern the transformations.
According to the National legislation referring to the Natural Heritage and
Landscape Code, approved as Legislative Decree No. 42 dated 22nd January
2004 and that came into force on 1st May 2004, the urban municipal plan is
assigned a special conclusive value in the process to construct the overall
safeguarding system, as also expressly stated in the regional law of Lombardy
No. 12 dated 15th March 2005 “Law to govern the territory”, requiring the
foregoing planning tool to also include landscape-related aspects.
The following aspects emerged from an analysis of the initial experiences of
local landscape planning:
o the lack of a simple, effective, easily applied and above all shared
operating methodology, with the consequent limited tangible use of the
tool;
o a lack of clarity and precision in describing the criteria used to
determine the classes of sensitivity of the local landscape;
o a limited or no participation and involvement of the local community in
the preparatory phases of the Plans and when determining the classes of
sensitivity of the landscape;
o no consideration has been given to shared tangible and schematic
parameters or indicators, as a tool to analyse and evaluate the landscape.
The fundamental methodological aspects of the proposed method have been
defined starting from this critical verification. The method must be:
- concise → a limited number of parameters to be assessed;
- readily understandable and applicable → clear, well-defined and
practical;
- general and flexible → applicable to different territorial environments
without too many difficulties;
- easy to manage → based on readily obtainable data;
- effective and useful → which enables tangible results to be attained.
However, the starting point remains the clarity of the concepts and the
objectives which are to be achieved. Therefore, it is essential to specify what is
meant by landscape, which of the infinite definitions and interpretations of the
term is taken into consideration in relation with the scope of the methodology
proposed.
It was deemed appropriate to consider the concept of landscape proposed by
the European Convention (Florence, 2000), since it has been formulated recently,
incorporates various aspects addressed in previous years and expresses the
complex nature of the concept in the best possible way.
The landscape is everything that surrounds us and that we perceive with our
eyes and with our emotions, and more in detail it is “a determined portion of
territory which is perceived by man, the appearance of which is the result of the
action of human and natural factors and of their interrelations […] a shared
asset, the basis of the cultural and local identity of populations, an essential
component of the quality of life and expression of the richness and diversity of
the cultural, ecological, social and economic heritage”. In this sense we can
interpret the landscape as the series of signs left on the territory by natural and
human history, read and interpreted by an observer.
And it is at this point that the main problem to be addressed is encountered: if
the landscape is “everything”, how can a practical procedure be found that takes
into consideration each constituent feature? How can such an ambiguous
concept be rendered concise and schematic?
There is a need to deliberately limit the analysis of the landscape by
considering the most significant and important aspects for its planning and
management. But what is significant in a landscape? To make a selection
among the features which surround us entails a choice that is undoubtedly
arbitrary and subjective. Furthermore, in this way the understanding of the
landscape will never be complete, because some factors or aspects that could be
significant will always be excluded.
A possible solution exists if one observes that complexity is a factor which
cannot be eliminated in order to know the landscape, and it is pointless to try to
understand everything. It is necessary to be aware of the fact that an
“approximate knowledge” exists.
2 Presentation of the proposed procedure

The method that is illustrated below intends to foster the awareness of both the
designer and the local populations as regards safeguarding the landscape and can
be a useful help in landscape planning at the municipal level.
The methodological path is characterised in stages:
1. a first stage to identity the objectives associated with the;
2. an analysis stage, based on collecting data and processing the data in a
reference list;
3. an assessment stage, which entails describing the aspects which
combine to attributing value to the features identified as being
characterising features;
4. drawing up a landscape map, which is created by using GIS
(Geographical Information System) software, which enables some
alphanumeric information to be associated with the graphic elements
represented, thanks to the use of a descriptive database that contains
information regarding the features necessary to appreciate the
landscape’s overall value.
3 Landscape analysis stage

The scale of the intervention is selected, which in this case is on a municipal
scale, after having defined the objective of the landscape analysis, which could
be to identify the special local features in order to verify the sensitivity to

changes associated with planning choices.
The processing activity consists in selecting the data collected. The most
important factors to be considered in the analysis stage are outlined in the
following list, which can be enriched and extended:
1. physical and natural features;
2. anthropic features: these represent the landscape components that
derive from the action of man;
3. perception features: these represent the landscape components subject
to perception;
4. critical areas: these represent factors which identify particular
situations.
The check-list structured in this way includes all the landscape features
considered to be most significant and above all shared. This has been possible
by involving the local community, also in the stage of preparing the checklist.
Table 1 shows a number of the items adopted, for purposes of illustration.
A value must be assigned after the landscape has been analysed and after all
its constituent factors have been identified. This entails defining criteria to
establish an opinion about the single feature. This represents an extremely
subjective and delicate operation.
In this session it was deemed more important to define the aspects of the
feature that can contribute to establishing the value.
An attempt was made to break down the feature into its fundamental aspects
to appreciate fully all the aspects that can influence the value of a given feature.
The concepts of shape, meaning and use were selected.
First of all, a landscape feature is characterised by a shape, which represents
its physical constitution, its aesthetic features such as size, colours, contours, etc.
The value opinion in relation to the shape can depend on the presence of rare or
unusual features, the diversity and variety of the features, their integration, etc.
The meaning is the result of the individual's perception and is influenced by a
series of interrelationships which are established among the object, the
surrounding context and the person making the assessment. The relationship with
the objective of the analyses must also be established to assess the meaning. The
following criteria have been identified:
• Readability → meaning the quality that confers on a feature of the landscape
a high probability of evoking a well identified, structured and functional
environmental image in an observer;
• Use → has a dual meaning. To be understood as the possibility of using the
feature in question or the possibility of using other features via this feature;
• Context → represents the set of features that surround the feature under
examination and can be “natural”, or “man-made”;
• Visual → is understood as the type of view presented by observing the
landscape feature in its context;
• Value → this term is associated with the social and cultural aspects and with
the local history, as well as the importance of the landscape heritage;
• State → this term refers to the degree of naturalness and integrity of a given
feature.
Table 1: Brief description of the landscape features and the data sources.
Feature Brief description Sources

NATURAL FEATURES
Mountain, ridge Represent the watershed lines
Technical regional
surveys of the principal catchment
map
basin.
Morphological factors Characteristic forms of the
land due to past or current Technical regional
morphological map
transformations.
ANTHROPIC FEATURES
Historical centres and Built-up areas which represent Municipal
settlements the oldest district of the urban cartography
system. Historical archives
PERCEPTION FEATURES
Panoramic views Visual overviews of shared
importance
Panoramic points Place from which it is
Municipal and
possible to enjoy a panoramic
provincial
view
cartography
Routes to enjoy the Road or footpaths, pedestrian- Site visits
landscape bicycle lanes and routes from
which the landscape can be
enjoyed
CRITICAL AREAS
Degraded areas Areas subject to change Municipal
documents and
cartography
Site visits
The function is associated with the how the feature is used. The function
aspect is associated with a value related to the following criteria, as in the case of
the meaning assessment:
• Accessibility → this represents the possibility of reaching the landscape
feature, therefore of enjoying it and exploiting it from the recreational
or useful point of view;
• Use → this is understood as the use of the land referred to the area in
which landscape feature under examination is located;
• Profitability → this refers to the possibility of generating income
associated with the landscape feature as a tourist venue or as a source of
commercial resources or again as a place for recreation.
The most obvious and important novelty introduced by the methodology

proposed, consists in characterising the landscape features directly at the
cartographic level: the features are not only represented by locating them on the
map, but an assessment is also associated with them through a combination of
symbols associated with the aspects which define their value.
Some examples of the map detailing the shape, meaning and function aspects
of a number of features that constitute the landscape are shown in the
illustrations below.
LEGEND
• source
• eternal
• precise feature
• clear water
Figure 1: Shape table.
LEGEND
• panoramic point
• significant view
• strong readability
• limited use
• natural context
• panoramic view
• perceptive value
• in good conditions
Figure 2: Meaning table.
LEGEND
• hollow
• wood
• difficult access
• object of
speleological studies
Figure 3: Function table.
4 Conclusions
An attempt has been made to create a practical and flexible procedure to assess
the quality of the landscape, using on a knowledge database to manage the
territorial transformations by referring to the methodologies and the operating
methods proposed by the regional regulations which relate to landscape planning
and managing landscape transformations, and their application to a number of
municipal territorial contexts, (in particular the guidelines for landscape-related
examination of projects, approved with Regional Council Resolution (D.G.R.)
No. 7/11045 dated 8th November 2002).
The starting point of this study is represented by the endeavour to establish an
in-depth knowledge of the landscape features, to focus attention on the most
tangible and sensitive aspects and on their possible representation.
The greatest difficulty found was associated with the need to represent the
landscape concisely, without reducing the significance and complexity.
The landscape is analysed by specifying its shape, its meaning and its
function and the methodology proposed enables its potential, values and
criticalities to be interpreted.
The procedure proposed has been applied to a case study: the Municipality of
Polaveno, located in North Italy (Province of Brescia) and a number of aspects
worthy of note have been highlighted. On the one hand, problems and
difficulties emerged due to the limits of the form to represent the landscape that
is necessarily partial. On the other hand new potential for new fields of
application have emerged.
The need to make significant choices regarding the features to be taken into
consideration stems from the impossibility of representing “everything”. In
relation to the objective that is to be achieved, the best and most effective criteria
to be adopted in making this selection, is undoubtedly the criteria of interpreting
the landscape based on the social perception of the persons that benefit directly
from the landscape.
Difficulties were experienced in retrieving all the useful information, both
with reference to the location of the landscape features, as well as to determine
their meaning. The available and up-to-date cartographic sources are very
limited. The best solution is to verify the landscape features by site visits, and
above all to become acquainted through the eyes of those that live there and
benefit from the landscape directly every day.
The methodology proposed does not impose any rigid schemes. Using
flexible check-lists becomes a simple reference and a guide in the selection of
the features which describe the landscape and for their assessment.
It was found useful to exploit the advantages and the potential offered by the
GIS (Geographical Information System) digital tool to be able to consider the
continuous transformations of the built up landscape framework, in the best
possible way.
The data collected and summarised can be changed or consulted at any time,
also solving the problem of updating the maps.
References
[1] R.Arnheim, La dinamica della forma architettonica, Feltrinelli, Milano, 1981
[2] B.Badiani, Movimento e qualità dello spazio urbano, Ed. Bios, Cosenza,
2004
[3] A.Clementi, Revisioni di paesaggio, Meltemi ed., 2002
[4] J. Gehl, Vita in Città: spazio urbano e relazioni sociali, Maggioli Editore,
Rimini, 1999
[5] C.Lévy-Leboyer, Psicologia dell’ambiente, Laterza, Bari, 1982
[6] A.Sestini, Il paesaggio, TCI, Milano, 1963
[7] F.Steiner, Costruire il paesaggio. Un approccio ecologico alla
pianificazione, McGraw-Hill, 2004
[8] L. Quaroni, Il volto della città, in Comunità n.25, 1954
[9] M.Tira (a cura di), Metropoli e mobilità: il caso di Brescia, vol.4, Sintesi
Ed., Brescia 1996
Evaluation and analysis criteria of the

environmental risk factor of the anthropic
perturbation in the infrastructure works
G. Gecchele & G. Pizzo
Land, Environment and Geo-Engineering Department,
Turin Polytechnic, Italy
Abstract
The present study represents a methodological example of the knowledge of the
territory through the identification of the qualitative parameters that give
technical information of the territory characteristics, of the environmental effects
and changes due to anthropic perturbation.
In this work some analysis and evaluation criteria of the environmental risk
factors are presented, due to the infrastructure works, in order to supply the
planning choices and to manage the execution phase, referring to great
infrastructure works.
In particular, this work, relating to the meaning of impact and environmental
quality established by the literature, wants to present a check list for the
environmental analysis, which is applied as an example to a phase of an
infrastructure work and which represents an important scheme to identify, to
analyse and finally to evaluate the necessary information to reach the
environmental safety.
Keywords: environmental analysis and evaluation, infrastructure works,
1 Introduction
The great civil infrastructure works need specific analyses aimed to identify,
evaluate and manage the environmental impacts.
It is then necessary to establish which is the environmental state before the
beginning of the works, in order to understand its possible evolution.
doi:10.2495/GEO060351
The study of the relation between the anthropic activities and the
environment [1] needs to take into account not only the physical, chemical and
biologic components of the environment self, but also the social and economic
dynamics of the community, the relations, the structure, the energetic balances
[2], in order to understand the whole induced impact.
Every action which is done on the territory produces an alteration of the
environmental parameters. Such alteration normally presents positive and
negative aspects and has to be compared to the induced benefits and to the
environmental and social costs.
The present study wants to study the negative environmental effects produced
by the great and continuative infrastructure works; it will be taken into account
the only environmental system, leaving to the specific disciplines the
complementary analyses of the social and economic dynamics and the relevant
costs and benefits.
The criticities of the environmental alterations have to be analyzed at first in
the planning phase, by the application of the study models of the ecosystem
response deriving to the anthropic pressure levels. The consequent decreasing of
the environmental state quality will produce a decreasing of the environmental
safety, which is the condition of absence of danger and of safeguard of the
population and ecosystem integrity. Such decreasing will be related to the check
and control procedures, put into act to manage the impacts.
2 Environmental quality, environmental impact

The environmental impact, according to the definition of Malcevschi [3] is the
consequence of the action of a source that, after a more or less complex sequence
of events, generates pressures on environmental targets, which could also be
altered.
In order to limit the impact due to an intervention on the territory, it is
necessary to study in the planning phase the possible modifications induced on
the environment and on the existing anthropic activities.
The dimension and the importance of the impacts is related to the possible
increasing of the risk on the safety and health of the exposed population [4], to
the variation of availability of the existing resources and to the impossibility to
reach the aims of safeguard and of environmental quality expressed by the local
administration and by the scientific and cultural community. The sum of such
factors implies a decreasing of the environmental safety.
These evaluations have to be integrated with the principles of sustainable
development, put into act by the local authorities and aimed to the integration of
all the interested social, politic, economic, administrative and legal subjects [5].
If the local authorities can put into act the criteria for a local sustainable
development, the real problem is the interaction with the global system (national
or international) [6] whose dynamics of sustainable development could be
different or incompatible with the local decisions.
The infrastructural networks are a typical example of action that often shows
the contradiction between the local and global development models that,
although pursuing the same principles, however they don’t collaborate on the
dynamics of environmental and social economic development.
The great infrastructure works show how the natural system factors involve
in particular the local system, without interacting, sometimes not at all, with the
global system.
If it is not possible to understand what will be the destiny of the global
ecosystem, it is possible and necessary to study the different scenarios of the
evolution deriving from the different strategies and politics of safeguard in the
local system.
The analysis of the ecologic systems and their interaction with the social and
economic systems, have moreover permitted to put into act adequate
programming and control instruments aimed to pursuit a sustainable
development of the territory.
In the next paragraph are underlined the correlation between some of these
instruments, with particular attention to the evaluation and control instruments.
3 Correlation environmental state – environmental evaluation

The instruments for the environmental evaluation are aimed to analyze the
criticities deriving from the perturbation of one or more elements of the
environmental sphere, of the health and of the territory.
In order to apply such instruments, it is at first necessary to analyze the level
of perturbation of the environment due to a specific pressure and to correlate
such perturbation to the existing environmental state. It is then necessary to
introduce some definitions that represent the modification of the environmental
state.
The “persistence” describes the state in which an environmental system does
not modify its inner structure, due to external interferences.
The “resilience” indicates the capacity of a system to maintain its own
structure, due to external interferences.
The “inertia” is the capacity of a system to maintain its own dynamicity.
The “collapse” is a fast modification of the system directed to a definitive
destruction.
It is possible to represent the evolution of the environmental state, due to
pressures generated by the realization of an intervention on the territory, with
different cases.
1) In the case 1, there is an event (construction of a work) on time t0. The
environment component passes from a condition of persistence to a condition of
resilience. If the system can support the intensity of the pressure, then the
environmental state, when the work is finished (time tf), could come back to
same beginning level and stay in a condition of persistence. The environmental
safety is safeguarded.
The environmental impact is represented by the colored area between the
straight line without intervention and the curve after the intervention (case1).
persistence
CASE 1
CASE 2
Limit of
resilience
inertia
resilience
collapse
CASE 3
t0 tf time
Environmental Environmental Safety Plans

Strategic Impact Environmental Management
Assessment Assessment System procedures
Figure 1: Effects of the impacts on the system.
2) If the pressure on the system is important (i.e. the great infrastructure

works), after the phase of resilience, the environmental state can reach different
levels of persistence, according to the entity of the impact, represented by the
area between the straight line without the intervention and the curve after the
intervention (case 2). In this case the system has been impacted, but it can come
back to a condition of persistence.
3) If the system cannot support the pressure of an intervention, the resilience
limit will be exceeded and when the work is finished, the environmental state
will present a lower level respect to the beginning condition of persistence, and
moreover it will decay (case 3). Such decay will reach the collapse in different
times, depending on the inertia of the system.
Considering the complexity of the conflicts due to the interaction between the
human activities and the environmental dynamics, the application field of the
environmental evaluation is wide [7].
From a general to a particular application, the evaluation can interest plan

and territorial programs (Strategic Environmental Assessment, Ecologic
Incidence Assessment), preliminary and definitive projects (Environmental
Impact Assessment, Ecologic Incidence Assessment, Safety Plans), existing
plants or executive projects (Integrated Environmental Prevention, Operative
Safety Plans, Environmental Management System).
It is possible to find a relation between such evaluation instruments and the
effects of the impacts on the examined system. The Strategic Environmental
Assessment, that represents a preventive instrument of global evaluation, is
directly related to the beginning state of the system, since it studies the phase of
programming, before the work is planned.
In this programming phase, the Environmental Impact Assessment studies the
system after the Strategic Environmental Assessment, but before the perturbation
state of the system, since it evaluates the project.
The results of the planning choices after an environmental impact evaluation
should make the system maintain its own structure, safeguarding not only the
compatibility, but also the sustainability. The Operative Safety Plans and the
Environmental System Management (ISO 14000, EMAS) can be acted in the
executive phase and are aimed to avoid the exceeding of the resilience limit.
For that situations on the limit of the environmental collapse, the operation of
control and mitigation of the impacts are more complex, since the system lost its
elasticity.
Such instruments should then be able to manage the impacts from the first
phase of execution of the infrastructure works, in order to make the system
maintain its inner elasticity.
4 Criteria for the environmental analysis and evaluation

The environmental evaluation of a project or the construction phase of an
infrastructure work is not a rigid scheme, but it is an dynamic instrument that
depends on the meaning of evaluation. In every case the evaluation is aimed to
check the relation between the intervention and the safeguard of the
environment, according to the principles of the environmental sustainability.
In order to make detailed evaluations of an area interested by a project, it is at
first necessary to make a descriptive analysis of the characteristics of the project
self. After a first descriptive analysis there should be a phase of qualitative
evaluation and, when it is possible to get data from available sources or direct
measurements, the evaluation can be quantitative. Such evaluation, through the
use of adequate instruments, can give and estimation of the alterations due to the
realization of the project.
The descriptive analysis can be done by check lists, yet used by the control
authorities [8], which are a selected list of environmental factors, structured on
the different target to be reached (description of the impact related to the project,
knowledge of the possible alternatives of the project, indication of the
environmental components with indicators).
The present study wants to show a check list which is structured beginning
from the description of the activities indicated by the project to the identification
of the risk, in order to represent a reference guide for the environmental
evaluation.
In tab. 1 there is an example of check list for the environmental analysis in
the construction of streets, applied to a classical macro phase, expressed by the
literature [9], of preparation, boundary and cleaning of the area interested by the
infrastructure work.
The activities are divided into phases and subphases and are linked to the
environmental aspects, which are defined by the EMAS as the elements of the
activities that have an interaction with the environment [10]. For each
environmental aspect the interested environmental component and the
consequent risk for the environmental safety have been identified. The risk for
the environmental safety is the possible modification of the environmental state,
of safety and health of the population that can be represented by pollution, by
consumption of raw material and energy, by an annoyance or an environmental
disaster.
In the check list specific indicators have been also identified. Such indicators,
which can be used in the following evaluation phase, represent and measure
particular environmental impacts. Moreover this checklist links the impact to the
interested phase of the activity (correlation source – effect).
The items in the check lists have to be identified according to the specific
activity and territory which is studied and have to be separated in sub items in
order to permit a complete evaluation of the environmental impacts.
In order to apply the indicators to the specific areas which have to be studied,
they should be able to represent the measures of the environmental monitoring. It
is important to take into account not only the possible damage conditions, but
also the annoyance ones. The indicator should then represent the beginning
environmental state and measure the variation of the examined parameter. The
comparison with the existing situation permits to evaluate if the environmental
condition is more or less critical (disturbance – annoyance – damage).
It is then necessary to get information on the environmental system
concerning the area interested by a project. Such information, with the analysis
of the environmental aspects of the project, make possible to identify and
evaluate, through specific indicators, every impact of an infrastructure work.
If the beginning environmental state is good, the impact of an infrastructure
work can not be critic, but its effect will be a decreasing of the environmental
quality. Moreover if the beginning environmental state is not good, a little
decrease of the environmental quality could generate the collapse of the system,
that means a condition of damage. The indicators should then be related to the
examined system in order to give important information for a correct
environmental evaluation [11].
Table 1: Example of check list for the environmental evaluation.
Detail of the Environmental Environmental Identification Indicators for

operative aspects components of the risk for the
phases the environmental
environmenta evaluation
l safety
MACRO PHASE: PREPARATION, BOUNDARY AND CLEANING OF THE AREA
Boundary Modification of Antroposphere, Local Sound level
the vehicular biosphere acoustical
traffic pollution
Atmosphere Local Concentration of
emission combustion
products
Installation of Antroposphere Visibile Reduction of
boundary impact visible cones
structures
Installation - - - -
of mobile
and
stationary
equipment
Elimination Elimination of Lithosphere Increasing of Concentration of
of loam soil, loam soil, land chemical
disafforestati disafforestation vulerability substances
on, ground
leveling Hydrosphere Increasing of Concentration of
land porosity chemical
substances
Ground leveling Atmosphere Dust Concentration of
production dust
Mucking Lithosphere Waste Quantity of
removal production waste
Movement of Atmosphere Release of Concentration of
dumpers and combustion combustion
equipment products products
Antroposphere, Acoustical Sound level
biosphere pollution
Conveyance Modification of Antroposphere, Reduction/incr Variation of flow
and watercourse biosphere ement of flow rate
movement of rate
surface
waters
5 Conclusions
The meaning of environment as the sum of the physical, chemical and biologic
conditions is reductive, a linear relation cannot take into account every interested
part [12], nor represent the relations of an Ecologic, Social and Economic
System. The best definition of environment is a whole of systems, with their
inner dynamics, relations, structures and energetic balances.
It is important to choice adequate indicators to make a correct evaluation, in
order to establish administrative and planning procedures aimed to face the
environmental impacts, due to the realization of a project inside a territory.
Such indicators should find a correlation in the different level of schemes for
the environmental evaluation, from the Environmental Strategic Assessment, to
the Environmental Impact Assessment arriving to the Operative Safety Plans, in
the case of the infrastructure works.
The final target is to pursuit a condition of environmental safety,
safeguarding the health and integrity of the population and the ecosystem
quality, operating then not only for the control of the pollution, but for the
environmental, economic and social sustainable development.
The only compatibility is not enough to assure a development finalized to
maintain a high ecosystem quality level. It is necessary then to be able to
evaluate the capacity of a territory to support the impacts, in order to establish
management procedures for the environmental safety.
To reach this target, the methodological instrument for the knowledge of the
territory, which is necessary to get information on the effects and modification
induced on the environment, should use the analysis and evaluation criteria of
the environmental risk described in this study. The operative steps to get
information for the environmental safety in the case of an infrastructure work,
should be following:
- analysis of the documents. This phase is necessary to understand the
legislation and the technical procedures which should be applied to an
infrastructure work. It is then necessary to identify and analyze the compulsory
and voluntary norms for the environmental, safety and health control and
safeguard. Moreover it is necessary to analyze the documents belonging to the
local administration and to the company concerning the risk for the
- environmental analysis. The pollution sources are identified and related to
their effects. The environmental state is evaluated by the use of appropriate
indicators.
- environmental evaluation. This phase wants to underline the environmental
criticities of the infrastructure work, by evaluating the environmental impacts,
through the studying of an evaluation criteria.
The application of such operative scheme to the infrastructure works will
permit to calibrate the analysis methodology indicated by the present study, and
will be object of thorough examination.
References
[1] Taylor, William M., The vital landscape : nature and the built
environment in nineteenth-century Britain, William M. Taylor. -
Aldershot ; Burlington : Ashgate, 2004, 252 pp.
[2] Afgan, Naim H. Sustainable development of energy, water and
environment systems : proceedings of the conference on sustainable
development of energy, water and environment systems, 2-7 June 2002,
Dubrovnik, Croatia, Naim H. Afgan, Zeljko Bogdan, Neven Duic'. -
Lisse : Balkema, 2004, 367 pp.
[3] Malcevschi S., Qualità ed impatto ambientale, Etaslibri, Milano, 1991,
355 pp.
[4] Gerrard, Simon, Environmental risk planning and management, Simon
Gerrard, R. Kerry Turner and Ian J. Bateman. - Cheltenham : Elgar, 2001,
615 pp.
[5] Clerico M., Pizzo G., Environmental issues for sustainability development
in underground works, IABSE Symposium ‘Towards a better built
environment – innovation, sustainability, information technology’,
Melbourne 11-13 September 2002; p. 106,107; CD p.1-8
[6] Naveh Z., Lieberman A., Landscape Ecology, Springer-Verlag, New
York, 1994
[7] Ravetz, Joe, City - Region 2020 : Integrated planning for a sustainable
environment, Joe Ravetz. - London : Earthscan, 2000, 307 pp.
[8] Arpa Piemonte, Sostenibilità ambientale dello sviluppo, Arpa Piemonte,
Torino 2002, 367 pp.
[9] Comitato Paritetico Territoriale Per La Prevenzione Infortuni, L’igiene E
L’ambiente Di Lavoro Di Torino E Provincia, Conoscere per prevenire
n.12, Valutazione dei rischi nel settore delle costruzioni, vol.1, CPT,
Torino, 2005, 802 pp.
[10] Regulation (EC) No 761/2001 of the European parliament and of the
council of 19 March 2001 allowing voluntary participation by
organisations in a Community eco-management and audit scheme
(EMAS), Official Journal of the European Communities L 114 Vol.44,
pag. 1 on 24 April 2001
[11] Australian And New Zeland Environment And Conservation Council
1996. Canberra : Dept. of the Environment and Heritage, 1996
[12] Bottero M., Mondini G., Valle M., Verso una pianificazione ambientale
compatibile: i piani territoriali di coordinamento, GEAM rivista della
Associazione Georisorse e Ambiente, No. 104, December 2001, pp 265-
269.
Section 9
Natural hazards and risks
M3 (Monitoring, Management and Mapping) –

tool for the solution of the conflict:
“people and environment”
Z. Boukalová, V. Beneš & P. Kořán
Cross Czech, G-Impuls, Czech republic
Abstract
Natural disasters are typical examples of people living in conflict with the
environment. Vulnerability of populated areas to natural disaster is partly a
consequence of spatial planning policies that failed to take account of hazards
and risks in land use zoning/development decisions. Thus it is important to
combine knowledge, technology, M3 and actors in the field of risk assessment
and land use zoning to achieve effective natural disaster prevention and
mitigation. Understanding geologic processes is essential to research fields, such
as engineering, environmental management, land preservation and restoration,
urban environment, soil and water pollution, soil erosion and landscaping.
Keywords: monitoring, management, mapping, natural disasters, maintenance
of dikes, geophysical methods, dipole electromagnetic profiling, sustainable
hazard mitigation, multi-hazard risk assessment.
1 Monitoring, management and mapping

In recent years, due to more and more frequently occurring weather effects of
extreme nature which cause disastrous floods, increased attention has been paid
to two main issues, in the context of the conflict “people and environment”:
- Inspection and maintenance of dikes and embankments.
- Risk mapping of natural hazards (floods) and spatial planning policies
definition.
doi:10.2495/GEO060361
1.1 Monitoring, management and maintenance of dikes
Inspection, monitoring, management and maintenance or remediation of dikes

and embankments is very important task, in recent Europe. Our experience
gained in the Czech Republic (projects IMPACT, FLOODsite, ARMONIA)
shows that inadequate attention has so far been paid to the documentation of dike
breaches and failures after extensive floods. Basic data on the reasons for, and
the extent and course of dike breaches are missing in the majority of the cases.
Exact data are seldom known, even from the recent disastrous floods in central
Europe that occurred in 1997, 1998 and 2002. The data are often incomplete and
of insufficient authenticity. However, it is evident that analyses of such
information, followed by appropriate adjustments and repairs of the dikes, may
significantly reduce the risk of occurrence of new dike breaches and failures. We
particularly talk about those dike segments where the reasons for destruction
were, for example, inappropriate dike structure, inappropriate material or
reduced stream channel capacity due to clogging. Furthermore, after analyzing a
database, it often turned out that dike breaches in these sections had occurred
repeatedly.
Statistical analysis of dike breach parameters may also allow some important
generalizations related to the causes and characteristics of breach in specific river
basins (catchment areas). For example, it turns out that the prevailing reason for
dike breach occurrences in Slovakia is liquefaction caused by seepages in the
underlying beds. The main reason for dike failures in Hungary is overtopping.
Entirely different mechanisms of dike breach occurrences of course require
different types of preventive dike modifications.
At present, dike maintenance and preventive repairs are based on a system of
visual inspection complemented by analyses of airborne or satellite photographs,
or sometimes on slow intrusive methods (boreholes drilling etc.) results. Only
rarely is the project documentation of dikes and embankments complemented by
detailed information on structures and material properties, i.e. information
acquired by engineering-geological investigation, drilling, laboratory tests of
soils, etc. The reason for this is the considerable cost of such investigation and
the large extent of the dikes. However, we believe that information on the nature
of materials and basic dike structure is essential for efficient failure prevention.
This particularly applies to old dikes for which construction documentation is
missing. Furthermore, in some countries (for example, developing countries or
countries of former East Europe) we may expect low quality of construction
work that may contribute to dike breach when stressed (see Fig. 1).
It is in this area that a package of geophysical methods can be of particular
value. Geophysical methods investigation and monitoring provide a continuous
image of physical properties of a dike body and, furthermore, this type of
investigation is relatively inexpensive. Last years, we concentrated on testing the
possibilities of application of the following geophysical methods:
- Geo-electric methods
resistivity profiling (RP), self potential method (SP), multielectrode method
(MEM), electromagnetic frequency method (EFM).
Figure 1: Example of inappropriate material in the core of a damaged dike.
- Seismic methods
shallow seismic method (SSM), seismic tomography (ST), multi-channel
analysis of seismic waves (MASW).
- Microgravimetric method.
- GPR method.
- Geomagnetic survey, gamma-ray spectrometric survey.
Potential now exists to apply an innovation: a Geophysical investigation tool,

which is based on generating electro-magnetic (EM) fields and mapping their
propagation through soils and structures. The method finds perturbations in the
EM fields arising from concealed boundaries or changed materials. Previous
tools and approaches have shown limited effectiveness and poor data
interpretation accuracy compared with conventional intrusive geotechnical
investigation and description methods. Speed has been affected by the need to
repeatedly remount transmitters and receivers.
The new technology, so-called Geophysical Methods Suite – GMS (using
GEM-2 tool), could provide a breakthrough in an area of science in which
infrastructure managers and engineers have been highly skeptical. Experience
gained in the Czech Republic shows that the new technology GMS could support
national government asset owners across Europe. The GMS is proposed to
enable rapid, economic and repeatable identification of non-homogeneities
(possible points of disruption) in line embankments and is especially promising
for the identification of problem or weak spots in line embankments for water
management and flood defense, where embankments could be subject to rupture
under extreme hydrological conditions. Thanks to fairly fast and economical
monitoring procedures it should also be possible to use the technology to

facilitate the maintenance of line embankments in a state which should prevent
leakage and ruptures. More, this technology should make it possible to check
fairly long sections of line embankments in relatively short time (about 10 km
per day), whilst maintaining adequate levels of precision.
What is innovative about the proposed GMS? Having used the new tool to
define the hot spots of an existing embankment system, detailed investigation,
maintenance and renewal efforts can be concentrated in a cost-effective way on
the critical parts of the embankments. The core part of the GMS is Dipole
Electromagnetic Profiling with the unique brand new and innovative apparatus
the GEM-2 (multi-frequency device).
In order to incorporate the geophysical methods into a complex of dike
prevention and maintenance, we first have to identify the effects that can be
monitored by these methods. Figure 2 illustrates an approach to incorporation of
geophysical methods into a dike and embankments maintenance program. From
the viewpoint of dike maintenance – dike breach, timing of the action is of
central importance.
The breach formation itself takes place at a time scale of hours, max. a few
days. A hazardous segment is evident, application of geophysical measurements
is not assumed here. However, except for overtopping, the remaining defects
mostly show somewhat hidden PRE-breach formation stage (for example,
seepage through the underlying beds, repeated seepage at an increased water
level, structure defects, etc.) which predisposes the point of future dike breach.
This stage often lasts for even tens of years and is our area of interest for the
application of geophysical methods.
The database of quick testing measurements, which is the basic component of
the monitoring system (see Fig.2), provides a basic description of dike materials
and structures, division of dikes into quasi-homogeneous blocks (i.e. dike
segments showing similar geotechnical and physical properties). Productivity of
measurement is rather high, based on the dike character ranging between 10 and
20 km of a dike per day. From the viewpoint of dike maintenance, these data are
an appropriate complement to a visual inspection, allowing us to assess relative
permeability of the dike material and its homogeneity and to detect subsurface
distribution systems reaching a dike, etc. This allows us to more precisely
identify problematic dike segments that are disturbed and weakened inside. All
types of dike-testing measurements should be linked to GPS.
The GMS system: is composed of 3 basic building blocks:
Quick testing measurement – fast and cheap measurement for basic evaluation
of the dike condition and homogenity within the whole river-basin. This method
is also the core for repeated (monitoring) measurement. As a method for this
purpose we suggest DEMP using multi-frequency tool (for example GEM-2).
Diagnostic measurement – detailed measurement of the eroded (non-

homogeneous) sections aimed at finding hidden defects of the dikes. The method
is based on the application of the set of geo-electric methods, especially multi-
probe resistance method MEM complemented by another independent method
based on the type of the defect searched for accordingly.
Geophysical Monitoring System

newly included in inspection and maintenance of the
dikes (together with airborne photographs analyses,
visual inspection, etc.) input
data
- better
fit
Large-scale,
Database of quick
laboratory and
testing easurements
mathematical
simulation
Monitoring of Diagnostics of
geomechanical problematic
properties segments better
understandi
ng of
breach
effective and timely maintenance P R E - Breach Formation

prevents the defects from occurring
Dike structure defects Piping Overtopping
Slope deformation
Actual dike condition
Repeated high water attacks Internal erosion
Seepages through or
below the dike
YEAR MONTH DAY HOUR
Time
Figure 2: A diagram of incorporation of geophysical methods into dike

maintenance.
Measurement of geotechnical conditions – geophysical measurement to

monitor geomechanical conditions of eroded dike sections. For the analysis of
dikes geomechanical characteristics; especially seismic methods and micro-
gravimetry will be used.
The GMS asset lies in the possibility of objective evaluation of dike

homogeneity and condition. Geophysical methods are suitable supplement for
current methods of checks (visual check, aerial and satellite pictures analysis).
Monitoring function of GMS lies in the analysis of relative changes of
geophysical parameters. GMS database construction requires so called initial
stage and so called following check stages.
Initial stage of GMS is based on quick testing measurement of dikes within the
whole area of the river-basin and following diagnostics measurement of selected
problem sections. These checks result in complex evaluation of dikes condition
in the river-basin including the suggestion of the necessary repairs. Special
selected sections may be checked using measurement of geotechnical condition.
Check stages are planned for the dikes which are constructed to protect some
place against high water level (flood protection). In Europe, the check stage
should be carried out after 3 years at the latest without the reference to flood
conditions. Check stages may of course be carried out based on agreement at
whatever times according to the needs of the dike owner/caretaker, e.g. in limited
time setting during floods. The check stage includes repeated quick testing
measurement and comparison of the acquired data with the data from previous
stages (when using the results to eliminate the influence of climatic conditions).
Thanks to analysis of repeated measurements we are able to locate time unstable
anomaly areas which often coincide with the places where the dike ruptures
occur. The check stage can be supplemented with diagnostic measurement if
needed.
The GMS system success is largely based on narrow co-operation between

geophysics specialists and dikes caretakers. They have large quantity of
information which can help in making the geophysical measurements
interpretation much more precise. Without mutual trust and communication the
GMS database program has no meaning.
1.2 Mapping of natural risks and hazards
The implementation of flood risk assessment procedures for spatial planning in

order to prevent and mitigate floods in urban areas is very important issue. From
October 2004, the project ARMONIA is trying to find right solution regarding
the implementation of multi-risk assessment procedures for spatial planning in
order to prevent and mitigate natural disasters in an urban environment. The
main objective of the project is to provide guidelines for EU standards on the
harmonisation of data, methodologies and maps related to all main natural
hazards acting on urban areas, incorporating consideration of climate change
impacts.
ARMONIA seeks to achieve outcomes that can mitigate the adverse effects of
natural phenomena through joint effort of the scientific community, technology
experts and users. The target is not only a scientific output, but a measurable
impact on policies/practices for disaster mitigation initiated within the period of
the project, which fits with Europe's goals regarding sustainable development in
supporting environmental and security policies by facilitating and fostering the
timely provision of quality data, information and knowledge, developing tools
and improving management practices.
Disaster management in Europe currently suffers from following main
reasons: There is no common shared strategy at European level for the
prevention and mitigation of natural disasters, mainly when dealing with the
integrated and combined impacts of natural hazards on modern society, including
secondary social effects. Main important initiatives are on regional or national
levels. As disasters are often affecting several countries with heavy
transboundary effects (as for example the floods 2002), there is important to
support common multidisciplinary EU activities focused on prevention and
mitigation of natural disasters.
There is the need for a modern “disaster science” which can better deal with
the complexity of reality in Europe (taking into consideration dynamic
systematic interactions). The common accepted “natural and induced
technological disaster management cycle”, including stages of prevention,
mitigation, preparedness, response and recovery is a theoretical view that is
usually marginally working in practice: the synergic interaction and implication
of various hazards and risks are missing. A new approach for the harmonization
and further development of all component of the disaster cycle is needed.
The effective use of available technological tools in the field of disaster
management by end users is not common in many European countries, especially
in some of the new member states. More, the sustainable hazard mitigation
programmes addressing both the short and long term consequences of their
implementation in a holistic manner have not been adequately included in the
agenda of European research, end users and stakeholders.
The project ARMONIA should bring the important outputs and solve the
issues as harmonisation of different risk mapping processes for standardizing
data collection/analysis, monitoring, outputs and terminology for end users and
optimisation of methodologies for hazard/risk assessment for different types of
disastrous events. More, the project will suggest the design of a harmonised
decision-making tool for applying hazard and risk mitigation in spatial planning
and optimisation of a guideline on natural hazard mitigation in the context of the
EU Environmental Assessment Directive (2001/42/EC), till end of year 2007.
Human and economic losses due to natural disasters continue to increase
world-wide. The high importance of disaster reduction policies was stressed at
the UN-World Conference on Disaster Reduction (January 2005, Kobe, Japan).
As a consequence there is a need to integrate knowledge, technology and actors
and to update state of the art into a disaster management approach that reflects
the complexity of the modern society in a realistic way.
2 Conclusions
There is an urgent need for better disaster and vulnerability reduction support
actions, in the catchment scale. In order to achieve overall management of water
resources, at river basin level in particular, it is of prime importance for decision-
makers (Directors of River Basin Organizations and Administrations, Basin
Committee members, representatives of Local Authorities and associations of
users) to have easy access to comprehensive, representative and reliable
information, at all relevant levels. The M3 (Monitoring, Management and
Mapping) tool, adapted on the modern society would be the best start to find
realistic and end users friendly “disaster and vulnerability reduction support
actions” in the context of the conflict: “people and environment”.
References
[1] ARMONIA: Applied multi Risk Mapping of Natural Hazards for Impact
Assessment, FP6 project n. 511208; web: www. armoniaproject.net/html4
[2] Boukalova Z., Beneš V.(2004): Case studies and geophysical methods.
Association of State Dam Safety Officials: Dam Safety 2004. Phoenix,
Arizona, USA, September 2004.
[3] Boukalova Z., Beneš V. (2005): Long-term monitoring of geotechnical state
of flood protection dikes using non-destructive geophysical methods, report.
Cross Czech a.s., Prague, 2005.
[4] Margottini C. (2006): Natural hazards and economic impact in Europe:
some concepts for the 7FP. International Symposium on Climate
Challenges. Brussels, 2-3 February 2006.
Gas hazard: an often neglected natural risk in

volcanic areas
W. D’Alessandro
Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo,
Palermo, Italy
Abstract
Volcanic areas release huge amounts of gases, which apart from having
important influences on the global climate could have strong impact on human
health. Gases have both acute and chronic effects. Carbon Dioxide and Sulphur
gases are the main gases responsible for acute mortality due to their asphyxiating
and/or toxic properties. On the contrary Mercury and Radon have important
chronic effects respectively for its toxicity and radioactivity.
The problem has long been neglected until the “Lake Nyos” catastrophe in
1986, in which about 1700 people were killed by a volcanic CO2 emission,
attracted the worldwide attention of the mass media.
In this paper we present some studies on gas hazards in three different
volcanic systems chosen for their different activity status: Mt. Etna (Italy),
characterised by frequent activity with a mean CO2 emission of about 450 kg s-1;
Pantelleria island (Italy) at present in quiescent status and a CO2 emission of
about 12 kg s-1; and Sousaki (Greece) a recent (Quaternary) but now extinct
volcano with a CO2 emission of about 0.6 kg s-1. In all three systems the main
problems arise from CO2 emissions while secondary problems are due to SO2
and Hg (Etna), H2S (Sousaki) and Rn (Pantelleria).
Keywords: gas hazard, Carbon Dioxide, Sulphur gases, Radon, Mercury.
1 Introduction
Pliny the Elder, who perished during the 79 A.D. eruption of Vesuvius in Italy, is
possibly the most famous known victim of volcanic gases. The description of his
death, made by his nephew Pliny the Younger (Letter 16, 6), clearly points to
cardio-respiratory collapse of a person with chronic respiratory disease.
doi:10.2495/GEO060371
Although it is unclear if the fatal causes were volcanic gases or fine suspended
ash or even a heart attack, the description closely resembles that reported for one
of the victims of volcanic gases at Aso volcano in Japan in recent times, who
suffered from pulmonary emphysema [1].
Gas manifestations have been recognized since olden times and often
sanctuaries were built near them since the Neolithic to get protection against
them from divinities. In these sanctuaries sometimes oracles gave their responses
under the narcotic effect of elevated CO2 concentrations. An important sanctuary
for example was placed at the Palici Lake in Sicily (5th cent. B.C.). The abundant
CO2 released from its shores and trough its waters was used to test the truth of
persons [2]. Standing on the shore of the lake the person had to swear in the
name of the sanctuary’s gods looking to the sky, and then had to touch the waters
with his hands. If he had no consequences he was saying the truth. Probably in
wind free days there was a higher proportion of ‘liars’. Nowadays fluxes in the
area are still very high [3] but no signs of the sanctuary is still present, the lake
was dried up and CO2 is industrially exploited for gas addition to beverages.
On February 20th 1979 at the onset of eruptive activity of Sinila crater on
Dieng Plateau, Indonesia, 142 persons, fleeing from the nearby Batur village,
died asphyxiated by a cloud of CO2-rich volcanic gas rolling down the flank of
the volcano [4]. But it was not until 1986 that a wide public knew about the risk
of volcanic gases. On August 21st of that year about 1700 people were killed and
850 injured by a massive gas release from Lake Nyos, Cameroon [5]. The gas
release of this volcanic crater lake, composed almost exclusively of CO2, was not
related to any paroxysmal volcanic activity. The release occurred on a calm night
and the gas cloud, heavier than air, flowed in a valley that was more than 250 m
below the surface level of the lake.
Recently Witham [6] published a new database of volcanic disasters and
incidents of the 20th century. Of the 491 reported events 11% were referred to
volcanic gases, accounting for 2016 (2% of the total) of the killed and 2860
(18%) of the injured people. As the author himself states, this database has to be
improved and completed and indeed only one of the many incidents occurred in
Italy has been reported. But it is equally impressive to learn that the Lake Nyos
disaster ranks at the 8th place for the number of dead and at the 5th place for the
number of injured in this database.
An important research project aiming at the study of diffuse degassing of the
whole Italian peninsula financed by Italian Civil Protection is currently in
progress [7]. The new awareness of the importance of determining the risk
connected with natural gases mainly of volcanic origin probably derives from the
expansion of Rome’s metropolitan area well within the area of Albani Hills. This
recent, possibly still active, volcanic area is characterised by anomalous gas
fluxes, which led to some fatal outcomes in the last years [8].
Sometimes the risk associated to gases is enhanced by improper human
operations. For example an area within the urban area of Rome is naturally
protected from high gas fluxes from the soil by shallow impermeable sediments.
But excavation or drilling operations sometimes perforate these impermeable
layers leading to massive gas outflows [8]. One of these episodes evidenced that
both of the medical and fire brigade structures were (and probably still are)
unprepared to face this kind of emergency. The doctors in fact didn’t recognize
the CO2 intoxication of 7 persons that lived close to the drilled well, and
attributed ailments to collective hallucination because no clinical signs were still
present at the emergency station. Returning to their homes these persons escaped
death only by chance and were eventually hospitalised. On the other hand the
remedy applied by fire brigade, to drill additional wells trying to exhaust the gas
source, had as the only consequence to exacerbate the problem that was fixed
only sealing all the drilled wells with the injection of special cements [8].
2 Hazardous gases
2.1 Carbon dioxide
Carbon dioxide is the more abundant gas, after water vapour, released by
volcanic activity. Total CO2 output of the different volcanic systems span over a
wide range of values, from about 450 kg s-1 of Mt. Etna (Italy) and Popocatepetl
(Mexico) down to values of less than 0.1 kg s-1 (Iwoyama, Japan) [9]. Open
conduit volcanoes, like Mt. Etna and Popocatepetl, emit this gas mainly through
the craters. On the contrary volcanoes with closed conduit, even if characterised
by intense fumarolic activity, release CO2 almost exclusively through diffuse soil
degassing [9]. Carbon dioxide annually released by volcanoes at the global scale
was estimated in about 300 Tg, which represents only a small fraction (1%) of
the CO2 released by human activities [10]. The geographical distribution of
natural CO2 release corresponds prevailingly to areas of active or recent (< 10
Ma) volcanism and seismicity [10]. Carbon dioxide has different origins:
biogenic (respiration), hydrocarbon oxidation, thermal or chemical breakdown of
limestone and mantle degassing. The first two processes very rarely create
dangerous natural CO2 accumulations. Carbonate minerals in the crust can
release huge quantities of CO2 both through thermal breakdown due to high
geothermal gradients and through reaction with acid hydrothermal fluids.
Magma ascent from the earth’s mantle is a very efficient CO2 degassing process.
Mantle-generated magmas contain up to 1.5% by weight of CO2, but its low
solubility leads to early (deep) gas separation during ascent of magmas.
Considering that CO2-depleted magmas could eventually solidify within the
crust, some volcanic system, like Stromboli Island (Italy), emit on long time
average by weight more gas than lava.
Normal CO2 concentration in the atmosphere at sea level is about 350 ppm
(by volume) but its concentration can rise if production exceeds consumption
and dissipation. Being heavier than air, in high flux areas, CO2 can accumulate in
topographic depressions and enclosures reaching concentrations as high as
100%. Carbon dioxide concentrations higher than 10% can be lethal to humans
and animals, and at concentrations above 20-30% even a few breaths can very
quickly lead to unconsciousness and death from acute hypoxia, severe acidosis
and respiratory paralysis [11]. Therefore, poorly ventilated places below and
immediately above ground such as caves, galleries, cellars, water wells, etc. can
be very dangerous in areas of anomalous CO2 emissions.
2.2 Sulphur gases
Sulphur gases follow carbon dioxide in order of abundance in volcanic gases.
The most important species are SO2 and H2S and their relative abundance is
fixed by the thermodynamic parameters within the volcanic system (pressure,
temperature, oxygen fugacity). Low temperatures and the presence of a
hydrothermal system favour H2S. Open conduit volcanoes display high SO2/H2S
ratios (> 20) while on the other hand volcanoes with low temperature fumarolic
activity display low ratios (< 0.1).
Sulfur dioxide is a highly irritant gas. In healthy persons 5 to 10 ppm SO2
cause eye, nose and throat irritation while 30 to 40 ppm can lead to respiratory
failure. In individuals with bronchial asthma or other chronic lung diseases
exposure to much lower levels (0.25 to 0.5 ppm) can be life threatening [1].
Concentrations of up to a few ppm can easily be achieved close to active
volcanic craters. Indeed six fatalities due to SO2 have been documented at Mt.
Aso volcano, Japan in the period 1989-1997, while other 59 persons had to be
hospitalized in the period 1980-1995 [1]. Five of the dead and 29 of the injured
suffered from chronic lung diseases, evidencing the higher risk for this class of
people. In 1996 a monitoring system for the measurement of SO2 in air was set
up in the crater area. No one could enter the area when SO2 level was exceeding
5 ppm, while an advise was given when it was exceeding 0.2 ppm for more than
5 min discouraging the visit of the crater. People suffering of respiratory and
cardiovascular diseases were asked to vacate the area when SO2 was exceeding
2.5 ppm for more then 5 min. But such precautions were not able to save the
lives of two persons on 23 November 1997. So it was decided to apply a more
rigorous criterion, forbidding the access if SO2 was exceeding 0.2 ppm for more
than 1 min or in the presence of instantaneous peaks exceeding 5 ppm [1].
Hydrogen sulphide is both an irritant and asphyxiant gas. Levels of up to
20 ppm have generally no effects on healthy people while for asthmatic persons
this level have to be reduced to 2 ppm [12]. Concentrations above 20 ppm cause
irritative effects on eyes and respiratory tract, above 50-100 ppm neurotoxic
effects appear and 500-1000 ppm are considered of immediate life danger [12].
Although the human odour threshold is very low (0.02 ppm) the warning signal
is lost above 150 ppm because of olfactory nerve paralysis by H2S itself. Like
CO2 also H2S, being heavier than air, tend to accumulate in closed and/or
depressed areas. Some of the fatal incidents with volcanic gases were attributed
to the effect of H2S released by low temperature fumarolic vents or by gas
bubbling through thermal springs. Furthermore, although not conclusive, some
studies evidenced also health effects from chronic exposure to H2S (0.1 – 2 ppm)
in the city of Rotorua (New Zealand), which is built on a geothermal field [13].
2.3 Radon
Radon is a natural radioactive gas being and intermediate product of the
radioactive decay series of Uranium and Thorium. Radon can easily enter the
human body by inhalation. Radioactivity of Radon and of its decay products has
been linked to an increase in the risk of developing lung cancer. Only smoking is
a greater risk factor for lung cancer with respect to Radon inhalation [14].
The concentration of Radon in the indoor air and in soil atmosphere depends
from many factors. One the most important is the content of parent elements
(Radium, Uranium, Thorium etc.) in the rocks of the subsoil or of the building
material. Furthermore, release and transport of radon are controlled by nature
and alteration of the containing minerals, moisture content, and by the nature
(i.e. carbon dioxide gas or groundwater) and flux of a carrier fluid [14].
In volcanic and geothermal areas higher risk of Radon accumulation occur
where differentiated magmas, enriched in parent elements, are involved, where
geothermal alteration is widespread and where anomalous gas fluxes are present.
2.4 Mercury
Mercury is a highly volatile, bioaccumulating toxic trace metal strongly enriched

in volcanic and geothermal emanations. The contribution of volcanic activity
(about 700 Mg yr-1), although highly debated, represents 20-40% of the global
natural emission [15]. Very high mercury concentrations in air (up to 40 µg m-3)
have been measured in volcanic and geothermal areas of Hawaii and of Iceland
[16], well above the guideline value of 1 µg m-3 recommended by WHO for
general population exposition and sometimes also above the occupational long-
term exposure limit of 25 µg m-3. Although mercury accumulation in
professionally exposed persons (volcanologists, guides, employed of the
geothermal industry) has been demonstrated [17], no study on the neurotoxic
effects of volcanic-derived mercury in this population has been until now
performed.
3 Gas hazard at selected volcanoes
3.1 Mt. Etna
Mount Etna, located in eastern Sicily, is the largest strato-volcano in Europe

(3300 m a.s.l.; base, 60-40 km) and one of the most active in the world. It grew
in proximity to the collision boundary of the African and Eurasian continental
plates, from repeated eruptions of alkali basalts-hawaiites over the last 200 ka
[18]. Its recent activity is characterized by permanent open-conduit passive
degassing, interrupted by paroxysmal activity (effusive to moderately explosive)
at the summit craters and/or newly formed flank craters. At present, Etna’s
central conduit feeds four summit craters called Voragine, Bocca Nuova, South-
East and North-East. Degassing at the summit craters has continued without
interruption in the last few decades, although only rarely have all the craters been
degassing contemporaneously. Mt. Etna is considered, on long time average, the
greatest point source of many volatile compounds to the atmosphere [18].
3.1.1 Lower flanks

Etna emits yearly about 1 Tg of CO2 diffusely through the soils of its flanks. This
huge amount represents about 10% of the total CO2 output, the greatest part
being released through the summit craters. Spatial distribution of diffuse
degassing is strongly controlled by the tectonic setting of the volcanic system
and two anomalous degassing areas on the SW and on the E lower flanks have
been recognized [18]. In these areas very high values of CO2 fluxes from the soil
(up to 500 g m-2 d-1) and of CO2 partial pressure in groundwaters (up to 5 atm)
are measured. These areas are also densely settled and intensively cultivated due
to the high fertility of the soils and to their huge groundwater resources.
Groundwater is exploited through wells and drainage galleries. Until recent times
wells were excavated with large diameters (generally 2 m) up to depths of 400
m. These wells, still diffused and in use, in the two anomalous degassing areas
display often dangerous accumulations of CO2 on their bottoms. People in the
area know the problem very well and all operations inside the wells are generally
made by expert persons under steady air-pumping, in worse cases with two or
more independent pumping systems connected to autonomous electric
generators. Such precautions warrant a reasonable safety and no fatal accidents
have been registered at least in the last 30 years period. But in the last years an
increasing trend of land abandonment left many of these wells unattended so that
unprepared persons could get inside unaware of the high risk. For the mitigation
of this risk it would be very important: i) to take census of all the old wells, at
least, in the two anomalous degassing areas and ii) regularly control their state
and if abandoned close all possible access and evidence the danger with adequate
warning notices.
Groundwaters are also extracted from the flanks of Mt. Etna with drainage
galleries (or horizontal wells). A few tens of these galleries have length of more
than 1000 m and water yields of 0.1-1 m3 s-1. Some of these galleries, built in the
anomalous degassing areas, emit together with the groundwater also huge
quantities of CO2, which represents a big problem not only inside the gallery but
also outside its exit. The gas, in fact, tend to follow the water flowing in the
canal at the exit of the gallery and in wind free days maintain dangerous
concentrations of some % up to distances of some tens of meters.
In May 1993 one fatal incident was registered at the gallery known as
‘Ponteferro’ near the village of S. Venerina on the eastern flank of the volcano.
There was no eye-witness of the episode but the reconstruction points to the huge
gas release as one of the main causes. The victim was an old man that kneeing
down to collect water from the canal with a bottle, loosed consciousness due to
the high CO2 concentration and eventually drowned falling in the canal. The
access to the area, where many people come to collect drinking water, was
immediately restricted to supervised periods and the canal was covered with a
metal grid.
3.1.2 Summit crater area

The access to the summit area is generally unrestricted in periods of low volcanic
activity. On the crater rims in the downwind direction lethal gas concentrations
can easily been reached. But only volcanologists have sometimes the necessity to
go inside the volcanic gas plume to collect samples and in this case they have to
wear efficient gas masks and leave the area as soon as possible. Strong winds
and water vapour condensation that evidences the plume help all other people to
avoid areas with dangerous gas concentrations. Dangerous conditions could arise
in rare cases, when wind blows very mildly or changes rapidly direction, but
fortunately only healthy people reach the crater rims after a hard climb and they
can stand higher concentrations. But there’re two areas close to the summit
craters, Pizzi Deneri (2800 m a.s.l. about 2 km NE from the summit craters) and
Torre del Filosofo (2900 m a.s.l. about 1 km S from the summit craters), where
tourists are transported with all-wheel drive busses. Often between these tourists
there’re elderly people that possibly suffer of chronic lung disease. No danger
advice is given to these people of the possible effects of volcanic gases. In my
opinion it is only by chance than until now there have been no serious
consequences. The few measurements made in these two areas, in fact, gave
monthly average values between 0.2 and 0.5 ppm [19], which are life threatening
for persons affected by chronic lung disease [1]. For this reason it would be
desirable to set up a warning and monitoring system like that of Mt. Aso in Japan
[1].
Torre del Filosofo is also an area of anomalous soil degassing connected to
old eruptive fractures. There is also an abandoned hut that was used for long time
as shelter for volcanological surveillance instruments. One of its rooms, partially
below ground level, used in winter times as emergency toilet, suffered for long
time of CO2 accumulation but fortunately ventilation was always enough to
prevent dangerous concentrations. In November 2003, during an eruption the hut
was completely covered by volcanic ash. In the following summer the hut was
partially uncovered to show its rests to the tourists, creating an access to some of
its rooms. But these rooms that could represent an attractive shelter for
excursionists during winter, in recent times (summer 2005) displayed very
dangerous CO2 concentrations up to some percent.
Recent measurements evidenced also high concentrations of mercury in the
summit crater area (E. Bagnato and S. Giammanco – personal communication)
with values of some µg m-3 in the volcanic plume and some hundreds of ng m-3
in the atmosphere in the Torre del Filosofo area. In this case the most exposed
persons are not the tourists but the guides that spend a lot of time in the area.
3.2 Pantelleria Island
The island of Pantelleria represents the top of a large active volcano developed
on the African continental crust. The volcano is located between Sicily and
Tunisia, within NW-SE trending tectonic depressions related to the opening of a
rift system developed during the Neogene–Quaternary [20]. The island is
dominantly composed by volcanics showing compositional variations ranging
from basalts to peralkaline rhyolites (pantellerites) ranging in age from 320 ka to
the present. The volcanic activity of the island is mainly characterized by violent
explosive eruptions and secondarily by basaltic effusive activity [20]. Its last
eruption occurred in 1891 about 5 km NW off its coast. The island is at present
characterized by widespread occurrence of surface hydrothermal manifestations

(fumaroles and thermal springs) [21].
Also Pantelleria island displays two anomalous degassing areas with cold
(mofettes, bubbling gases) and hot (fumaroles) gas manifestations and areas of
strong CO2 fluxes from the soil [21]. One of these mofettes is placed in a
topographically depressed area near Lake Specchio di Venere. In wind-free days
CO2 accumulates near the ground being a lethal trap for small animals. Such
conditions could be very dangerous also for human that would lay down resting
on the shores of the lake. Fortunately in recent times warning signals have been
set up, but it would probably be safer also to fence the most dangerous area
(some tens of m2). Furthermore in the inner of the island there’re many natural
saunas exploiting the heath of fumaroles. Almost all of them are fed by
fumarolic gases composed exclusively by water vapour and atmospheric air, but
a few, within one of the anomalous degassing areas, contain also dangerous
concentrations of CO2. These improvised saunas are built by foreign people that
live on the island only on holidays and are not aware of this type of risk.
Recently Radon measurements in soils and in dwellings of the island
evidenced for most of the island very high values [22]. The highest values were
detected in central and southern part of the island where the most radionuclide-
rich rocks crop out. The higher natural radiation doses absorbed by the local
population has been considered an additional risk factor in developing some kind
of cancer and indeed a statistical study confirmed their higher incidence [Brai M.
personal communication].
3.3 Sousaki
The Sousaki area is located in Greece, about 65 km west from Athens, near the
Isthmus of Corinth and represents the NW end of the active Aegean volcanic arc.
Here, sparse outcrops of dacitic rocks are the remnants of late-Pliocene to
Quaternary volcanic activity (4.0 – 2.3 Ma [23]), while widespread fumarolic
alteration and warm (35 – 45°C) gas emissions are still recognizable. Drilling
exploration assessed the presence of a low enthalpy geothermal field.
In the area showing the highest hydrothermal alteration, located along a
narrow valley, several small caves were dug in the past century to extract
hydrothermal alteration minerals (alunite, magnesite, sulfur). Some of these
caves display at present hydrothermal gas emission from their bottoms. The
gases, being denser than atmospheric air, flow on the grounds of the caves and
eventually spill out from the mouth of the caves dispersing in the atmosphere
after descending the flanks of the valley. The gases are composed of more than
90% of CO2 and have also high concentrations of H2S (1000 – 6000 ppm).
Hazardous concentrations of CO2 and H2S are measured inside the caves were
dead small animals are always found. But H2S dispersing in the surrounding
atmosphere displays concentrations of some ppm that produces surely annoying
smell and represents also a potential chronic health impact for the nearby living
persons.
4 Conclusions
In the 20th century more than 2000 persons died and nearly 3000 were injured by
volcanic gases. The most dangerous gas species is CO2, responsible of more than
90% of the victims and of the worst episodes (Lake Nyos and Lake Monoun,
Cameroon and Dieng Plateau, Indonesia), but lethal episodes are also attributed
to sulphur gases (SO2 and H2S). Furthermore also Mercury and Radon have
dangerous chronic effects on human health due to their toxicity and radioactivity.
Gas hazard is often disregarded because it is almost always connected to low or
absent volcanic activity when attention is low. But we know that volcanic gases
and especially CO2 can be released by volcanic systems up to some million years
after volcanic activity ended. It is therefore important not to underestimate
potential risks and the effort of the Italian scientific community under the
patronage of Civil Protection for the understanding and mitigation of this natural
risk goes in the right direction.
References
[1] Ng’Walali, P.M., Koreeda, A., Kibayashi, K., Tsunenari, S., Fatalities by
inhalation of volcanic gas at Mt. Aso crater in Kumamoto, Japan, Legal
Medicine 1, 180–184, 1999.
[2] Ferrara, F., Memoria sopra il lago de Palici ora lago Naftia in Sicilia,
Reale Stamperia, Palermo, 1805.
[3] De Gregorio, S., Diliberto, I.S., Giammanco, S., Gurrieri, S. & Valenza,
M., Tectonic control over large-scale diffuse degassing in eastern Sicily
(Italy), Geofluids 2, 273-284, 2002.
[4] Le Guern, F., Tazieff, H. & Faivre–Pierret, R., An example of health
hazard: people killed by gas during a phreatic eruption: Dieng Plateau
(Java, Indonesia), February 20th 1979. Bulletin Volcanologique 45, 153–
156, 1982.
[5] Baxter, P.J., Kapila, M. & Mfonfu, D., Lake Nyos disaster, Cameroon,
1986: the medical effects of large scale emission of carbon dioxide?
British Medical Journal 298, 1437-1441, 1989.
[6] Witham, C.S., Volcanic disasters and incidents: A new database, Journal
of Volcanology and Geothermal Research 148, 191–233, 2005.
[7] http://www.ingv.it/progettiSV/Progetti/Vulcanologici/vulcanologici_con_
frame.htm
[8] Carapezza, M.L., Ranaldi, M. & Tarchini, L., Gas hazard in the Roman
area: soil CO2 discharge and accidents induced by uncontrolled shallow
boreholes, Abstract Book of the 8th International Conference on Gas
Geochemistry ICGG 8, Palermo and Milazzo, Italy, 2-8 October 2005, p.
12, 2005.
[9] Pecoraino, G., Brusca, L., D’Alessandro, W., Giammanco, S.,
Inguaggiato, S. & Longo, M., Total CO2 output from Ischia Island
volcano (Italy). Geochemical Journal 39, 451-458, 2005.
[10] Mörner, N.A. & Etiope, G., Carbon degassing from the lithosphere,
Global and Planetary Change 33, 185-203, 2002.
[11] Henderson, Y. & Haggard, H.W., Noxious gases, Reinhold Pub. Co, 1943.
[12] World Health Organization, Hydrogen Sulfide: Human health aspects,
Concise International Chemical Assessment Document 53, 26 pp., 2003.
[13] Bates, M.N., Garret, N. & Shoemack, P., Investigation of health effects of
Hydrogen Sulfide from a geothermal source, Archives of Environmental
Health 57, 405-411, 2002.
[14] Appleton, J.D., Radon in air and water, in: Essentials of Medical Geology
(Selinus, O., Alloway, B. Centeno, J.A., Finkelman, R.B., Fuge, R.,
Lindh, U. & Smedley, P. eds.), Elsevier, 2005.
[15] Pyle, D.M. & Mather, T.A., The importance of volcanic emissions for the
global atmospheric mercury cycle, Atmospheric Environment 37, 5115-
5124, 2003.
[16] Siegel, S.M. & Siegel, B.Z., Geothermal hazard. Mercury emission,
Environmental Science and Technology 9, 473-474, 1975.
[17] Siegel B.Z., & Siegel, S.M., Mercury in human hair: uncertainties in the
meaning and significance of ‘unexposed’ and ‘exposed’ in sample
population, Water Air and Soil Pollution 26, 191-199, 1985.
[18] Calvari, S., Bonaccorso, A., Coltelli, M., Del Negro, C. & Falsaperla, S.
(eds.) Etna Volcano Laboratory, Geophysical Monography Series, AGU,
143, 2004.
[19] Aiuppa, A., D’Alessandro, W., Federico, C., Ferm, M. & Valenza M.,
Volcanic plume monitoring at Mount Etna by diffusive (passive)
sampling. Journal of Geophysical Research 109(D21), D21308, 2004.
[20] Civetta, L., Cornette, Y., Gillot, P.Y. & Orsi, G., The eruptive history of
Pantelleria (Sicily Channel) in the last 50 ka, Bulletin of Volcanology 50,
47-57.
[21] Favara, R., Giammanco, S., Inguaggiato, S. & Pecoraino, G., Preliminary
estimate of CO2 output from Pantelleria island volcano (Sicily, Italy):
evidence of active mantle degassing, Applied Geochemistry 16, 883-894,
2000.
[22] Cinti, D., Pizzino, L., Galli, G., Quattrocchi, F. & Voltattorni, N., Radon
and thoron soil survey in the Pantelleria island: first results. Abstract book
of the Annual Assembly of Italian National Group for Volcanology,
Naples, 20-22 December 2004.
[23] Pe-Piper, G. & Hatzipanagiotou, K., The Pliocene volcanic rocks of
Crommyonia, western Greece and their implications for the early
evolution of the South Aegean arc, Geological Magazine 134, 55-66,
1997.
Section 10
Remote sensing
Identification of As-bearing minerals

associated with mine wastes from former
metalliferous mines in France using
laboratory reflectance spectra
V. Carrère
Laboratoire de Planétologie et Géodynamique, UMR CNRS 6112,
Université de Nantes, France
Abstract
Arsenopyrite-rich wastes from former metalliferous mines in the Massif Central
Region, France, contribute, through intensive leaching, to the formation of thin
layers of As-Fe crusts on the tailing surfaces acting as a cement. When subject to
rainfall, acid mine drainage develops, As is remobilized and trapped secondarily
by oxyhydroxides or sulphates such as goethite, hematite, jarosite, etc.,
depending on water pH. The aim of this study is to assess the potential of
hyperspectral remote sensing in this particular environment (small outcrops,
important leaching process, vegetated environment). Samples were collected in
the field from tailings and stream bottoms in various sites. Reflectance spectra of
the samples were acquired in the laboratory using a field spectrometer under
artificial illumination (0.4–2.5 µm spectral range). Mineral composition was
estimated by comparing reflectance spectra to reference spectral libraries. This
was done after continuum removal, using two different techniques, in order to
minimize the influence of factors such as grain size or moisture content.
Geochemical and SEM-EDS analyses were performed to gather information on
the mineralogy. Preliminary results from spectral analyses show clear evidence
of jarosite, goethite or limonite, schwertmannite and scheelite in various
proportions, depending on crust types, confirmed by SEM-EDS analyses.
Keywords: surface spectral reflectance, mine wastes, continuum removal, iron
oxyhydroxides.
doi:10.2495/GEO060381
1 Introduction
Mining activities generate a multiform pollution due to sulphides oxidation
present in the tailings. Their contact with atmospheric conditions can lead to
Acid Mine Drainage (AMD) (Johnson and Thornton [1]; Webster et al. [2]), a
remobilisation of toxic elements (As, S, Pb, …) which could be immobilized in
the sediments, the soils, by a more or less stable trapping. Some of these
elements can be temporarily fixed in different ways: precipitation as oxides,
hydroxides or sulphates (Herbert [3]), and absorption onto different materials
(Karathanasis and Thompson [4]). Due to its presence as a major element
accompanying many types of ore deposits, As is present in numerous mining
areas. Because of its toxicity, its solubility has been extensively studied (e.g.
Azcue et al. [5]; Hunt and Howard [6]; Roussel et al. [7]). Richardson and
Vaughan [8] studied the early stages of arsenopyrite oxidation using
spectroscopic investigations. They showed that in conditions of low pH, such as
those found in tailings dumps, arsenopyrite oxidizes rapidly and forms highly
soluble As compounds as well as Fe(II) or Fe(III) arsenites or arsenates.
According to local chemical conditions, As, Fe and S may be exported in
drainage waters (Nesbitt et al. [9]; Savage et al. [10]). These elements could
precipitate as authigenic minerals (Scott [11]; Jambor [12]) or amorphous
material or can be adsorbed onto Fe oxides, hydroxides, or suspended material
(Pierce and Moore [13]; Tsung-Hui et al. [14]; Manning et al. [15]).
In the French Massif Central, mining activities were greatly developed in the
twentieth century. Thus, before the 1980’s abundant mining tailings were
generated and laid in dumps without any remediation. The present study focuses
on one site, the site of Enguialès, located in the Aveyron Department, France
(Figure 1), which was slightly remediated. Courtin-Nomade et al. [16] provided
a detailed description of mineralogical and geochemical conditions of the tailings
which is used as a reference for analyzing the spectral information. As-rich Fe
cements are well developed as weathering crusts within the dumps. The size of
the site (roughly 1 km2) is well suited for estimating the efficiency of
hyperspectral reflectance data, a non-destructive approach, to identify and map
the various minerals that could potentially be used as markers for As
contamination. Previous studies on other sites clearly showed (see for example
Swayze et al. [17]) that hyperspectral reflectance data allow to identify mineral
phases onto which As can be adsorbed (iron oxyhydroxides, clays) based on
diagnostic absorption features. Moreover, mineral assemblages can be used as
indicator of acidity or chemical conditions that can control As mobility in soils
and rivers.
We present preliminary results on the potential of hyperspectral surface
reflectance measurements to identify mineralogical phases susceptible to have
high As concentration and/or mineral associations indicative of soil acidity in
such a specific context. More classical approaches such as mineralogical,
chemical and SEM-EDS analyses were also used to validate these identifications.
1.1 Site description
This tungstiferous deposit is located in the “La Châtaigneraie” district, in the

south of the French Massif Central (Figure 1). Mineralized veins are mainly
composed of wolframite, scheelite, pyrite and arsenopyrite. The mining activities
ceased in 1979. The tailings represent 300 000 t directly spread out in a steep
slope (~35°) on about 22 000 m2. The tailings are disposed as a pile ravined by
AMD and meteoric waters. The remediation that was done on this site consisted
of closing galleries access and gathering all the waters in only one dewatering
point, which induced AMD phenomenon. Currently only the meteoric waters
circulate across the tailings. Nevertheless, the Eh-pH conditions are oxidant and
always very acidic with pH = 2.8 +/- 0.3. The oxidation of the tailings is
expressed by local iron oxyhydroxides crusts, especially well developed in the
gullies. Grains of ore gangue (quartz, muscovite and tourmaline) mainly
constitute these iron crusts. These primary minerals are bound by a Fe-As rich
cement. The richest As product corresponds to an amorphous iron “arsenate”
which is mainly observed in the ochre cement rather than in the red cement. In
the current Eh-pH conditions in Enguialès, the estimated solubilities show a high
potentiality to release As from the identified amorphous arsenate. On the
contrary, the most crystallized products as goethite or jarosite, in which As is
also trapped (Courtin-Nomade et al. [18]), appear as stable minerals in the long
term (Stahl et al. [19]).
Figure 1: The Enguialès mine test site: General view and site location (after
Courtin-Nomade et al. [16]).
1.2 Sample collection
The various surface types that can be observed at the tailings surface were
sampled in the field in December 2002. Only the surface was scraped in order to
be representative of what can be observed by field or airborne hyperspectral
instruments. Samples collection was based on surficial visual characteristics
(color and aspect). When compared to descriptions published in the literature for
this site (Courtin-Nomade et al. [16]; Courtin-Nomade et al. [18]), they could be
attributed to 5 main categories: (1) micaschists (in place bedrock); (2) grey sand;
(3) ochre cement; (4) red cement and (5) dark-reddish cement.
2 Spectral analyses
Spectral absorption features observed in the visible to short-wave infrared
wavelength region result from several distinct processes. In the spectral range
from 0.4 to 1.2 µm, absorption features are produced mainly by energy level
changes in the valence of transition metals (crystal field transitions), by paired
excitations of metal cations, or by charge transfer between metal cations and
their associated ligands (Hunt and Ashley [20]; Sherman and Waite [21];
Burns [22]; Clark [23]). Absorption features in the SWIR region (1.3–2.5 µm)
are generated by molecular vibration processes: for example, the 1.4 and 1.9 µm
absorptions are related to the first overtones of the water O-H stretching
fundamentals and the combination water O-H stretching with H-O-H bending
vibrations, respectively (Hunt et al. [24]).
Laboratory spectral measurements in the VIS-SWIR spectral range of
samples of the five surface types were recorded using a Spectralon TM white
reference standard and an ASD FieldSpec FR spectrometer under artificial
illumination. Figure 2 shows the resulting reflectance spectra. The five spectral
signatures are rather different and show most of their distinctive absorption
features in the visible-near infrared (VNIR). The drop-off toward the U.V.
corresponds to charge transfer between ferric cations and adjacent oxygen
anions. The strong absorption edge between 0.4 and 0.6 µm is related to paired
excitations between magnetically coupled ferric cations (Sherman & Waite [21],
Crowley et al. [25]), while the broad absorption bands around 0.9 µm can be
attributed to crystal field absorption bands in ferric (0.75–0.95 µm) or ferrous
(0.9–1.1 µm) iron bearing minerals. Features observed around 2.20 µm are
attributable to the Al-O-H combination bands while those observed between 2.25
and 2.38 µm are due to Fe-O-H and Mg-O-H combination bands
(Crowley et al. [25], Clark [23]). Based on these observations, it is clear that the
different surfaces can be discriminated from one another. The next step consists
of identifying the various minerals that contribute to their composition using the
wavelength position of the absorption features. Comparison with spectra of pure
reference minerals was performed on “raw” spectra and after continuum
removal.
Figure 2: Reflectance spectra of typical surface types.
2.1 Continuum removal
The overall shape of a reflectance spectrum (or continuum) is related to physical

properties of the surface (e.g. grain size, roughness, moisture content, local
slope, etc.) which may change from place to place despite a homogeneous
composition. In order to better identify a spectral feature by its wavelength
position, it must be isolated from these other effects. The first step is continuum
definition and removal. Following the definition of Clark and Roush [26], the
continuum needs tie points at local maximum reflectance but there might be no
stable tie points toward the blue or the infrared for natural surfaces. Therefore,
the continuum can be defined in different ways: (1) using straight lines in
wavelength, linking local maxima between absorption features, including
starting and ending wavelength; this can be called the “straight line” approach
(Figure 3A); (2) using a gaussian decomposition for the absorption features and a
straight line in wavenumber for the continuum to overcome the instabilities at
both end of the spectrum; this is done by the Modified Gaussian Model (MGM)
developed by Sunshine et al. [27] (Figure 3B). The spectral deconvolution by
MGM uses a non-linear least-squares analysis. The measured spectrum
reflectance at a given wavelength is fitted by a superposition of n Gaussian
distributions (each is defined by three parameters: central wavelength, half-width
and negative strength) and a continuum curve (slope and intercept for a straight
line in wavenumber). The inversion process must start with an initial set of
parameters that includes absorption band centers and full width at half
maximum. These parameters are chosen by observing the spectral shapes and
following a priori knowledge about spectral characteristics (Figure 3B).
Once the continuum line is established, the continuum-removed spectra are

calculated by dividing the original reflectance values by the corresponding
values of the continuum line. As the tie points can be different for the various
surfaces depending in their mineralogical composition, the spectra were treated
globally.
Figure 3: A - Continuum removal using the “straight line” approach; solid line:
original spectrum (bottom) and after continuum removal (top), dashed
line: continuum. B - Principle of the spectral modelling using the
MGM; solid line: original spectrum (bottom) and after continuum
removal (top), dashed line: continuum, dotted line: gaussian functions
used to model the original spectrum.
2.2 Mineral discrimination based on spectral features
A reference spectral database, published by Crowley et al. [25] containing the

main minerals that can be encountered in sulphide-rich mine waste piles was
used. Only minerals supposed to be stable in the environmental conditions of our
test site (exposed settings, dry conditions), were selected for comparison.
Primary ore minerals (muscovite, scheelite or dravite) were also included but not
quartz or feldspar since these minerals, although present, do not show any
absorption features in the wavelength range considered here. A list of reference
minerals and formula is presented in Table 1. The 1.4 and 1.9 µm absorption
features attributable to H2O were excluded of the identification procedure since
they are present in most minerals and therefore not discriminant.
Table 1: Reference database mineral names and formula.
Mineral Formula
Dravite (Tourmaline) NaMg3(Al,Fe)6Si6O18(BO3)3(OH)4
Ferrihydrite 5Fe23+O3*9H2O
Fibroferrite Fe3+(SO4)(OH)*5H2O
Goethite α-Fe3+O(OH)
Hematite α-Fe23+O3
Jarosite KFe33+(SO4)2(OH)6
Limonite Mixture of hydrated iron oxides
Muscovite KAl2(AlSi3)O10(OH)2
Paracoquimbite Fe23+(SO4)3*9H2O
Rozenite Fe2+SO4*4H2O
Scheelite CaWO4
Schwertmannite Fe163+O16(OH)12(SO4)2
Szomolnokite Fe2+SO4*H2O
2.2.1 “Raw” spectra

Precise location of absorption features is more difficult to define on raw
reflectance spectra, particularly in the case of broad absorptions. In a first
approximation, jarosite was identified in grey sand, micaschists and dark-reddish
cement. Goethite is found in ochre and red cement, schwertmannite in
micaschists and possibly in dark-reddish cement. Scheelite is present in grey
sand, micaschists and ochre cement and muscovite in dark-reddish cement.
2.2.2 “Straight line” continuum removal

Removing the continuum helps locating more precisely the absorption positions.
Results show that it is more likely that hematite rather than goethite is present, or
more probably a mixture of hydrated iron oxides such as limonite, based on the
position of the broad absorption feature around 0.9 µm. The presence of jarosite
in micaschists and grey cement, and of schwertmannite in dark-reddish cement is
confirmed. Dravite is clearly present in micaschists and possibly in grey sand
and ochre cement. Exact match with scheelite and muscovite is difficult probably
because of mixing effects, and of the close positions of their absorption features
around 2.20 µm.
2.2.3 Using MGM gaussian function positions

Deconvolution of spectra using the MGM produces more gaussian functions than
“realistic” absorption features. Therefore, it is harder to decipher the results. This
technique appears also highly sensitive to input parameters which require some a
priori knowledge. Several runs were necessary to adjust the parameters and reach
a reasonable fit. Although the wavelength positions of the gaussian function did
not exactly match the absorption positions determined in the previous steps, the
presence of scheelite in grey sand and of hematite rather than goethite was
confirmed. This approach requires further improvement in order to be
operational on spectra of natural surfaces. A thorough sensitivity analysis of the
deconvolution to input parameters is also required.
3 SEM-EDS and XRD data

XRD, performed on a limited number of samples due to time constraints,
confirmed the presence of jarosite, dravite and muscovite in the ochre cement.
Indications of quartz and chlorite or serpentine were also found but not for clays.
SEM-EDS analyses were performed on minerals appearing to be statistically
representative of the samples. These analyses confirmed the presence of primary
ore minerals such as muscovite, quartz, feldspar and garnet as well as pyrite,
ferberite or wolframite [(Fe, Mn)WO4]. Hydroxides, hydrated arsenates and
phosphates, also present, were harder to identify since (OH) or (H2O) quantities
cannot be estimated accurately with this technique.
4 Conclusion
These preliminary results show that identifying specific minerals present in
cements developed at the surface of tailings can be complicated. In principle,
many ferric oxide and sulphate minerals should be distinguishable by using field
spectral measurements, as well as existing hyperspectral sensors. In practice
however the ability to discern individual iron sulphate or oxide species is limited
by the occurrence of the minerals in complex mixtures with each other and with
other surficial materials. Mineral mixing, grain size and mineral composition are
important factors influencing the spectral signature, making direct comparison
between reference and measured reflectance spectra difficult. Selecting the
proper endmembers to perform spectral unmixing is rather complicated because
of the different sources of changes previously mentioned. This study showed that
removing the continuum improved identification as it allows one to focus on
absorption wavelength positions. It might be necessary to work on specific
absorptions individually to improve the results. Using the MGM technique
should give access to more quantitative information such as absorption depth and
area that can be related to concentration or fractional composition but this
approach needs to be more thoroughly evaluated when applied to natural
surfaces. Its sensitivity to input parameters needs to be characterized. Finally,

detailed analysis of the chemical and mineralogical composition of the samples
is necessary to better understand the optical properties of such surfaces.
References
[1] Johnson, C.A. & Thornton, I., Hydrological and chemical factors
controlling the concentrations of Fe, Cu, Zn and As in a river system
contaminated by acid mine drainage. Water Res., 21, pp. 359-365, 1987.
[2] Webster, J.G., Nordstrom, D.K. & Smith, K.S., Transport and natural
attenuation of Cu, Zn, As, and Fe in the acid mine drainage of Leviathan
and Bryant Creeks, Environmental Geochemistry of Sulfide Oxidation,
eds. Alpers, C.N., Blowes, D.W., Am. Chem. Soc. Symp. Series, 550, pp.
244-260, 1994.
[3] Herbert, R.B., Properties of goethite and jarosite precipitated from acidic
groundwater, Dalarna, Sweden. Clays and Clay Minerals, 45, pp. 261-
273, 1997.
[4] Karathanasis, A.D. & Thompson, Y.L., Mineralogy of iron precipitates in
a constructed acid mine drainage wetland. Soil Sci. Soc. Am. J., 59, pp.
1773-1781, 1995.
[5] Azcue, J., Muroch, A., Rosa, F. & Hall, G., Effects of abandoned gold
mine tailings on the arsenic concentrations in water and sediments of Jack
of Clubs Lake, B.C. Environ. Technol., 15, pp. 669-678, 1994.
[6] Hunt, L.E. & Howard, A.G., Arsenic speciation and distribution in the
Carnon Estuary following the acute discharge of contaminated water from
a disused mine. Marine Poll. Bull., 28, pp. 33-38, 1994.
[7] Roussel, C., Bril, H. & Fernandez A., Arsenic speciation: Involvement in
evalution of environmental impact caused by mine wastes. J. Environ.
Qual., 29, pp. 182-188, 2000.
[8] Richardson, S. & Vaughan D.J., Arsenopyrite: a spectroscopic
investigation of altered surfaces. Mineral. Mag. 53, pp. 223-229, 1989.
[9] Nesbitt, H.W., Muir, I.J. & Pratt, A.R., Oxidation of arsenopyrite by air
and air-saturated, distilled water, and implications for mechanism of
oxidation. Geochim. Cosmochim. Acta, 59, pp. 1773-1786, 1995.
[10] Savage, K.S., Tingle, T.N., O’Day, P., Waychunas, G.A. & Bird, D.K.,
Arsenic speciation in pyrite and secondary weathering phases, Mother
Lode Gold District, Tuolumne County, California. Appl. Geochem., 15,
pp. 1219-1244, 2000.
[11] Scott, K.M., Solid solution in, and classification of, gossan-derived
members of the alunite-jarosite family, northwest Queensland, Australia.
Am. Mineral., 72, pp. 178-187, 1987.
[12] Jambor, J.L., Nomenclature of the alunite supergroup. Can. Mineral, 37,
pp. 1323-1341, 1999.
[13] Pierce, M. & Moore, C., Adsorption of arsenite and arsenate on
amorphous iron hydroxide. Water Res., 16, pp. 1247-1253, 1982.
[14] Tsung-Hui, H., Shang-Lien, L., Cheng-Fang, L. & Dar-Yuan, L.,

Characterization of arsenate adsorption on hydrous iron oxide using
chemical and physical methods. Colloids and Surfaces A: Physiochem.
Eng. Aspects, 85, pp. 1-7, 1994.
[15] Manning, B.A., Fendorf, S.E. & Goldberg, S., Surface structures and
stability of arsenic (III) on goethite: spectroscopic evidence for inner-
sphere complexes. Environ. Sci. Technol., 32, pp. 2383-2388, 1998.
[16] Courtin-Nomade, A., Bril, H., Néel, C. & Lenain, J.F, Arsenic in iron
cements developed within tailings of a former metalliferous mine –
Enguialès, Aveyron, France. Applied Geochem., 18, pp. 395-408, 2003.
[17] Swayze, G.A., Smith, K.S., Clark, R.N., Sutley, S.J., Pearson, R.M.,
Vance, J.S., Hageman, P.L., Briggs, P.H., Meier, A.L., Singleton, M.J. &
Roth, S., Using imaging spectroscopy to map acidic mine waste.
Environmental Science and Technology, 34, pp. 47-54, 2000.
[18] Courtin-Nomade, A., Néel, C., Bril H. & Davranche, M., Trapping and
mobilisation of arsenic and lead in former mine tailings – Environmental
conditions effects. Bull. Soc. Geol. France, 173, pp. 479-485, 2002.
[19] Stahl, R.S., Fanning, D.S. & James, B.R., Goethite and jarosite
precipitation from ferrous sulfate solutions. Soil Sci. Soc. Am., 57, pp.
280-282, 1993.
[20] Hunt, G.R. & Ashley R.P., Spectra of altered rocks in the visible and near
infrared. Economic Geol., 74, pp. 1613-1629, 1979.
[21] Sherman, D.M. & Waite, T.D., Electronic spectra of Fe3+ oxides and
oxide hydroxides in the near IR to near UV, American Mineralogist, 70,
pp. 1262-1269, 1985.
[22] Burns, R.G., Mineralogical application of crystal field theory. Cambridge
University Press, Cambridge, 1993.
[23] Clark, R.N., Spectroscopy of rocks and minerals and principles of
spectroscopy (Chapter 1). Manual of Remote Sensing, ed. A.N. Rencz,
John Wiley and Sons, New York, pp. 3-58, 1999.
[24] Hunt, G.R., Salisbury, J.W. & Lenhoff, C.J., Visible and near-infrared
spectra of minerals and rocks: IV. Sulphides and sulphates, Modern Geol.,
3, pp. 1-14, 1971.
[25] Crowley, J.K., Williams, D.E., Hammarstrom, J.M., Piatak, N., I-Ming
Chou & Mars, J.C., Spectral reflectance properties (0.4 - 2.5 µm) of
secondary Fe-oxide, Fe-hydroxide, and Fe-sulphate-hydrate minerals
associated with sulphide-bearing mine wastes, Geochemistry:
Exploration, Environ., Analysis, 3, pp. 219-228, 2003.
[26] Clark, R.N. & Roush, T., Reflectance spectroscopy: quantitative analysis
techniques for remote sensing applications. J. Geoph. Res., 89, pp. 6329-
6340, 1984.
[27] Sunshine J., Pieters, C.M. & Pratt, S.F., Deconvolution of mineral
absorption bands: an improved approach, J. Geoph. Res., 93, pp. 6955-
6966, 1990.
Rapid mapping with remote sensing data

during flooding 2005 in Switzerland by
object-based methods: a case study
Y. A. Buehler, T. W. Kellenberger, D. Small & K. I. Itten
Remote Sensing Laboratories (RSL)/NPOC, Department of Geography,
University of Zurich UZH, Switzerland
Abstract
Rapid mapping and monitoring with remote sensing techniques is an important
source of information for decision-makers faced with large scaled disasters. In
August 2005 several regions in central Switzerland were affected by severe
flooding. The “National Emergency Operations Centre” (NEOC) of Switzerland
invoked the International Charter on “Space and Major Disasters”, requesting
support by remote-sensing data for disaster-management. In this paper, a SPOT-5
and a RADARSAT-1 satellite-scene acquired a few days after the flood-peak are
described and processed to map the affected area with object-based methods. The
main difficulties with optical images are the small extent of the Swiss landscape,
the large height-differences in the relief, shadows, clouds and snow covered
areas. Comparing different approaches, it was found that beside the NIR and
SWIR band the application of a DEM and a VISNIR water index showed the
most promising results for a fast discrimination of the affected areas. The object-
based classification is mainly dependent on threshold-values and Boolean
operators. As investigations show, the spatial resolution of the acquired radar
dataset in this case is not sufficient to map the affected areas. The results of the
classification based on SPOT-5 scenes are up-to-date maps of flooded areas. We
show that maps for decision-makers can be produced using auxiliary
topographical and land use data. Aerial infrared images acquired from
SWISSTOPO after the flooding were used to test the accuracy of the result.
Procedures for rapid mapping applications in future disaster cases are discussed.
Keywords: rapid mapping, flooding, disaster response, international charter on
space and major disasters, natural hazards, object-based classification.
doi:10.2495/GEO060391
1 Introduction
Disaster mapping with remote sensing data has been investigated for several
years. In the past, the poor availability of appropriate sensors, low temporal
resolution, time-consuming data acquisition and distribution, limited image
analysis techniques and computer resources – and in addition, the lack of missing
awareness about the existence of products limited widespread application
Iglseder et al. [1]. Today, major advances reduce these obstructions, enhancing
possibilities for a useful disaster-management tool. In general, two types of rapid
mapping products are required by end-users: overview-maps of affected areas
and damage-maps combined with additional information (e.g. land-use type of
flooded area, destroyed traffic routes, changes in flood-levels) Allenbach et al.
[2]. This paper focuses on the detection of flooded areas and a raw differentiation
of water depth. The creation of overview-information is not required due to the
area-wide availability of 1:25’000 topographical maps.
Organisations active in the domain of rapid mapping associated with the
International Charter on “Space and Major Disasters” in Europe today are for
example UNOSAT in Geneva Switzerland, SERTIT in Strasbourg France and the
ZKI at the DLR in Oberpfaffenhofen Germany. The majority of past charter calls
concerned areas of a large extent in regions of mainly even topography.
Examples are the Elbe-River flooding in autumn 2002 and the destructive Indian
Ocean Tsunami in December 2004. In contrast to these disasters types, the
flooded areas investigated in this paper are of small spatial extent in the rough
topography of the Swiss Alps. An overview of mapping natural hazards in alpine
region is provided in Metternich et al. [3].
In August 2005 a 5b-cyclone carried large quantities of humid air from the
Adriatic Sea to the central Alps. The rainfall that ensued reached intensities
above 300 l/m2 over two days, which had never been measured in situ before.
Due to a snow line over 3000 meters above sea level and wet conditions in the
days before the disaster, the soils were already saturated and could not absorb the
additional intense rain Frei [4]. The levels of rivers and lakes rose and flooded
farmland, roads and urban areas. The steep terrain of the central Alps caused
landslides and mudflows so that many people had to be evacuated. An overview
of the main affected regions in Switzerland and the study area is shown in Figure
1. Especially the large extent of affected areas is exceptional for a flood event in
Switzerland. Three days after the rainfalls the “National Emergency Operations
Centre” (NEOC) invoked the International Charter on “Space and Major
Disasters” requesting support by remote-sensing data for disaster-management.
The “National Point of Contact” (NPOC) of the “Federal Office of Topography”
(SWISSTOPO) as project manager produced rapid mapping products and handed
them out to support the decision-makers. Due to the federal organization of the
Swiss Civil Protection, several different organisations e.g. cantonal crisis
committees, federal offices, the police, fire departments and the army are all
possible end-users of rapid mapping products Federal Office for Civil
Protection [5]. This paper focuses on the optimisation of rapid mapping products
in this particular case.
To acquire useful satellite-data, a very short timeframe is available in which

the affected area is only viewed by a small number of sensors. Additionally, the
Swiss Alpine area presents specific challenges to the use of satellite data. The
small extent of the landscape requires a high spatial resolution of better than
15 meters, which is only available from a few optical sensors (e.g. SPOT,
IKONOS). The large height-differences in the relief cause shadows in optical
images and radar-shadow, layover and foreshortening in radar images. In optical
images, clouds, cloud-shadow and snow complicate the classification of flooded
areas. Due to these difficulties, there are presently only a few datasets available
for a rapid classification of the flooded areas in Switzerland.
Figure 1: Overview of the main affected areas (ovals) in Switzerland and the
study area (dashed rectangle).
2 Remote sensing data

2.1 RADARSAT-1
The “Canada Centre for Remote Sensing” (CCRS) submitted in response to the
Charter invocation a RADARSAT-1 standard mode C-band dataset with a spatial
resolution of 28 meters. A small subimage covering the Lake Brienz region is
shown in Figure 2. The advantage of radar data compared to optical sensors is its
capability to acquire data during the night and through clouds, which is of great
advantage for flood monitoring. Investigations with a pre-flood and a post-flood
RADARSAT-1 image showed that the spatial resolution is not sufficient to detect
small flooded areas of several square meters in the Swiss landscape. The fine
mode with a spatial resolution of 9 meters was not acquired but would have been
a preferable alternative. Further problems such as radar shadow, layover and

foreshortening, increased by the rough topography complicate the classification.
By the time the Charter was activated and the RADARSAT1 image planed and
acquired, much of the flooding had receded, leaving only residual ponds.
Additionally, the archive of images acquired with the same geometry for a
particular region is small and it is difficult to find pre-disaster scenes for change
detection. Further information about geometric and radiometric correction of
radar data in mountainous terrain is provided in Small et al. [6]. Due to
difficulties mentioned above, the RADARSAT-1 data was not used for the
classification of affected areas in this investigation.
Figure 2: Radarsat-1 geocoded terrain corrected standard mode image in full

resolution, acquired on August 28th 2005. The oval highlights a
known flooded area.
2.2 SPOT5
SPOTIMAGE delivered two SPOT5-scenes, acquired on August 26th and 30th 2005.
As base image for change detection, an image from the data archive acquired on
June 12th 2003 was applied. SPOT5 carries two sensors HRG providing two
panchromatic bands (spatial resolution 5 m), two visible bands, a near infrared
band (spatial resolution 10 m) and a short wave infrared band (spatial resolution
20 m). The image swath is 60 km x 60 to 80 km depending on the acquisition
angle Spotimage [7].
The utility of optical images is highly dependent on characteristics like the
revisit time, the repetition rate, the acquisition angle and the scene’s cloud-cover
Campbell [8]. After the heavy rainfall, clouds obscured more than 80% of the
scene acquired on August 26th 2005. By chance, the main parts of the affected
valleys were fortunately cloud-free. The scene from August 30th 2005 has nearly
no cloud-cover but was recorded more than a week after the peak of the flood-
event. Figure 3 shows quick-looks of the two post-flood images.
Figure 3: Quicklooks of the Spot5 HRG images acquired after the flooding.
3 Methodology
3.1 Data preprocessing
Usually the first step in the preprocessing chain for rapid mapping is the
geometric calibration. It is the foundation for change detection methods and
visualisation by overlaying the images with auxiliary geographical data. The
required accuracy lies below the extent of one pixel (SPOT5 10 m, RADARSAT1
28 m). The rough terrain of the Swiss Alps makes accurate geometric calibration
both more important and more difficult. Use of a digital elevation model (DEM)
improves the accuracy remarkably Meier [9], Kellenberger [10].
Correction of atmospheric and illumination effects is not always possible
within the fast preprocessing required in rapid mapping. Due to the
inhomogeneity within a single scene, the lack of atmospheric parameters and
time consuming processing, these corrections were not conducted in this
investigation. To normalize the datasets, the pixel values are transformed from
digital numbers to physical units (at sensor radiance) applying the calibration
coefficients in equation (1) Schowengerdt [11].
Ls = C0 + C1 * DN (1)
Ls = at sensor radiance
C0 = calibration offset or bias
C1 = calibration gain
DN = digital numbers
3.2 Segmentation
Traditionally, classification of remote sensing data has been conducted using a

pixel-based approach, where the spectral response of one pixel and possibly its
very close surroundings is used for classification. Object oriented approaches
first merge adjacent pixels of high homogeneity in spectral reflectance or shape

to image objects, dependent on a chosen scale factor. Classification on the basis
of image-polygons allows use of the spectral statistics in addition to the
integration of intrinsic, topological and contextual information. Furthermore, the
approach improves the possibilities to include knowledge from sources other
than remote sensing data Benz et al. [12]. A disadvantage in the field of rapid
mapping is the large amount of required processing power and memory, slowing
down the segmentation and classification process.
The aim of segmentation is to create image objects that accurately represent
the shapes and sizes of the water-covered areas in the SPOT5 pre-disaster and
post-disaster scenes. Investigations showed that best results are achieved based
on a VISNIR water index (2), the near infrared band NIR and the Swiss digital
elevation model DHM25 on a low scale. According to the normalized difference
vegetation index NDVI Campbell [8], the VISNIR water index is based on the
band containing high spectral response (green) and the band that contains low
spectral response (NIR) for water. Figure 4 shows part of the calculated water
index band from the pre-disaster scene superimposed with the computed
segments.
green − NIR
VISNIR water index = (2)
green + NIR
Figure 4: Computed water index of the pre-disaster scene superimposed with

the generated image-objects.
3.3 Classification
Based on the computed segments, the first classification is performed on the pre-
disaster scene. The aim is to generate a precise representation of the areas
covered by water before the flood. The classification is based mainly on the
VISNIR water index (2), the NIR band and the standard deviation of the height
of the pixels merged in one polygon, which is an estimate of the flatness of a
segment.
The second classification, performed in the post-disaster scene, highlights the
areas that are covered by water now but were not in the pre-disaster scene. The
detection of flooded areas is based on the water index (1), the SWIR band, the
NIR band, the standard deviation of the height of the pixels merged in one
polygon and Boolean combinations. In addition to the flooded areas, clouds,
shadow and snow are separated to avoid intersections with the flood classes.
To save time, the classes are defined by thresholds and not by samples. An
advantage of this method is that later images from the same sensor can be
classified with the same or slightly modified class definitions. The dataset from
August 30th 2005 (Figure 3) was acquired under different atmospheric conditions
and has a larger offnadir view angle, leading to differences in the measured
spectral response. A classification with the same class definitions applied in the
first post-disaster scene shows accurate results for flooded areas. This approach
allows an uncomplicated monitoring of the affected areas, and production of
dynamic flood maps. The same approach could be applied during future flood
events, helping to speed up the production of rapid mapping products. More
research has to be done to check the potential for adapting the established class
definitions to further scenes from the same or different sensors.
To control the achieved accuracy of the object-based classification, infrared
aerial images flown on August 29th 2005 by SWISSTOPO were used. The scanned
aerial infrared images have an average scale of 1:10’000 and have been
resampled to a pixel size of 0.25m to accomplish the accuracy assessment.
Flooded areas of different water levels are perfectly identifiable on these images,
which makes them the best groundtruth available to assess the accuracy of the
classification from the SPOT5 scene. Table 1 shows the results of this evaluation.
Table 1: Accuracy assessment for the two flood classes.
Flooded area low water level Flooded area high water level
Producer’s
0.907 0.929
accuracy
User’s accuracy 0.985 0.995
Kappa Index per
0.864 0.816
class
Overall accuracy 0.921
Kappa overall 0.839
Especially the high Kappa coefficient value of 0.839 verifies the good results
of the classification that a visual evaluation promises. Minor misclassifications
are mainly caused by the time difference between the acquire dates of the
satellite scene and the aerial images. Additionally, flooded areas of a very small
extent or with a very high fraction of debris cannot be detected in the Spot5
image.
4 Results
Rapid mapping products aim to be quickly available and easy to read for end
users. The polygons classified as flooded are combined with topographical maps
to generate a layout that the decision makers are familiar with, and allow them to
add further information. This procedure enables extraction of information about
the type of affected area (e.g. affected traffic routes or flooded settlements).
Figure 5 shows an example of a topographical map with flooded areas
superimposed, as detected within the SPOT5 post-disaster image.
To produce overview maps on a larger scale, the flood-polygons could be
plotted over the original satellite data or over satellite data combined with
topographical maps. These products can be valuable tools when planning rescue
efforts shortly after the disaster.
Figure 5: Rapid mapping product to support the disaster management,

topographical map ©SWISSTOPO (BD053189).
5 Conclusions
The results from this investigation show that object-based classification of
remotely sensed data can produce valuable base data for rapid mapping products,
supporting decision makers in time of crisis. An estimation of the time required
for the production steps after the arrival of the remotely sensed raw data is given
in Table 2. The extracted objects representing the flooded areas can be exported
to a GIS and intersected with land-use or population density maps to estimate
economic damage and other valuable information. These possibilities can

provide an important support for disaster management.
Table 2: Estimation of the time needed for the main steps of rapid mapping
(for a single scene).
Basic work Geometric calibration: ca. 1 h
Segmentation ca. 1 h
Adaptation of the classification for ca. 1 h
the pre-disaster scene
Adaptation of the classification for ca. 1 h
the post-disaster scene:
Validation of the classification ca. 1 h
result:
Map production Map generation: ca. 1/2 h per map
Map layout: ca. 1/2 h per map
Total ca. 6 h
The temporal availability of high spatial resolution satellite data has improved
within the last decade and this trend is still going on. The “International Charter
on Space and Major Disasters” provides the infrastructure to acquire the
appropriate data a short time after a disaster and enables the support with rapid
mapping products. To improve the international success of this institution it is
important that the regions, which are affected by a disaster, invoke the charter
immediately after the event and that more satellite operators join the charter to
raise the probability of image availability.
To enable the detection of small flooded areas in rough terrain, the spatial
resolution ought to be better than 15 meters. The acquisition date should be as
close as possible to the disaster event so that most of the affected areas are still
covered by water.
Common cloud cover in the time after flood-events triggered by heavy
rainfall is the main problem in rapid-detecting affected areas in optical satellite
data. The spectral appearance of urban areas under cloud shadow is very similar
to the appearance of water in SPOT5 data and reduces the accuracy of the
classification. In a further step of research a fast automatic extraction of clouds
and shadow should be developed to improve the speed and quality of the
classification. Application of radar data can bypass the problems of clouds and
cloud shadow, but is limited by layover, foreshortening and radar-shadow in
mountainous areas. Although these limitations exist, there is a need for research
on object-based classification in rapid mapping with combined data from radar
and optical sensors.
Acknowledgements
The authors like to thank swisstopo for data-support and the geometric
calibration, SPOTIMAGE and the Canadian Space Agency for the satellite data.
References
[1] Iglseder, H., Arensfischer, W., Wolfensberger, W., Small Satellite
Constellations for Disaster Detection and Monitoring, Natural Hazards,
volume 15, issue 11, pp. 79-85, 1995
[2] Allenbach, B. et al., Rapid EO Disaster Mapping Service: Added value,
feedback and perspectives after 4 years of Charter actions, SERTIT, 2005
[3] Metternich, G., Hurni, L. & Gogu, R., Remote Sensing of landslides: An
analysis of the potential contribution to geo-spatial systems for hazard
assessment in mountainous environments, Remote Sensing of
Environment, volume 98, issues 2-3, pp. 284-303, 15 October 2005
[4] Frei, C., August-Hochwasser 2005: Analyse der Niederschlagsverteilung,
Bundesamt für Meteorologie und Klimatologie (MeteoSchweiz), 2005
[5] Federal Office for Civil Protection (Bundesamt für Bevölkerungsschutz,)
http://www.bevoelkerungsschutz.admin.ch (access December 12th 2005)
[6] Small, D. Meier, E. Nüesch, D., Robust Radiometric Terrain Correction
for SAR Image Comparison, Geoscience and Remote Sensing Symposium
IGARSS '04. Proceedings, IEEE International, volume 3, pp. 1730 - 1733,
2004
[7] Spotimage, http://www.spotimage.fr (access December 12th 2005)
[8] Campbell, J. B., Introduction to Remote Sensing third edition, Taylor &
Francis London, pp. 272-285, 465-467, 2002
[9] Meier, E., Geometrische Korrektur von Bildern orbitgestützter SAR-
Systeme, Remote Sensing Series volume 15, 1989
[10] Kellenberger, T. W., Erfassung der Waldflächen in der Schweiz mit
multispektralen Satellitenbilddaten, PhD Thesis, RSL Department of
Geography, UZH University of Zurich, pp. 142-164, 1996
[11] Schowengerdt, R. A., Remote Sensing – models and methods for image
processing, Academic Press, San Diego, pp. 311-313, 1997
[12] Benz, U. et al., Multi-resolution, object-oriented fuzzy analysis of remote
sensing data for GIS-ready information, ISPRS Journal of
Photogrammetry & Remote Sensing volume 58, pp. 239 – 258, 2004
Using spatial technology for analyzing

disturbed areas and potential site selection in
Chihuahua, Mexico
V. M. Tena1, A. C. Pinedo2, A. H. Rubio1, P. de L. G. Barragán3,
A. A. Pinedo3, M. V. Hernandez3 & C. Velez2
1
Investigadores del INIFAP, Chihuahua, Chihuahua, México
2
Profesores-Investigadores, UACh, Chihuahua, México
3
Estudiantes de Doctorado, Chihuahua, Chihuahua, México
Abstract
The objectives were to: 1) analyze the physical-chemical variables of a site in
order to characterize the disturbed area’s suitability for planting; 2) detect
disturbed areas through spectral transformation techniques; and 3) generate
cartographic maps with the location of the disturbed areas suitable for planting.
The field data, Landsat TM 7 images, and digital elevation models were analyzed
with IDRISI and ArcView. Multivariate Cluster Analysis, and Principal
Components Analyses were applied to valuate the biometric and physical-
chemical variables of the soil, and detection of disturbed areas. Physical-
chemical analyses showed similar characteristics for the whole area in the study;
only soil depths are considered important for the establishment of forest
plantations. The combination of bands 3, 4 and 5 allowed detection of, in a
preliminary way, the disturbed areas. The Principal Component Analysis showed
that the first component reduced the dimensionality of the data while the second
component detected the disturbed areas.
Keywords: reforestation, disturbed areas, detection, spatial, Chihuahua, México.
1 Introduction
Worldwide, deforestation is associated with increasing demands for forest
products due to increased human populations [1]. Mexico has a deforestation rate
estimated at 600,000 hectares per year [2]. This is one of the highest rates of
doi:10.2495/GEO060401
deforestation in the world with the presence of associated problems like erosion,
desertification, less biodiversity, and lost productivity of the land. This problem
hurts the economy of the forested regions of the State of Chihuahua, where
ecological damage includes irregularities in weather, water, soil, flora and
wildlife among others. In addition, protection measures and commercial tree
planting activities showed a series of limitations and obstacles to success,
because they do not result in clear and well defined objectives due to the lack of
reliable and up to date information related for possible areas to be planted.
To establish successful plantations that provide wood sources for the forest
industry and to diminish pressure in the natural forests, environmental factors
must be considered such as the biophysical and chemical features of the planting
sites. Field data analysis combined with satellite imagery and digital elevation
models (DEM) of high spatial resolution allow the generation of up to date and
reliable information to make decision making easier [3,4]. This helps in making
adequate planning and execution of recovery programs in disturbed areas,
looking to reactivate their productivity as well as conserving and preserving the
associated natural resources.
The objectives of this study were to analyze the physical and chemical
variables of a site to characterize the disturbed areas for suitability for planting,
to detect the disturbed areas through spectral transformation techniques, and to
generate cartographic maps with the location of the suitable areas to be planted.
2 Regional setting
The study was conducted in four areas: (1) San Juanito, (2) Bocoyna, and (3)
San Ignacio de Arareco, all in the Municipality of Bocoyna, and (4) the
commonly owned area called Cusarare in the Municipality of Guachochi,
Chihuahua, located in the Sierra Tarahumara. Geographically, these areas are
located in the UTM coordinates 228369 E, 3046635 N and 265236 E, 3105332
N.
2.1 Climate
The study areas are humid and temperate (C(E)W2) with a mean annual
precipitation of about 600 mm. The mean annual temperature is 12.9 °C with a
minimum of -9.9 °C and a maximum of 26.7 °C. Extreme temperatures varied
from -18 °C in December of 2003 to 40 °C in June of 1992.
2.2 Geology
The area has outcrops of extrusive igneous rocks dominated by the acid types of
rhyolite and tuff of the superior tertiary Ts (Igea). Rough topography of the study
area includes a great diversity of topographic shapes such as plateaus, canyons,
mountains, and hills. Most of the surveyed sites where on plateaus with elevation
ranges from 2,212 meters to 2,572 meters.
2.3 Physiography
Rough topography of the study area includes a great diversity of topographic

shapes such as plateaus, canyons, mountains, and hills. Most of the surveyed
sites where on plateaus with elevation ranges from 2,212 meters to 2,572 meters.
2.4 Soils
The soils of the municipality of Bocoyna are the eutric regosol type and those in
the Cusarare in the Municipality of Guachochi are haplic feozem type. In
general, the soils have shallow depths fluctuating from 35 through 250 cm with
textures that range from sandy clay loamy soils to sandy loams that are medium
high in organic matter content and nitrogen, and deficient in phosphorous, with a
highly acid pH and a hydraulic conductivity from moderate to moderately fast.
2.5 Vegetation
The study areas are characterized by a mixture of tree species that are dominated
by Pinus arizonica, Pinus duranguensis, and Pinus engelmannii and associated
with Quercus spp., Populus spp., Juniperus depeana, Pinus leyophila, Pinus
chihuahuana, Pinus ayacahuite, and Picea chihuahuana.
3 Analysis
Information was analyzed with IDRISI and ARC/VIEW 8.x programs. There
were four data sources: (1) Landsat TM7 January 2003 images, (2) four digital
elevation models (DEM) in 1:50,000 scale, (3) edaphology, topography, and soil
use thematic charts, and (4) field data collected for variables of interest.
A radiometric analysis procedure was applied to the bands used from Landsat
TM7 with geometric correction to fix possible noise effects, pollutants
attenuation, relative humidity, and cloudiness levels [5]. The sub-scene of
interest was adjusted according to the analysis of the study areas for better details
in the analysis of the variable. Quadrangles of the DEM in 1:50,000 scale were
segmented in sub-scenes according to the location of each site. Once the DEM
was generated, two conversion algorithms were applied to generate the Slope
Digital Model (SDM) and the Exposure Digital Model (EDM). Through the
Boolean operator (AND), the slope was reclassified in the next range: 0–2%
commercial plantations, 3–5% forest recovery, and >5% for protection of other
existing resources. For a preliminary detection of disturbed areas, a monoband
analysis of landsat TM was explored, and a band combination was selected
according to the suggested procedures of Pinedo [6] to better discriminate the
disturbed areas. Once the areas were detected, the SEEDTOOLS module from
ERDAS was used, interpolated by the ARCVIEW program to generate in an
automatic way the masks of the areas, the same that were overlain in the sub-
scenes derived from landsat TM as well as from the DEM, SDM and EDM.
Under a stratified random sample, data from the variables of interest were
collected using three sampling techniques: (1) a 1 square meter plot for location
(GPS), elevation (altimeter), exposure (Silva compass), slope (clinometer), and

soil variables, (2) 12.56 square meters (2 meters radius) for the arboreal
vegetation regeneration data, and (3) 200 square meters plots (4 meters radius)
where forest variables included plant species, total and commercial tree heights,
normal diameters, product distributions, ages, annual increments, and year pass).
Statistical analyses were conducted with a multivariate technique of Cluster
Analysis of the MINITAB program. To know the relation of each of the landsat
TM bands and reduce the possible dimensionality of the spectral data, a principal
component analysis was applied beside the newly derived images which allowed
the detection of disturbed areas suitable of being planted [7].
4 Results
4.1 Disturbed areas analysis
The correlation matrix of the principal component analysis (PCA) showed, with
exception of band 4, a high degree of association with the rest of the spectral
bands of Landsat TM (Table 1), in a way that the combination of any of the three
bands can generate a cover type map with an easy detection of the disturbed
areas. Although, band 4 was chosen as the least correlated and combined with an
infrared and one visible band, it was used to generate a composition of false
color using band 3 in the blue channel, band 4 in the green channel, and band 5
in the red channel. This process was useful in this study since it allowed the
generation of actual land use cartography to discriminate in a preliminary way
the disturbed areas, locate the sample sites, and then, quantify the polygons of
areas suitable for planting.
Besides attenuating the effects of the differences of variability between bands
and diminishing the dimensionality of the data between them, the PCA allowed
the generation of new images, making as many components as bands that were
introduced in the analysis.
The first component synthesized 94.37% of the total variability while the
second component explained 3.55% of the total variability. The two components
represented 98% of the total variability of the original variables, while the rest of
the components synthesized 2.08% of the total variability. Even when the first
component reduced the dimensionality of the data preserving almost all the
information, the second component showed clearly the disturbed areas of
detection.
4.2 Cluster analysis
4.2.1 Distribution and grouping of the sites

Variables from the 40 sites were grouped among the four resulting clusters of the
multivariate analysis. (Clusters analysis) based on the similarity in values of the
studied variables, in a way that cluster 1 grouped 6 sites located in the lands of
San Juanito and San Ignacio de Arareco, cluster 2 grouped 9 sites in San Juanito,
San Ignacio de Arareco and Cusarare, cluster 3 grouped 10 sites of the four study
areas while cluster 4 grouped 15 sites from the four areas (Table 2).
Table 1: Matrix of correlation between bands of the sub scenes of landsat tm7
image for the study area.
Scene TM1 TM2 TM3 TM4 TM5 TM7

band_1 1.0000
band_2 0.9814 1.0000
band_3 0.9725 0.9877 1.0000
band_4 0.7520 0.8112 0.7833 1.0000
band_5 0.9162 0.9421 0.9467 0.8408 1.0000
band_7 0.9307 0.9468 0.9591 0.7661 0.9830 1.0000
Table 2: Created clusters based on the interest variables.
Clusters Sites
1 9,10,13,14,15,17
2 1,2,5,8,16,18,36,38,40
3 4,11,12,26,27,30,31,33,37,39
4 3,6,7,19,20,21,22,23,24,25,28,29,32,34,35
4.2.2 Physical and chemical analysis of the sites

The grouped variables in the four clusters showed similar physical and chemical
qualities for the whole area under study, with a relative difference shown by the
variable “soil depth”, which values are higher in the sites grouped in clusters 1
and 2 with regard to the sites grouped in clusters 3 and 4. Because of this, they
showed better conditions for plantation establishment. (Table 3).
4.2.3 Cartography of the planting susceptible areas

The masks of the disturbed areas that originated in component 2 of the PCA
were overlaid for their visualization and analysis in the composite image. This
cartographic product represents the base map that in a thematic and digital
format contains the necessary information to cover the necessities of other
variables that were not analyzed in this study as the estimation of the disturbed
areas, commercial type of affected threes, severity degree of the disturbance, and
ecological impact analysis, among other variables.
The planting areas have a total of 5,820 hectares (Table 4) that are a
consequence of diverse factors that determine the conditions of the existing
forest. In all the studied areas, condition is associated with management history
and includes over-harvesting, illegal harvesting, forest fires, and changes in soil
use.
4.2.4 Surfaces analysis

Besides elevation, the processing and analysis of the DEM gave data for various
slopes, which is important in commercial planting and environmental restoration
programs. Slopes greater than 5% are present in approximately 88% of the areas
with gradients from 62% in at San Juanito to 94% in Cusarare. This
physiographic situation is associated in an inverted way to the class 0-2 % where
San Juanito has the highest percentage of plain or nearly flat land (23%) while
Cusarare has only 1% of this class. Diverse authors [8,9,10] assure that the
selection of the right species, slope, soil, and climate among other factors are
variables that determine the success of plantations since the intended commercial
plantations are meant to be established in preferred nearly flat soils, while the
areas dedicated to protection are meant to be planted on steep slopes.
Table 3: Values of interest variables and their grouping in the four resulting
clusters.
Variable Cluster
1 2 3 4
Elevation 2390.00 2410.28 2456.66 2318.41
Soil Depth 34.5000 33.7143 26.6000 25.0000
(cm)
Mould 0.9167 0.4429 1.0800 0.6667
(cm)
ApDens. 0.6220 0.6624 0.7226 0.8039
E. Poros 63.9307 64.6811 62.0351 60.4064
(%)
C. E. 0.2200 0.1743 0.1867 0.2000
(mmhos Normal Normal Normal Normal
/cm)
pH 4.18 4.32 4.36 4.20
Highly acid Highly acid Highly acid Highly acid
Org. 1.7183 1.6671 1.5407 1.1325

Mat.(%) M. high M. high M. high M. high
RAS 9.51 Normal 7.27 Normal 7.58 Normal 7.27 Normal

N-NO 404.69 469.29 235.75 230.62
(Kgm/ha) M. high M. high M. high M. high
P 12.78 14.88 11.84 9.43
(Kgm/ha) Deficient Deficient Deficient Deficient
K (ppm) 770.83 278.57 High 210.83 High 275.00 High
Excessive
C. H. 6.473 MF 6.68 MF 4.44 M 8.62 MF
(cm/hr)
Sand (%) 52.7280 45.1337 57.6853 60.9947
Clay (%) 23.1520 25.0720 20.0800 20.2787
Bare soil 39.0667 49.5571 27.3333 41.7250
(%)
Table 4: Planting susceptible surfaces based on the spectral analysis.
Property Area (ha) Planting area (ha)

San Juanito 5229 1877.83
Bocoyna 6055 956.46
San Ignacio de Arareco 21051 1370.04
Cusárare 29914 1616.09
TOTAL 62249 5820.42
5 Discussion
The creation of preliminary maps was very useful because it allowed the
generation of actual use of soil cartography to discriminate the disturbed areas,
locate the sampling sites, and then quantify the polygons of the areas suitable for
planting. The careful selection of these three bands was made according to the
suggestions of Richards [11] and Beaubien [12] who recommended the reduction
in the number of bands and use of those whose combination maintained the level
of discrimination for the types of cover of interest.
From a statistical point of view, the PCA easily made a first interpretation
over the variability axles of the image and allowed identification of those
features that showed up in almost all of the bands and those others that are
specific to some group of bands [7,13,14]. Although Muchoney and Haack [15]
referred to the temporary changes and disturbance in vegetation that are observed
in the last components of Landsat TM, in this study the component 2 allowed
discrimination in a clear way for the disturbed areas which indicates the
importance of using this multivariate tool to ease the processes for this type of
study.
Variables of interest among the 40 sites were grouped in the four resulting
clusters from the multivariate analysis (Cluster analysis) based on the similarity
of the values of the studied variables that allowed the distribution and grouping
of the sites. Johnson [16] noted that this multivariate technique involves
techniques that produce classifications from data that initially are not classified
and must not be confused with the discriminating analysis in which from the
beginning it is known how many groups exist, and it has data that comes from
each one of these groups.
The site analysis showed similar physical and chemical features for the whole
study area, with a relative difference expressed by the variable soil depth whose
values are greater in the grouped sites in clusters 1 and 2 (34.5 and 33.7 cm.), in
respect to sites grouped in clusters 3 and 4 (26.0 and 25.0 cm.), for which they
showed the best conditions for plantation establishment.
This matches another study [17] where soil depth was one of the edaphic
variables of greatest impact in determining the quality of sites for plantations,
since with greater depth the water retention increases in the soil and makes it
easier to incorporate organic matter from the decomposition of logs, branches,
and twigs among others. This permits an adequate radicular development for the
establishment of crops [18]. In respect to this, Narváez and Armendáriz [19]
noted that verticiles in plantations of Pinus arizonica, Pinus duranguensis, and

Pinus engelmanii depend on the total depth of the soil.
The planting area on the assessed properties, which together make a total of
5,820 hectares, is a consequence of diverse factors that determine the actual
forest conditions, which is associated with the management history of the
properties related to factors like over-harvesting, illegal harvesting, forest fires,
and changes in land use.
Besides elevation, other processes and analysis gave important features of the
terrain, mainly in the derivation of slope, which is an important variable in
commercial plantations and environmental restoration programs. Diverse authors
[8,9,10] pointed out that the selection of adequate species, slope, soil, and
climate among other factors, are variables that determine the success of
plantations.
References
[1] Rotmans, J. and R. J. Swart. 1991. Modeling tropical deforestation and its
consequences for global climate. Ecol. Modeling. 58: 217-247.
[2] CONAFOR, 2005. Comisión Nacional Forestal. Inventario Nacional
Forestal. Versión Preliminar s/p.
[3] Franklin, J. 1994. Thematic Mapper Analysis of Coniferous Forest
Structure and Composition. In. J. Remote Sensing of Environment. p. 35.
[4] Francois, 1999. Aplicación de Imágenes de Satélite para Análisis de
Recursos Forestales Deforestación y Fragmentación Forestal en la región
de la Laguna de Términos, Campeche: un análisis del período 1974-
91. Centro EPOMEX Universidad Autónoma de Campeche. Campeche,
Camp. México. p 12.
[5] Lillesand, T. M. y R. W. Kiefer. 1987. Remote Sensing and Image
Interpretation. John Wiley and Sons. New York.
[6] Pinedo, A. A. 2004. Análisis multitemporal de áreas deforestadas en la
región centro-norte de la sierra occidental, Chihuahua, México. Tesis de
Maestría. Facultad de Facultad de Zootecnia. UACH. p.53.
[7] Ferrero, B.S., M.G. Palacio, y R.O. Campanella, 2004. Análisis de
Componentes Principales en Teledetección. Consideraciones estadísticas
para optimizar su interpretación (Parte II). Universidad Nacional de Río
Cuarto, Córdoba, Argentina. p. 9.
[8] Campos, R. D., Caráves, S. I. y Varela O. R. 1998. Normas para el
Trabajo Técnico No 53. p 54.
[9] Moreno, S. R., Moreno, S. F. y Cruz, B.G. 1994. Determinación de áreas
potenciales para plantaciones forestales. IV Reunión Nacional de
Plantaciones Forestales. SARH. pp. 180 – 186.
[10] Escárpita, H. A. 2002. Situación actual de los bosques en Chihuahua.
Madera y Bosques 8(1), 2002:3-18.
[11] Richard, J. A. 1986. Remote Sensing Digital Image Analysis. And
introduction, Ed. Springer-verlang.
[12] Beaubien, J. 1994. Landsat TM Satellite Images of Forests: From

Enhancement to Classification. Canadian Journal of Remote Sensing. Vol.
11. No Québec, Canadá. p. 17-26. Bosques 8(1), 2002:3-18.
[13] Robin, M. 1995. La télédéction. Nathan Universite. Des satellites aux
systémés d´ information géographiques. Edition Nathan.
[14] SPIPR2. 1992. Sistema Personal Interactivo de Percepción Remota.
Versión 2.0. Aguascalientes, Ags. 1992.
[15] Muchoney, D.M. and Haack, B.N. (1994). Change Detection for
Monitoring Forest Defoliation. Photogrammetric Engineering & Remote
Sensing. Vol. 60, No. 10. pp 1243-1251.
[16] Johnson, E. D. 1998. Métodos Multivariados Aplicados al Análisis de
Datos.Kansas State University. p. 319.
[17] Carmean, W. H. 1975. Forest site cuality evaluation in the United States.
Advances in Agronomy. 27:209-269. USA.
[18] Davel, M. y Ortega, A, M. 2002. Predicting site index from environmental
variables in Douglas-fir plantations in the Patagonian Andes, Argentina.
En línea Disponible: mdavel@ciefap.cyt.edu.ar. Accesado: 6 de
Noviembre del 2005.
[19] Narváez, F. R. y Armendáriz ,O. R. (2005). El suelo y clima en relación
con la calidad de sitio de las plantaciones forestales en los municipios de
Bocoyna, Guerrero y Madera del estado de Chihuahua. En Memoria del
VII Congreso Mexicano de Recursos Forestales. Chihuahua, México. pp
94-95.
Section 11
Soil and rock properties
The pedoecologial conditions of natural and

opencast peat fields in Estonia
M. Noormets1, T. Köster1, T. Tõnutare1, K. Kauer1, R. Kõlli1,
T. Paal2 & M. Oder1
1
Department of Soil Science and Agrochemistry,
Estonian University of Life Sciences, Estonia
2
Forest Research Institute, Estonian University of Life Sciences, Estonia
Abstract
Peat soils in five different bog areas were studied during 2004. The study areas
were chosen according to their exploitation type; natural, cultivated and milled
peatlands. The peat soil nutrient content, plant associations and their nutrient
content were examined. The peat soils were analysed for the CHWE, total N, P, K,
Ca, Mg, C, for ash content and for plant-available P, K, Ca, and Mg. For the
distribution of nutrients in the peat soils, samples were taken from 0-5, 5-10 and
10-15 cm depths. The dominating plant species and the type of plant associations
were determined in areas where it was presented. The plant samples were
analysed for N, P, K, Ca and Mg. The study results showed that in natural peat
areas the variability in the ash content was high; the same was found for the Ca
and Mg (%) content. With increasing depth, the nutrient content decreased, but
this depth relationship was not significant for every parameter examined. The
plant cover in cultivated peat areas had the highest nutrient contents.
Key words: Fibri Dystric Histosol, natural peat bog, milled peatland, cultivated
peat soils, plant association, total nutrients, available nutrients, pH, C, N, C:N
ratio, CHWE, ash, P, K, Ca, Mg.
1 Introduction
Nowadays we consider the peatlands as a valuable ecological biotope that should
be protected and rationally used because it is essentially a nonrenewable
resource. Currently the peatlands are used for growing wild berry species like:
Rubus chamaemorus, Oxycoccus palustris and Vaccinium species [1, 2, 3]; they
doi:10.2495/GEO060411
are used for forestry [4, 5], and as a growing media source for horticulture [6].
Historically and presently, peatlands have been heavily utilized and degraded by
their agricultural and mining use. After drainage, the self-regulation system of
mires is changed and the natural peat formation process is destroyed and then the
dominating process is peat mineralization [7, 8]. Due to various agricultural
activities in peat lands, the peat loss as a result of mineralization could reach 10
to 15 tons of organic material per hectare annually [9]. One possible strategy to
stop or slow the mineralization is to cultivate wild berries (Vaccinium species).
However, even with this strategy, mineralization and surface deflation are fast
compared with peat formation. The preservation of peatland resources is critical
because they are reservoirs of clean water, important sinks for carbon
sequestration, important for maintenance of the hydrologic regime, and lastly
they are valuable biotopes.
The Northern mires represent the largest store of carbon in the terrestrial
ecosystem. Therefore, these areas play an important role in the Earth’s gas
emission budget that affects the atmospheric CO2 concentration. According to
Kreshtapova and Maslov [10] the CO2 storage in boreal peat deposits is
estimated at 455 Pg. Moreover, Horn et al. [11] report that 30% of global
reserves of soil carbon are preserved on natural bog areas. These areas can
contribute up to 7% of the global annual emission of the greenhouse gas
methane. If the natural condition of mires is destroyed by drainage, peat harvest,
or agricultural use, they will change from being a carbon sink, to an important
CO2 emission source.
Currently the annual peat production of different types of peat is
approximately 4.5 – 5 million m3 in Estonia. Of the produced peat, 2.5 – 3
million m3 is comprised of the weakly decomposed peat used as a growing media
in various horticultural production operations. The peat for growth substrate is
produced mechanically by tractor drawn vacuum machines or by block cut
harvesting. Most of the well decomposed peat (1.5 – 1.8 million m3) is used for
energy consumption. Exports from milled peatlands are in excess of 90% for
plant growing media, 80% for briquettes, and approximately 60% for block cut
peat [12].
In this article will identify the peat soil properties as found under ‘natural’
conditions, milled peatland and under the cultivation in order to find out their
characteristics and suitability for natural revegetation. An additional purpose of
this study was to compare the natural, milled and cultivated peat soils
agrochemical properties.
2 Study area
The study areas are located in the County of the Tartu, in the southern part of
Estonia. According to the WRB soil classification system, the soil of the
experimental fields belongs subgroups of Fibri–Dystric Histosols - HSdy(fi). The
plant associations of the study areas were determined according to the
classification system developed by J. Paal [13]. The study areas Sangla I and
Sangla III are used currently for peat mining, therefore no plant cover is
presented.
Table 1: The type of study area use and plant associations or dominating
plant species.
Number of Name of Type of use Plant association or dominating

study area study site plant species
1 Sangla I Milled peat land Without plant cover.
2 Sangla II Technogenic No clear plant association is

stripe presented
3 Sangla III Milled peat land Without plant cover.
4 Ilmatsalu I Natural peat bog Ledo-Pinetum, Ledum pine heath

moor
5 Ilmatsalu II Cultivated peat V. angustifolium, V.
bog area. angustifolium X V. corymbosum
Sangla I. The peat processing started in this particular area in 1962 and currently
the mining area is increased up to 590 ha. The Sangla mire peat deposit thickness
is 6.5-9.5 m [5, 14]. Before the natural plant cover was removed for mining in
Sangla I, the area was characterized as the heath moor growth site type. The
main processing commodities produced from this site are horticultural peat and
briquettes. For the package peat with the following criteria is used: 100% of
poorly decomposed Sphagnum sp. peat (white peat); the moisture content is
between 35% and 55%; the peat is free of weed seeds and radioactive or
chemical substances. The Von Post decomposition index is H1…H3, the ash
content does not exceed 5%; the bulk density is 130 - 200 kg/m3; and the pH is
3.0 - 4.0.
Sangla II is a small area surrounding the main peat milling ground where the
natural revegetation is occurring and for the purposes of this paper, it is
described as a technogenic stripe. The whole area is influenced by drainage and
in the past during the planning of the milling ground the plant cover was
removed and pushed by machines to the sides. Nowadays there is a succession
plant cover that could potentially function as the donor area for wild plant
species during the rehabilitation work. The dominating plant species are Pinus
sylvestris, Betula sp., and Eriophorum vaginatum. In the grass canopy, the
following were identified; Carex cespitosa, C. acuta. C. elata, C. nigra, C.
lasiocarpa, Phragmites australis, Phalaris arundinacea, Filipendula ulmaria.
Sangla III. The mining area is located in the Sangla mire (same as the Sangla I
peat deposit) and in the Laugesoo mire. The main commodities are fuel peat,
peat briquettes and peat extraction.
Ilmatsalu I is part of Sangla mire, named for the region where it is situated. This
peat deposit is 6.5 – 9.5 m thick (same as Ilmatsalu II) and the bog is classified
as heath moor. The somewhat more important plant association of this bog is
Ledo-Pinetum. The main tree species is Pinus sylvestris with accompanying

species of Ledum palustre, Calluna vulgaris, Andromeda polifolia, Empetrum
nigrum, Oxycoccus palustri, Vaccinium uliginosum and Vaccinium myrtillus.
The following bryophytes were represented Polytrichum commune, P. strictum,
Sphagnum sp. In the grass canopy Rubus chamaemorus and Carex sp is
presented.
Ilmatsalu II. In, 1994 the peat milling on current area was finished. Before the
peat excavation the growth type was heath moors, and important plant
association was Ledo-Pinetum. Currently in opencast peat field have the
dominant species Vaccinium angustifolium and in addition V. angustifolium X V.
corymbosum cultivar Northblue was planted. Randomly among cultivated
species appeared Pinus sp., Betula sp. and Eriophorum vaginatum. In the
bryophyte canopy Marchantia polymorpha, Ceratodon purpureus, Polytrichum
strictu, Funaria hygrometrica are randomly scattered. Drosera rotundifolia was
randomly found. In early spring the plantation received 50 kg ha-1 of (11:11:22)
N:P2O5:K20 fertilizers.

For the current investigation the soil sampling was carried out during the
vegetation period in 2004. The plant association description was made during the
summers of 2004 – 2005. Soils were described on the basis of test pits and the
samples for analysis were taken from different depths (0-5, 5-10 and 10-15 cm).
The peat soil samples were taken in four replicates and analysed in laboratory in
triplicate. Peat samples were analysed for organic C according to the Tjurin
method [15] and total N according to Kjeldahl [16]. For the direct estimation of
the organic matter content the loss-on-ignition (LOI) method was used.
Available P, K, Ca and Mg were analysed according to the Mehlich-3 method
[18]. The pH was measured from the soil suspension with 1M KCl (1:5 w/v). For
the total phosphorus, potassium, sodium, calcium and magnesium in peat and
plant determination samples were destroyed by wet digestion with sulphuric
acid. Total phosphorus was determined spectrophotometrically by
vanadomolybdophosphoric acid method. K, Ca and Mg were determined with
atomic absorption spectrometer [17]. Hot water extractable organic C from the
soil-water extract (1:100, w/v) was measured [16]. STATISTICA 7 [19] was
used for the statistical analysis and the standard error (± SE) is presented on the
figures. In the tables the standard deviation (± SD) is presented.

The study area peat pH values varied from 2.4 to 5.2. In Sangla III the pH was
relatively high (pH 5.2) it might be hypothesized that the higher values for pH
and ash are indicative of the increased importance of fen deposit. However, the
nutrient content is low at this location. Spiers [20] is found that in low acidic
soils at pH 3.5, the plant growth and yield is decreased. In the cultivated area
Ilmatsalu II despite to fertilizers application the soil nutrient content for total and
available nutrients was low when compared to other areas (Table 2).
Table 2: Peat soils agrochemical properties in the study areas (n=12).
Soil Number of study area

characteristic 1 2 3 4 5
pHKCl 4.0 ± 0.7 4.0 ± 0.7 5.2 ±0.1 2.4± 0.1 3.1± 0.9
Ctot. % 38.9± 2.4 38.3± 3.3 38.3± 3.1 38.4± 4.4 41.0± 2.9
Ntot. % 1.3± 0. 3 1.6± 0.2 2.1± 0.1 1.0± 0.3 1.0± 0.1
C:N 31.5± 10.2 24.4 ± 2.6 18.3±2.3 40.7±12.9 42.3±7.6
CWE. % 0.62± 0.16 1.75± 0.44 0.5± 0.1 1.1± 0.4 1.0± 0.3
Ash % 3.3± 1.6 27.1± 18.6 10.7± 4.7 3.3± 2.9 1.3± 0.6
Ptot. % 0.03± 0.01 0.05± 0.01 0.03± 0.02 0.03± 0.02 0.02± 0.01
Ktot. % 0.19± 0.05 0.64± 0.24 0.14± 0.13 0.29± 0.25 0.11± 0.08
Catot. % 0.7± 0.2 0.7± 0.2 1.2± 0.4 0.2± 0.1 0.4 ± 0.3
Mgtot. % 0.18± 0.05 0.26± 0.10 0.23± 0.03 0.06± 0.01 0.12± 0.08
Pavb. mg kg-1 23.5± 10.4 85.6± 36.9 8.9 ± 4.8 88.3 ± 85.8 34.8± 31.7
Kavb. mg kg-1 224.8± 86.4 207.2± 56.9 120.2± 19.7 297.4±129.5 190.8± 34.5
Caavb. mg kg-1 9258± 1375 9070± 2410 12889± 2319 3530± 959 5551± 1988
Mgavb. mg kg-1 1598± 328 1840± 381 1733± 217 766± 248 1253± 567
*tot- total content of nutrients, *avb- available content of nutrients.

*The numbers in tables are presented as average of different depths (upper 15 cm).
According to the soil nutrient assessment valuation done at Estonian

Agricultural Research Centre the content of plant available P is in Sangla I and
Ilmatsalu II low (23.5 and 34.8 mg kg-1, respectively), in Sangla III very low (8.9
mg kg-1). Unexpectedly, the content of available Ca and Mg was very high in all
study areas peat soils.
The plant nutrient content was higher in the cultivated area than on natural
area grown plants (Figure 1). Although, the agrochemical parameters of the
cultivated area were in the same range, or slightly lower, the plant nutrient
content was higher. It is reported that the availability of P could be enhanced
while the yield is decreased [21]. But Holmes [22] has indicated that this
situation could be improved by additional phosphorous application. Fertilization
of plants grown in peat soils is important [23, 24, 25] because of the low mineral
nutrient availability in soil. From Noormets et al. [26] in a study using
Vaccinium species have found that plant growth and yield formation on opencast
peat field (Fibri Dystric Histosol) was influenced by the fertilisation rates and
nutrient balance at pH level 2.8-4.0.
Significantly different content for total N and Ca, also available Mg was
found in milled peatland when compared to the milled and cultivated peatlands
(Figure 1). The natural peat soils were found to have higher levels of total K and
plant-available P and K when compared to mined locations. The content of ash
depended upon the decomposition level of peat material itself, and it varied from
3.3% to 27.1% in the areas examined. The higher content of ash in technogenic
stripe might be explained by the mineral material that is carried from the road
nearby. Some studies have found that the liming of peat soils could lead to the
pH increase but in natural areas, like an exhausted peat soil, it could increase the
possibility for the wild plant species expansion from the surrounding areas [27].
The revegetation of milled peat lands areas in Estonia with trees and bushes
occurs rapidly, but the restoration of natural Sphagnum mosses would seem to be
very complicated. According to Salonen [28] the plant succession in opencast
peat fields differs from the natural sites. As demonstrated in trials in the Cacouna
Station (Canada), the revegetation in some plots by trees and bushes occurred
very quickly. The scarcity of Sphagnum and other mosses indicates that the bog
is not returning to the functional peatland ecosystem [29]. The revegetation
depends on the peat harvesting technology, where the block-cut mined bogs were
rapidly recolonized (less than 5 years) by ericaceous shrubs, but the vacuum-
mined peatlands were covered by plant cover much more slowly (approximately
25 years) and not by typical peatland species. The study has pointed to the two
main problems occurring by revegetation: firstly, large areas are without any
plant cover; and secondly, the moisture deficiency in the top peat layers. Both of
these conditions contribute to the wind erosion of the peatland [29].
2,0 2000
N, % P, mg kg-1
1,8 P, % 1800
K, mg kg -1
1,6 K, % 1600
Mg, % Mg, mg kg-1
1,4 1400
Ca, %
1,2 1200
1,0 1000
0,8 800
0,6 600
0,4 400
0,2 200
0,0 0
-0,2 -200
milled peatland natural cultivated milled peatland natural cultivated
Type of use Type of use
A. B.
3,0 45
N, % C, %
P, % 40 Ash, %
2,5
K, % 35
2,0 Mg, %
Ca, % 30
1,5 25
1,0 20
15
0,5
10
0,0
5
-0,5
natural cultivated 0
milled peatland natural cultivated
Type of use Type of use
C. D.
Figure 1: The nutrient content of different investigation areas according to

the type of exploitation (M±SE). A- content of total nutrients. B-
content of available nutrients. C- content of total nutrients in plants.
D- content of total carbon and ash.
2,0 2500
N, %
P, mg kg -1
P, %
K, mg kg-1
K, % 2000
1,5 Mg, % Mg, mg kg-1
Ca, %
1500
1,0
1000
0,5
500
0,0
0
5 10 15 5 10 15
Depth, cm Depth, cm
A. B.
50
C, %
45 Ash, %
40
35
30
Values
25
20
15
10
5
0
5 10 15
Depth, cm
C.
Figure 2: The nutrient content in different depths (M±SE). A- the content of

total nutrients. B- the content of available nutrients. C- content of
total carbon and ash.
The mean content of some nutrients (Ntot, Catot, Mgavb) decreased with
increasing depth, but as the variability was high, this decrease was not significant
(Figure 2). However, the decrease of some nutrient by depth like Ktot and Mgtot
was observed. Also, there was found to be a significant decrease in the content of
Pavb.
5 Conclusions
Currently, in Estonia there is a tendency to expand the working milling areas of
peatlands and then the rehabilitation work is completed afterward. This study has
demonstrated that the content of available P was highly variable, ranging from
10-88.3 mg kg-1, and the variability was highest in natural bog areas. In the
technogenic stripe areas, the ash content (27.1%) was significantly higher than in
other study areas where it ranged 1.3-3.3%. The content of available Ca and Mg
was also very high. The nutrient content in the plant cover was higher than in
natural area, likely due to fertilization. The studied depths for nutrient content in
the peat deposit showed significance difference for total K and available P, for
the rest of parameters the variability found to be very high.
Acknowledgements
We thank colleagues from the Department of Soil Science and Agrochemistry
EMU for their collaboration with the laboratory analyses, especially Mrs. Imbi
Albre. We thank the farmers and companies for their kind permission to use their
property for the research. This study was supported by the Estonian Ministry of
Education and Research, No 0172613s03. The conference participation fee was
supported by RAK and by the Estonian Science Foundation (Project 4726).
References
[1] Noormets, M., Köster, T., Karp, K., Paal, T. & Tõnutare, T., The
recultivation of opencast peatland in Estonia. Proc. of the 12th
International Peat Congress. Wise Use of Peatland. ed. J. Päivänen,
Finland, pp. 1195-1201, 2004.
[2] Jaadla, T., Cranberry production on peatlands. Estonian Peat, 1, pp. 19,
1994.
[3] Paal, T., On cultivating European cranberry in Estonia. Journal of Small
Fruit and Viticulture. 1(1), pp. 59 - 61, 1992.
[4] Hartman, M., Karisto, M. & S. Kaunisto. Carbon dynamics for a drained
peatland forest with drainage and nutrition gradients. Ecohydrological
Processes in Northern Wetlands. pp. 209-214, 2003.
[5] Valk, U., Estonian mires. Valgus, Tallinn-Estonia, pp. 270, 1988.
[6] Carlile, W.R., Horticultural developments, environmental challenges and
growing media. A global perspective. Proc. Int. Conf. Peat in
Horticulture, Quality and Environmental Callenges, eds. G. Shmilewski.
& L. Rochefort, Estonia, pp. 17-23, 2002.
[7] Ilomets, M., The peat mining on account of what? Estonian Nature, 2/3,
pp. 20-24, 2003.
[8] Armentano, T.V. & Menges, E.S., Patterns of change in the carbon
balance of organic soil wetlands of the temperate zone. Journal of
Ecology, 74, pp. 755-774, 1986.
[9] Tomberg, U., Breakdown of peat as a result to drainage, Saku, Estonia,
pp. 32, 1992.
[10] Kreshtapova, V., N. & Maslov, B. S., Contents of carbon compounds in
reclaimed peat soils as a function of the properties of peat organic matter.
Proc. of the 12th Int. Peat Congr. Wise Use of Peatland. ed. J. Päivänen.
Finland, 2, pp. 988-992, 2004.
[11] Horn, M.A., Matthies, C., Küsel, K., Schramm, A. & Drake, H.L.
Hydrogenotrophic methanogenesis by moderately acid-tolerant
methanogens of a methane-emitting acidic peat. Applied and
Environmental Microbiology, 69, pp. 74-83, 2003.
[12] Estonian Competition Board, 2002. http://www.konkurentsiamet.ee/
dokumendid/ko200272.htm.
[13] Paal. J., Classification of Estonian vegetation site types. Tartu-Tallinn. pp.
297, 1997.
[14] Orru. M., Širkova. M. & Veldre. M., Estonian peat resources. Estonian
Geological Survey. Tallinn, pp. 146, 1992.
[15] Vorobjova, L.A., Chemical analysis of soils. Textbook, Moscow
University Press, pp. 272, 1998.
[16] Procedures for soil analysis, ed. L.P. van Reeuwijk Wagening: ISRIC, pp.
112, 1995.
[17] Methods of soil analysis. Part 3. SSSA Book Series: 5. Madison.
Wisconsin. USA. pp. 1390, 1996.
[18] Handbook on Reference Methods for Soil Analysis. Soil and Plant
Analysis Council. Inc. pp. 202, 1992.
[19] Statsoft. Statistica 7,0. Copyright 1984-2005.
[20] Spiers, J.M., Influence of lime and sulfur soil additions on growth, yield
and leaf nutrient content of rabbiteye blueberry. Journal of the American
Society of Horticultural Science, 109, pp. 559-562, 1984.
[21] Bishko, A.J. & Fisher, P.R., The pH-Response of a Peat-based Medium to
Application of Acid-reaction Chemicals. Horticultural Science, 38(1), pp.
26-31, 2003.
[22] Holmes, R. S., Effect of phosphorous and ph on iron chlorosis of the
blueberry in water culture. Soil Science, 90, pp. 374-379, 1960.
[23] Noormets. M. & Karp. K., The influence of fertilization to vegetative
growth of the lowbush blueberry (Vaccinium angustifolium Ait.) in a
young and cropping plantation in peat bog. Transactions of the Estonian
Agricultural University, 212, pp. 149 – 154, 2001.
[24] Noormets, M., Karp, K., Starast, M. & Paal, T., The influence of
fertilization on the production of lowbush blueberry (Vaccinium
angustifolium Ait.) seedlings on opencast peat pits. Journal of
Agricultural Sciences, 5(8), pp. 293-303, 2002.
[25] Noormets. M., Karp. K. & Paal. T., Recultivation of opencast peat pits
with Vaccinium culture in Estonia. Ecosystems and Sustainable
Development IV. eds. E. Tiezzi. & C.A. Brebbia. Wessex Institute of
Technology. UK and J-L. USO. Universitat Jaume I. Spain. 2, pp. 584,
2003.
[26] Noormets, M., Karp, K., Kelt, K.,Tõnutare, T. & Paal, T., Fertilizers
effects on the lowbush blueberry (Vaccinium angustifolium Ait.) plants
and berry chemical composition, grown in opencast peat pits. Journal of
Agricultural and Food Science, (submitted), 2006.
[27] Paal, T., Starast, M. & Karp, K., The influence of liming on growth of
lowbush blueberry on exhausted peat fields. Uprawne rosliny wrzosowate,
ed. T. Ligocka, Skiernewice, Poland, pp. 71-76, 2003.
[28] Salonen. V., Penttinen. A. & Särkkä. A., Plant colonization of a bare peat
surface: population changes and spatial patterns. Journal of Vegetation
Science 3, pp. 113-118, 1992.
[29] Lavoie. C. & Rochefort. L., The natural revegetation of a harvested
peatland in southern Quebec: A spatial and dendroecological analysis.
Ecoscience 3(1), pp. 101-111, 1996.
[30] Naucke, W., Heathwaite, A.L., Eggelsmann, R. & Schuch, M., Mire
chemistry. Mires. Process, exploitation and conservation. ed. Heathwaite,
A.L. pp. 263-310, 1993.
Monitoring programme for underground rock

characterization facility
K. Lehto & J. Lahdenperä
Posiva Oy, Finland
Abstract
Posiva is a Finnish nuclear waste management company with the duty to take
care of the high-level nuclear waste produced in Finland. Pursuant to the
Decision-in-Principle of 2001 Posiva has moved to the phase of underground
characterisation of the repository site. The main objective of this phase is to
confirm the suitability of the Olkiluoto site by investigations conducted
underground. A programme of monitoring has been launched as part of the
investigations and to follow the changes occurring within the site due to the
construction. The programme of monitoring started in the year 2004, one year
before the construction work. Baseline monitoring was carried out prior to the
construction phase and the results gained from monitoring during construction
phase are compared to the baseline data. Rock mechanics, hydrology,
geochemistry, environment, and the use of foreign materials are included the
monitoring programme.
Keywords: monitoring, high-level nuclear waste, hydrology, rock mechanics,
geochemistry, environment, foreign materials, underground rock
characterization facility.
1 Introduction
Posiva Oy is a Finnish high-level nuclear waste management company owned by
the two Finnish nuclear companies Teollisuuden Voima Oy and Fortum Power &
Heat Oyj. Posiva’s duty is to take care of final disposal of spend nuclear fuel.
Posiva is established in the year 1996.
In the year 2001, the Finnish Parliament ratified the Decision-in-Principle on
the disposal of spent fuel from the Finnish nuclear power reactors. According to
the decision, Posiva may concentrate the investigation on one site, Olkiluoto, in
doi:10.2495/GEO060421
the municipality of Eurajoki in the western part of Finland. The Decision-in-

Principle also means that the repository would be located at Olkiluoto and the
disposal would be based on the KBS-3 concept.
On the year 2004 Posiva started to construct an underground characterization
facility called ONKALO, which will be a part of actual high-level waste
repository later on. The repository is planned to be in use in the year 2020.
Before Posiva started to construct ONKALO a baseline monitoring was done.
The monitoring results gained during the construction and use of ONKALO will
be compared to the data gained during the baseline monitoring.
ACCESS
TUNNEL
VENTILATION
SHAFT
MAIN
CHARACTERISATION
LEVEL
LOWER
CHARACTERISATION
LEVEL
Figure 1: The design of the underground rock characterisation facility at the

main drawings stage [6].
2 ONKALO: underground rock characterization facility

The construction of the underground rock characterization facility called
ONKALO was started in July 2004. By the end of the year 2005 about 990 m of
tunnel has been excavated and the first underground investigations program is
implemented. The target depth of the ONKALO is 520 m below the sea level.
The main characterization level is at the depth of 420 m below the sea level and
the lower characterisation level in depth of 520 m below the sea level, fig. 1. The
total length of the tunnels will be about 9 km. The entire facility is planed to be
ready in the year 2010. ONKALO is planned to be a part of the actual high-level
waste repository, which is planned to start its operation in the year 2020 [6].
The construction of the ONKALO and the high-level waste repository will
affect the surrounding rock mass and the groundwater flow system as well as the
chemical environment on the surface but especially at the greater depths. To
determine the magnitude and the extent of effects the programme of monitoring
was established. The programme of monitoring at Olkiluoto during construction
and operation of underground rock facility was established prior to the
construction phase [7].
3 Programme of monitoring
Before the programme of monitoring during construction of ONKALO begun a
baseline condition of Olkiluoto was reported [5]. The main purpose of the
baseline report was to establish a reference point for the following phases of the
Finnish spent nuclear fuel disposal programme. The focus of the baseline report
was to define the current surface and the underground conditions at the site, to
establish the natural fluctuation of properties that are potentially disturbed and to
provide the reference data to the models and modellers.
During the construction of ONKALO rock mechanical, hydrological,
chemical and environmental monitoring is carried out and the quality and the
quantity of the foreign materials used in the construction and investigations is
measured and followed [7]. The monitoring is carried out both on the surface and
underground. In the Baseline report the current understanding of the
investigation site before the constructions of ONKALO started was established.
The results gained from monitoring have been and will be compared to the
baseline data.
3.1 Hydrology
The following parameters have been included in the hydrological monitoring
programme:
o level of groundwater and seawater table
o hydraulic head
o flow conditions
o in situ salinity
o precipitation
o surface runoff
o infiltration
o soil frost
o water balance in the tunnels
The hydrological monitoring is carried out from the both on surface and
underground. Both the open and the packed-off boreholes are used in the surface
based monitoring. The flow conditions, in situ salinities and the hydraulic heads
are measured from the open boreholes by Posiva Flow Log measurement device
and Posiva’s hydraulic testing unit (HTU).
Figure 2: Map of deep drillholes (KR) and multilevel piezometers (EP) in

Olkiluoto.
The hydraulic heads and the level of the ground water table are measured
from the open and packed off boreholes as well as from the multilevel
piezometers, fig 2. The level of water table is measured manually from the open
boreholes once a week or month and the hydraulic head automatically from the
packed-off boreholes once every ten minute.
The rest of the parameters listed above form a background data for the
studies and the analysis of the actual monitoring results [1].
3.2 Rock mechanics
A local GPS-network was established at Olkiluoto in the year 1995. It consists of

14 stations one of which belongs to the Finnish permanent GPS network
FinnRef. Four of the stations are located outside the island. The purpose of the
measurements is to observe long-term large-scale rock movements and land
uplift.
Because the GPS-measurements are not sensitive enough, precise levelling is
carried out to measure the annual land uplift at the Olkiluoto.
A microseismic network of Olkiluoto was built in 2002. The network consists
of 12 stations. Four of the stations are located outside Olkiluoto. In the
beginning, the network monitored tectonic earthquakes in order to characterise
the baseline of seismicity of the Olkiluoto. When the construction started it also
monitors excavation-induced seismicity [9].
Later on when construction of ONKALO progresses more measurements

such as convergence measurements and acoustic emission will be done and
extensometers installed [7].
3.3 Geochemistry
The geochemical monitoring programme includes mapping of the changes in the

groundwater chemistry caused by the construction and mapping of the influences
caused by the engineering and the foreign materials entering ONKALO [3].
Most of the geochemical measurements have taken place from the surface based
holes, but some underground sampling is also done.
Groundwater samples have been taken both from the overburden and deep
bedrock by different methods. The chemical properties of the water used in the
ONKALO and pumped out from there have been analysed.
Possible movement of the saline waters in the bedrock is studied by Gefinex
400s measurements.
3.4 Environment
Environmental monitoring has been carried out for a long time: the
environmental monitoring performed by Teollisuuden Voima Oy (TVO) started
in the 1970’s. TVO is obliged to carry out monitoring of marine environment
and also of radioactivity of the environment around Olkiluoto, both in the marine
and terrestrial systems and human food chains. Noise, dust and water quality of
the private wells are monitored in campaigns. In addition, aerial photographing
and characterisation of the flora and fauna is included in the programme [7].
The data gathered through the monitoring is used as input for the biosphere
modelling in the long-term safety analysis as well as for the broader
environmental impact analysis [2].
3.5 Foreign materials
During the construction and operation of the ONKALO several types of foreign
materials are introduced to the underground facility, mainly for the engineering
purposes. The amount of the materials, such as cement, rock bolts, labelled water
etc. is monitored and recorded.
4 Results
The monitoring at Olkiluoto is part of the site investigations. The monitoring
activities are done continuously or regularly and always in the same way. The
programme of monitoring will be updated at the latest in 2012, but maybe
earlier.
4.1 Hydrology
The results of hydrological monitoring are studied and analysed to determine the
effects of the construction of ONKALO on the water balance of the reservoir.
The total leakage water inflow into ONKALO is about 16 l/min when length of
tunnel in 990 m. Because of the fairly small inflow into ONKALO no long-term
hydrological changes have occurred so far, fig 3.
PP1 PP2 PP3 PP4 PP5 PP6 PP7
PP8 PP9 PP10 PP31 PP32 PP34 PP35
Shallow boreholes PP36 PP37 PP38 PP39 PA1/1 PA1/2 PA1/3
PA2/1 PA2/3 PA3 PA4/1 PA4/2 PA4/3 PA5
(L, PP, PA, PR) L1 L2/1 L2/2 L2/3 L3/1 L3/2 L3/3
L4/1 L4/2 L4/3 L5 L7 L8/1 L8/2
L8/3 L9 L13/1 L13/2 L13/3 L14/1 L14/2
L14/3 L15/1 L15/2 L15/3 L16/1 L16/2 L16/3
L26 L27 PR1 PR2 PR3 PR4 H7
Sea level ONKALO chainage KR3 L4 KR3 L8 Pori Kuuminainen Daily precipitation
12 140
11 130
0 15 71 733 175 253 307 342 426 502 572 621 688 792 886 940 990
10 120
9 110
8 100
groundwater table level (m.a.s.l)
7 90
precipitation (mm)
6 80
5 70
4 60
3 50
2 40
1 30
0 20
-1 10
-2 0
1.9.2004 10.12.2004 20.3.2005 28.6.2005 6.10.2005 14.1.2006
time
Figure 3: Level of the groundwater table in the shallow boreholes.
4.2 Rock Mechanics
The annual land uplift at the Olkiluoto is 6 mm/year [5]. Precise levelling shows
that two areas rise faster than the rest of the island. The reason to this is that lots
of the rock masses are moved away from these two areas.
Five of the microseismic events are classified uncertain or unusual. After
Saari [10] it is very unlikely that they are microseismic events, but they are not
typical examples of an artificial seismic signals, either, table 1.
No other rock mechanical changes have occurred in Olkiluoto since the
construction started.
4.3 Geochemistry
Some changes in the main ions can be noticed in the groundwater samples taken
from the surface. All the noticed changes are mainly minor and cannot be caused
by the ONKALO activities because most of the samples have been collected
before the construction took place [3].
No other changes in the geochemistry have been observed.
4.4 Environment
No unexplainable deviations from the reference data can be seen so far [2].
Table 1: Events of uncertain origin 2002-2004. Loc. Err = location error,

Mag Loc = local magnitude, Size = estimated size of the seismic
source, ESEP = Energy of S-wave/energy of P-wave and No of
Acc = number of recordings used in analysis [10].
No Date Origin X Y Z
Time (m) (m) (m)
(UTC)
1 19.4.2002 09:41:36.5 6792766.0 1525527.0 -152.8

2 23.4.2002 17:55:47.4 6792785.5 1525545.9 -143.4
3 8.5.2002 00:10:04.2 6792017.5 1525408.3 -81.4
4 24.5.2002 19:59:35.7 6791939.0 1525934.5 -6030.0
5 24.8.2004 14:45:40.2 6791924.0 1526224.5 -364.8
Loc. Mag Seismic Radiated Size ESEP No. No

Err. Loc Mom. Energy of
(m) (log10) (log10) (m) Acc.
(J)
0.0 -2.9 6.5E+00 -3.1E+00 10.2 2.4E-01 1 1

0.0 -3.1 6.1E+00 -3.0E+00 9.1 3.3E+01 1 2
0.0 -3.2 6.1E+00 -3.4E+00 9.1 3.8E+00 1 3
0.9 -0.4 9.0E+00 2.5E+00 4.1 0 5 4
3.4 -0.4 9.2E+00 2.3E+00 16.7 4.7E+00 4 5
4.5 Foreign materials
The use of foreign materials is restricted and regulated in the ONKALO and a
list of appropriate materials that can be used in the ONKALO is established and
continuously updated. Altogether 46 different materials have so far been
authorised.
During the first 990 m of tunnel length a total of about 321 tons cement has
been used. This compares well with the estimations made before the construction
of the ONKALO begun.
5 Conclusions
Many of the possible changes caused by the construction of ONKALO are

driven by the drawdown of the groundwater table, which is a consequence of the
water inflow into the tunnel. Because of heavy use of cement in the upper parts
of ONKALO, the water inflow is only about 16 l/min and because of that no
drawdown has occurred.
Another reason why there have not been any significant changes is a fairly
short time since the construction started.
References
[1] Ahokas, H., Klockars, J. & Lahdenperä, A-M. Results of Monitoring at
Olkiluoto in 2003-2004. Hydrology. Working report 2005-28. Posiva Oy,
Eurajoki, 2005.
[2] Haapanen, R. Results of Monitoring at Olkiluoto in 2004. Environment.
Working report 2005-31. Posiva oy, Eurajoki, 2005.
[3] Hirvonen, H. & Mäntynen, M. (ed.). Results of Monitoring at Olkiluoto in
2004. Geochemistry. Working report 2005-29. Posiva Oy, Eurajoki,
2005.
[4] Juhola, P. Results of Monitoring at Olkiluoto in 2004. Foreign materials.
Juhola, P. Working report 2005-27. Posiva Oy, Eurajoki, 2005.
[5] Posiva Oy. Baseline conditions at Olkiluoto. Posiva 2003-02. Posiva Oy,
Eurajoki, 2003.
[6] Posiva Oy. ONKALO Underground Rock Characterization Facility - main
drawings stage. Working report 2003-26. Posiva Oy, Eurajoki, 2003.
[7] Posiva Oy. Programme of Monitoring at Olkiluoto During Construction
and Operation of the ONKALO. Posiva 2003-05. Posiva Oy, Eurajoki,
2003.
[8] Posiva Oy. TKS-2006. Nuclear waste management of the Olkiluoto and
Loviisa power plants: Programme for Research, Development and
Technical Design for 2007-2009 (in prep).
[9] Riikonen, S. (ed.) Results of Monitoring at Olkiluoto in 2004. Rock
mechanics. Working report 2005-30. Posiva Oy, Eurajoki, 2005.
[10] Saari, J. Seismic Network at the Olkiluoto Site. Working Report 2003-37.
Posiva Oy, Eurajoki, 2003.
Geotechnical evaluation of Stabilized

Dredged Material (SDM) from the
New York/New Jersey Harbor
A. Maher1, A. Sarmad2 & M. Jafari1
1
Department of Civil and Environmental Engineering,
Rutgers University, USA
2
DMJM Harris Inc., New York, USA
Abstract
As a result of the 1997 ban on ocean dumping of dredged sediments, the States
of New York and New Jersey have embarked on a rigorous program of seeking
environmentally friendly solutions to the handling of dredged material, including
the beneficial use of stabilized dredged material (SDM) in roadway applications.
A pilot study was initiated in 1998 to construct two embankments on a site in
Elizabeth, NJ, where SDM was successfully used as a cover for more than 100
acres of commercial development area. The pilot study included a laboratory
phase for geotechnical evaluation of SDM, and a field phase for monitoring and
evaluation of the construction process, as well as the performance of the fills
following construction. The results of the laboratory study, as reported in this
paper, demonstrate that SDM satisfies most of the geotechnical criteria for fill
construction, except those for durability, requiring proper coverage and
protection similar to those provided for fills constructed on cohesive soils.
Keywords: dredged material; contaminated sediments; beneficial use;
embankments.
1 Introduction
The Port of New York and New Jersey is the largest Port on the East coast of the
United States, situated in the metropolitan center of the Hudson Raritan Estuary
complex. The New York / New Jersey Harbor complex is naturally shallow,
with an average depth of 19 feet at low tide. Due to the Port’s strategic position
in regional and international trade, the U.S. Army Corps of Engineers has
doi:10.2495/GEO060431
provided some 250 miles of engineered waterways at depths ranging from 20 to

45 feet. Plans are underway to deepen the main channels to 53 feet during this
decade. Maintenance of these waterways, crucial to safe navigation, requires
dredging 4-6 million yd3 of sediment, or “dredged material”, annually.
Unfortunately, at least half of the material scheduled for removal is contaminated
with industrial chemicals and trace metals from historical and ongoing sources,
making management of the material challenging.
Historically, dredged materials from the channels and berths in the Port have
been relocated to other parts of the Harbor, used to fill in shallows, or dumped in
the ocean. Following the London Convention, the United States Environmental
Protection Agency (USEPA) directed consignees to evaluate dredged material
for its potential environmental impact prior to dredging. Materials found
suitable for open water disposal were to be placed in one or more designated
sites. In the case of the NY/NJ Harbor, this meant placing the material at a 2.2
square mile area off Sandy Hook, NJ, known locally as the “Mud Dump”.
Starting in 1991, further modifications to the ocean disposal testing requirements
resulted in strict restrictions on disposal at the site. In 1993, environmental
groups began legally challenging even the most recent regulations, eventually
resulting in an outright ban of disposal of dredged materials at the site by 1997.
Today, only material considered to be completely free of potential to cause
environmental harm is placed at the site, doubling as a cap of older, more
contaminated materials. Unfortunately, these new regulations did nothing to slow
the rate of sedimentation in the Harbor complex. Berths and channels in this
heavily trafficked system require nearly continuous maintenance to ensure safe
passage of commercial vessels. The Port community was unprepared for the loss
of management options for dredged material. Managers were forced to either
delay dredging or pay sums 15-20 times higher than usual. Dredging has all but
ceased in the Port, threatening the maritime industry.
In response the States of New Jersey and New York, the U.S. Army Corps of
Engineers (Corps) and the Port Authority of New York and New Jersey
(PANYNJ) created teams to find alternative methods for management of
contaminated dredged material. One of the alternatives considered was to seek
beneficial use of stabilized dredged material (SDM) in upland disposal sites.
This entails the stabilization of dredged material with pozzolanic admixtures to
create structural and non-structural fills for various applications, including those
in brownfield development projects and transportation infrastructure systems.
The beneficial use of SDM as a fill has been demonstrated to be cost effective
for high volume usage. For example, approximately 600,000 cubic yards of
SDM were successfully used as structural fill for the construction of parking
areas for the Jersey Gardens Mall in Elizabeth, NJ. In this project, the developer
utilized dredged material amended with Portland cement for the grading, filling
and capping required for the remediation of the landfill. Amending dredged
material with Portland cement yields three benefits: it binds contaminants to the
sediment particles, it removes excess water and it improves the structural
characteristics of the silt and clay particles.
2 Objective
During the course of the Jersey Gardens development project, the Office of
Maritime Resources of the New Jersey Department of Transportation (NJDOT)
initiated a pilot study to evaluate the feasibility of SDM as a fill material for
roadway embankments. Two embankments were constructed on existing
municipal solid waste fills at the Jersey Gardens Mall site using SDM as the fill
material. The project had two phases: a laboratory phase (phase I) consisting of a
comprehensive geotechnical evaluation of SDM for beneficial re-use
applications, and a field phase (phase II) consisting of performance evaluation of
embankments following construction. This paper summarizes the first phase of
the study.
3 Geotechnical properties of SDM

The controlling parameters for the laboratory investigation were the type and the
content of admixtures (cement and fly ash) that were used in the field phase, as
well as the sequence of mixing, curing and placement activities specific to the
project. The preparation of SDM in the field was conducted on the Jersey
Gardens site using a pugmill system. After preparation, the stabilized dredge
material (SDM) was placed on various locations at the site for stabilization for
curing. Unlike typical soil-cement mixtures in which the soil and cement are
mixed and then immediately compacted, the SDM due to its high initial water
content was placed on holding sites while it dried and cured, and the final site
preparations were made. Once the SDM had cured, it was moved to the
embankment sites for final placement, molding and compaction. As a result, a
direct comparison between the SDM used in this project and typical soil-cement
materials could not be made. However, soil-cement properties are used as point
of reference for the evaluation of laboratory results.
Three different mixtures were prepared for the laboratory evaluation; each
using raw dredged material (RDM), Portland cement and fly ash. The recipes
were all mixed on a wet-weight basis. The three recipes used were: 1) RDM
with 4% Portland cement, 2) RDM with 8% Portland cement, and 3) RDM with
8% Portland cement and 10% fly ash. The following tests were conducted to
characterize each mixture:
Unified Soil Classification ASTM D-1140, and D-422

Shear Strength (tri-axial), ASTM D-4767, 2850-87
Swell Pressure ASTM D-4546
Consolidation Test ASTM D-2435
Resilient Modulus AASHTO T274
Hydraulic Conductivity (Permeability) ASTM D-5084
Compaction Test ASTM D-1557
Durability ASTM D-559
Cement Content Determination ASTM D-806-96
3.1 Classification
The dredged material tested in this investigation is mostly silt with low
percentages of fine sand and clay. Sediments dredged from navigational channels
do not naturally contain coarse or medium sand (although incidental pieces of
gravel were found in some samples), because sand will settle before it reaches
still waters. In addition, these sediments cannot contain high percentages of
clay, because clay particles will stay in suspension. However, deepening
dredging in undisturbed areas might result in the generation of material
containing significant amounts of gravel and rock mixed with fine material. This
study did not address this type of material.
The SDM samples tested consisted, on average, of 66% silt, 14% clay and
16% fine and medium sand (12.1% fine, 3.9% medium). Gravel content was
negligible except for one sample, which contained 6.5% gravel. The percentage
of clay size particles was higher for those SDM samples that had been mixed
with fly ash, presumably due to the fine nature of fly ash particles. The organic
content of the raw dredge material was determined to be around 8%, according to
ASTM D2974.The effect of increased curing time on particle size distribution
was minimal. Any variation in particle size is attributable to size variation in the
source material. In addition to the gradation test, SDM samples were also tested
for plasticity index. Based on the Atterberg Limits, all the samples tested are
below the A-line and to the right of the LL=50 line on the Plasticity Chart.
Therefore, the SDM could be classified as Elastic Silt (MH).
3.2 Moisture-density relationship
According to the test results, maximum dry densities ranged from 76.6 pcf to
78.8 pcf (1.23 to 1.26 Mg/m3), and optimum moisture contents ranged from 26%
to 31.5%. A slight reduction in maximum dry density was observed when the
percentage of cement and the curing time were increased prior to compaction of
the material. This is similar to findings made by Kezdi [4], where the maximum
dry densities of cement-treated silts were found to decrease slightly with
increasing cement content.
3.3 Consolidation
Laboratory consolidation tests were conducted according to the ASTM D-2435

method. The samples were prepared using RDM amended with 4% Portland
cement, 8% Portland cement, and 8% Portland cement with 10% fly ash. The
SDM mix was remolded into a consolidometer with different compaction efforts
applied. To determine the level of compaction achieved with each sample, a
compaction test conforming to ASTM D-1557 was conducted for each recipe.
According to the test results, samples were compacted to varying degrees
ranging from 59% to 90% of their maximum dry density.
The moisture contents used when the test samples were remolded were
chosen to represent the site’s average and approved layers that did not meet the
85% Modified Proctor criteria. Samples were tested after one and six months of
curing. The energy applied for remolding the sample prior to the test played a
major role in the consolidation behavior of the material. The test results indicate
pre-consolidation stresses (Pc) as high as 8.7 tsf (833 kPa) once the sample is
compacted to 87% of its modified maximum dry density. This means that the
compacted material will compress before experiencing 8.7 tsf (833 kPa) of
overburden (equivalent to approximately 170 feet ( 52 m) of SDM, unit weight
of 100 pcf (1.6 Mg/m3), or 133 feet (40 m) of compacted granular fill unit weight
of 130 pcf – 2.08 Mg/m3). However, Pc as low as 1.32 tsf (126.4 kPa) was
recorded for a sample compacted to 86% of its modified maximum dry density.
The average value of Pc, for samples compacted from 81% to 90% of their
modified maximum dry density, is higher than 5 tsf (478 kPa).
The compression index (Cc) values range from 0.22 to 0.9. Both of these
values were recorded for SDM with 8% Portland cement. In general, for all
recipes tested, once compaction reaches 81%, the compression index will not
exceed 0.5. In that case, a Pc of 2 tsf or more should be expected. The
compression ratio (CR =Cc/1+e0) varied from 0.085 to 0.24. This value did not
exceed 0.19 for samples compacted to 83% or above.
The results also show that based on consolidation settlement estimates, SDM
embankments could be constructed to a height of 50 feet (15 m) with negligible
settlement taking place within the SDM fill. This conclusion is supported by the
results of the field settlement program. In the case of the two embankments in
this study, and in similar cases where construction is proposed on marginal
foundation soils, settlement is primarily a function of the foundation soil and its
consolidation characteristics.
3.4 Permeability ASTM D-5084
Twenty-four samples were prepared and tested for permeability (hydraulic

conductivity). Three different recipes for amending RDM were used in the
sample preparation: 4% Portland cement, 8% Portland cement, and 8% Portland
cement with 10% fly ash. The three different recipes were sampled at one
month and at six months. Half of the samples were compacted to 85% and the
other half were compacted to 90% of their maximum dry density, as determined
by Modified Proctor (ASTM D-1557). The permeability results ranged from
1.25x10-6 cm/sec to 4.3x10 –7 cm/sec. The lowest values were recorded for
samples of RDM amended with 8% Portland cement and 10% fly ash. Also,
samples amended with 4% Portland cement generally had lower permeability
than did samples amended with 8% Portland cement. This may be due to the
apparent effect of cementation on imposing a flocculated fabric arrangement in
SDM. In general, tests results indicate that SDM could be considered for use as a
low permeability layer in landfill cap applications. For roadway applications,
however, building roadways on SDM would be similar to building on compacted
fine-grained sub-grades, such as those used in arid regions like Arizona, Texas,
etc. For roadway construction, proper coverage must be provided using an
appropriate base or sub-base materials.
3.5 Shear strength
The strength parameters of SDM were evaluated for feasibility of SDM as a fill
material, and specifically for the slope stability of the pilot embankments. The
consolidated undrained (CU) shear condition was determined to best reflect the
realistic field conditions both during construction and post-construction periods.
Both one and six-month-old samples of the three different recipes for SDM were
tested for shear strength characteristics under CU conditions (ASTM D-2850-
87). The samples were compacted to 85% and 90% of their modified maximum
dry density and total and effective strength parameters were determined for
stability analysis.
The effective C and φ or (C′ and φ′) were calculated after the Mohr circles for
effective stresses were plotted. As expected, the effective friction angle values
were generally larger than the total values for SDM. No significant change or
trend in the magnitude of the frictional angle, and, with the addition of cement
and fly ash could be observed. This is similar to previous findings by Balmer [1],
Clough, et al. [2] and Van Riessen and Hansen [6]; where different soil types,
amended with varying cement contents, were extensively tested and showed no
significant change in frictional angle as a function of the varying amount of
cement.
In general, an average angle of 34o can be estimated for long-term stability
analysis of embankments constructed with SDM. On average, there is an 8°
increase in the effective friction angle compared with the total friction angle.
Cohesion, however, decreases as the friction angle increases.
The test results also showed that compaction plays a significant role in the
magnitude of strength parameters. For all the samples tested, a 5% increase in
dry density resulted in increased strength. On average, the un-drained C values
increased by 35%. Moreover, the average increases in φ′ and C′ were 1 % and
50%, respectively. On this basis, it can be concluded that compaction is the most
important physical stabilizer of SDM with respect to strength parameters.
A general comparison of SDM with typical soil-cement and cement-modified
soils shows that with the same percentage of added cement, and similar
compaction efforts (90% of optimum for SDM, and optimum for soil-cement)
cement-modified soils are denser than SDM, have slightly higher friction angles,
and have a much higher cohesion intercept under triaxial shear conditions.
3.6 Resilient modulus (AASHTO TP46-94)
The resilient modulus is a dynamic soil property used in the mechanistic design
of pavements. The test provides a means of characterizing base, sub-base and
sub-grade materials under simulated field loading conditions and is the basis for
a deterministic approach to pavement design. In the resilient modulus test, the
materials are tested under a variety of conditions, some of which include stress
state, moisture content, temperature, gradation and density. A detailed
description of the test and sub grade resilient properties of NJ soils is given by
Maher, et al [5].
For this study, the resilient properties of SDM were determined for all the
mixture types used. Table 1 summarizes the resultant resilient modulus values
for SDM mixtures and those for three New Jersey sub-grade soils that currently
underlie roadways in New Jersey. According to the table, SDM compares
favourably to the soil taken from various subgrades in NJ, indicating sufficient
resiliency under dynamic loads.
Table 1: Comparison of resilient modulus values between SDM and typical NJ

base materials.
Sample Stockpiling Compaction Resilient Resilient

Type Period Effort Modulus Modulus
(months) (%) (Psi) (MPa)
4% PC 1 85 4827.5 33.28
90 7720.2 53.22
4% PC 6 85 5167.9 35.62
90 8752 60.34
8% PC 1 85 11911 82.12
90 12326 84.98
8% PC 6 85 8432 58.13
90 8945 61.67
8%PC + 10% FA 1 85 5610 38.68
90 9254 63.80
8%PC + 10% FA 6 85 1498 10.32
90 6601 45.51
Rt. 23 in NJ max dry density 9633 66.42
(medium to fine sand)
(medium to fine silty sand)
(silt with fine sand)
3.7 Swell potential
Samples of SDM were also tested for swell pressure in order to determine if
SDM could be used in applications where the material would be in contact with
structures sensitive to swell pressures and excessive deformations. For example,
if SDM were used as a base material in roadways, excessive swell pressures and
deformations will be detrimental to the integrity of the pavement.
For this study, samples of RDM were mixed with 4% Portland cement, 8%
Portland cement, and 8% Portland cement plus 10% fly ash. Samples were cured
in the laboratory for one month and for six months. These samples were then
compacted to different densities in order to determine at what point the density
level and moisture content would become critical in generating excessive swell
pressure and deformation. Swell tests were performed in accordance with ASTM
D-4546.
The laboratory data indicate several trends. The strain or percent swell was
not significant for any of the samples tested. The strain values ranged from 0.1
to 1.2 percent, with an average of 0.6. The maximum strain belonged to the
sample amended with 8% Portland cement plus10% fly ash (1.2%). This
magnitude of volume change is considered low and, therefore, not detrimental to
adjacent structures. The swell pressure, however, was high for samples
compacted to 94% or higher of their maximum dry density with moisture
contents on the dry side of optimum. For these samples, the overall average
swell pressure was 1.005 tsf (96.25 kPa). The average for one-month old
samples was slightly higher at 1.34 tsf (128.32 kPa), with an average strain of
1.1%. Although strains were not high for any of the samples tested, the swell
pressure generated was moderate. For SDM that was mixed with 8% Portland
cement and compacted to 95% of its maximum dry density, the swell pressure
was measured as high as 1.96 tsf (187.69 kPa). However, considering low
associated strains, SDM would not have any detrimental effect on adjacent
structures.
For samples compacted on the wet side of their optimum moisture content,
much lower swell pressures and strains were measured. The average swell
pressure for those samples was 0.14 tsf (13.41 kPa), and the average strain was
0.3%. This is due to the fact that fine-grained soils have a flocculated structure
at low moisture contents (below optimum moisture content). At moisture
contents above optimum, the structure of the soil particles becomes more
dispersed and layered. For dispersed structures, additional moisture does not
result in significant volume changes.
3.8 Durability
The major durability concerns regarding SDM include potential strength loss due
to freeze-thaw cycles and moisture variation. The freeze-thaw test simulates the
internal expansive forces that result from the moisture in fine-grained soils.
During freeze-thaw cycles, SDM experiences an increase in volume and a loss in
strength. Some soil-cement mixtures have the ability to regain strength under
certain conditions; specifically, the availability of reactive Calcium Oxide,
adequate temperature and a high pH environment. For SDM, these conditions do
not exist; therefore, any strength loss will be permanent.
In order to study the effects of freeze-thaw cycles on SDM, samples were
prepared from the three different recipes. The testing was performed in
accordance with ASTM D560. Samples were compacted to 85 and 90% of their
maximum dry density, as determined by Modified Proctor. To provide a point of
reference, a natural clay sample was also tested for its behavior during freeze-
thaw cycles.
According to the test results, none of the samples could withstand more than
three freeze-thaw cycles before failing. Significant volume change (ranging
from 1.8% to 58%) was experienced during testing. Considering that the
average volume change for the natural clay sample was 2%, it may be concluded
that the freeze-thaw effect is several times more severe for SDM than it is for
natural clay. As a result, all SDM should be protected against frost in order to
maintain the cement contents within the percentages used for this project. Frost
depth in New Jersey is approximately 2.5 to 3 feet (0.75 to 0.9 m). Under these
conditions, SDM should be kept at least three feet below the surface. This should
apply to both pavements and embankment slopes.
Wet-dry tests are conducted to simulate shrinkage forces in cement-modified
or soil-cement specimens. Wet-dry cycle tests were conducted on the three
different recipes of SDM. Tests were conducted according to ASTM D-559. All
of the samples with the exception of one (8% PC @ 90% Modified Proctor)
collapsed before experiencing 12 wet-dry cycles. Volume changes were in the
range of 10% to 48% of the original volume. Therefore, SDM should be
protected against frequent wet-dry cycles with placement of proper coverage for
roadway applications, or low permeability layers in general fill applications.
Furthermore, if SDM is compacted at moisture contents below the shrinkage
limit, the potential for the development of tensile cracks and a consequent loss in
strength could be minimized.
4 Conclusions and recommendations

Beneficial use of stabilized dredge material (SDM) has been shown to be a
practical option for the management of navigational dredged material in the Port
of NY and NJ. The laboratory study described in this paper evaluated the
geotechnical properties of stabilized dredge material (SDM) from the NY/NJ
Harbor for potential high volume applications in roadway construction. The
study was the first phase of a two-phase pilot project sponsored by the New
Jersey Department of Transportation for finding alternative methods for
beneficial use of the 2-4 million yd3 of contaminated sediments dredged
annually, to maintain the maritime the transportation system that serves the Port.
The results of the laboratory study demonstrate that stabilized dredge material
(SDM) satisfies most of the geotechnical criteria for construction of fills and
embankments, except those for durability: freeze-thaw and wet-dry cycles.
Proper coverage and protection need to be provided for SDM fills to address the
durability problem, similar to those addressed in the construction fills with
cohesive soils. A summary of the test results as described in the paper is as
follows:
1. The raw dredged material from the NY/NJ Harbor is mostly silt with
low percentages of fine sand and clay. The dredged material samples
tested in this study consisted of 66% silt, 14% clay and 16% fine and
medium sand (12.1% fine, 3.9% medium). The percentage of clay size
particles was higher for those stabilized samples that had been mixed
with fly ash. This is due to the fine nature of fly ash particles. The
organic content of the raw dredge material was determined to be around
8% according to ASTM D2974. Based on the Atterberg Limits, all the
samples tested are below the A-line and to the right of the LL=50 line
on the Plasticity Chart, classifying SDM as Elastic Silt (MH).
2. The maximum dry densities for the different mixes tested ranged from
76.6 pcf to 78.8 pcf (1.23 to 1.26 Mg/m3), and optimum moisture
contents ranged from 26% to 31.5%. A slight reduction in maximum
dry density was observed when the percentage of cement and the curing
time were increased prior to compaction of the material.
3. The compression index (Cc) values for SDM ranged from 0.22 to 0.9.
and did not exceed 0.5 for any of the samples, once the samples had
been compacted to 81% of their maximum dry density. Therefore, a Pc
of 2 tsf (191.52 kPa) or more should be expected. The compression
ratio (CR =Cc/1+e0) varied from 0.085 to 0.24. It can be concluded that
SDM embankments up to 50 feet (15 m) in height could be constructed
with only minimal settlement within the SDM fill.
4. The hydraulic conductivity ( permeability) results ranged from 1.25x10-
6
cm/sec to 4.3x10 –7 cm/sec. SDM could, therefore, be considered for
use as a low permeability layer in landfill cap applications. In roadway
applications, however, building on SDM fills would be similar to
construction on compacted fine-grained sub-grades, such as those in
arid regions like Arizona, Texas, etc. Proper coverage must be provided
using appropriate base or sub-base materials.
5. The addition of admixtures produced no significant change or trend in
the frictional properties of SDM. In comparison to soil-cement and
cement-modified soils, SDM has lower friction angle and much lower
cohesion intercept under triaxial shear conditions mainly due to the
sequence of sample preparation used in this study which followed the
field operations. Temperature had a major effect on the curing process
of SDM at temperatures below 40°F; it is recommended that SDM be
placed during warm seasons (e.g., April through October in New
Jersey).
6. The resilient modulus values for all of the samples tested compared well
with three sub-grade soils that are currently under New Jersey
roadways.
7. The strain or swell percentage was not significant for any of the samples
tested. The strain values ranged from 0.1% to 1.2%, with an average of
0.6%. This magnitude of volume change is considered to be low and,
therefore, not detrimental to adjacent structures.
8. The results from durability tests indicate that SDM is susceptible to
frost action (several times more susceptible than natural clay) and
should be placed below frost line. Based on the wet-dry tests, proper
soil cover needs to be provided at all times to minimize strength loss
due to erosion. Compacting SDM at moisture contents below the
shrinkage limit would minimize the potential for tensile cracks and
thereby minimize any further strength loss in the material.
References
[1] Balmar, G. G. (1958). “Shear strength and elastic properties of soil-cement
mixtures under triaxial loading.” PCA, Bulletin, D32.
[2] Clough, G. W., Sitar, N., Bachus, R. C., and Rad, N. (1981). “Cemented
sand under static loading.” JGED, ASCE, Vol. 107, No. GT6, 799-817.
[3] Cotton, M. D. (1962). “ Soil-cement technology – a resume.” Journal of the

PCA Research and Development Laboratories, Vol.4, No.1, 13-21.
[4] Kezdi, A. (1970). “Handbook of soil mechanic,” VEB Verlag fur Bawsen,
Berlin.
[5] Maher, A., Bennert, T., Papp, W., (2000). “Resilient properties of New
Jersey sub-grade materials,” Report submitted to NJDOT.
[6] Van Riessen, G. J., and Hansen, K. (1992). “Cement-stabilized soil for coal
retaining berms.” ASCE, Geotechnical STP 30, 981-992.
The impact of soil quality on cocoa yield

in Nigeria
O. A. Amusan1 & F. O. Amusan2
1
University of Bonn, Institute of Agriculture,
Water Engineering and Land Improvement, Bonn, Germany
2
Alcon Labs, UK
Abstract
The Southwestern region of Nigeria is the largest administrative region in
Nigeria; it occupies about 30 percent of Nigeria and has an estimated population
of about 40 million. Over 50 percent of the area is categorised as rural areas of
which about 90 percent of the people depend on farming for their livelihood.
Over 90 percent of Nigerian cash crop cocoa is produced in the cocoa belt of the
Southwestern region, but both cash and food crops have consistently declined in
the last few years. This phenomenon constitutes a threat to food security and
calls for efforts to explain the downward trend and make recommendations for
improvement. The objectives of this study were to evaluate the soils of some
areas in Southwestern Nigeria for cocoa on the one hand, and identify factors
affecting cocoa yield on the other. A novel technique that combines soil survey
with socio-economic analyses was adopted. Socio-economic surveys covered
resource quality and constraints to agricultural production, whereas soil sampling
and analyses were carried out to assess the contribution of soil to yield. Three
locations having similar agro-ecological features were selected, namely Ibadan,
Ife and Akure. Cocoa farmers were randomly selected and interviewed on their
farms using standardised questionnaires to elicit information on factors affecting
crop yield. Relationships between cocoa yield and variables presumed to
influence yield were determined using linear multiple regressions. Soil organic
C, Age of farm soil, and ECEC were identified as the major constraints to yield.
Other variables are related to biophysical and management factors. It is
recommended that emphasis should be placed on soil management techniques
that conserve organic matter and enhance the nutrient and water holding capacity
of the soils. Policies that would enhance sustainability of agricultural land use
and crop marketing are also required.
Keywords: cocoa yield, food security, land use, linear multiple regressions, soil
organic C, soil quality.
doi:10.2495/GEO060441
1 Background and aim of the study

The main soil factor affecting cocoa in Nigerian conditions is the clay content.
This is important chiefly because a relatively high amount of clay in the soil
gives it a fairly high capacity for retaining moisture, which remains available to
the trees during the dry season. Soils in Southwestern Nigeria exhibit varying
degrees of fertility. They are derived from hard crystalline rocks of the
“Basement Complex” (gneisses, granites, schists, quartzites, and amphibolites),
which have been eroded in varying degrees in different parts of the main cocoa
belt. Fertile soils are found next to areas with poor soils and contiguous tracts of
fertile cocoa soil are rare. Pockets of fertile soils can be found around Ife, Ibadan
and Akure (Figure 1).
Figure 1: Map of the cocoa growing regions of Nigeria.
Cocoa grows best in soils with the following properties (Rehm and Epsig [1]):
1. A dark, grey brown, crumby humid surface layer 3.5–7.5 cm in
thickness consisting of medium texture and free from coarse sand, grit
or small gravel.
2. About 30cm of medium-textured earth, brown or reddish-brown in
colour.
3. A layer of friable red clay 60–90 cm in thickness
4. A layer of friable clay mottled whitish, orange, red and pale yellow, and
60–90 cm thick which grades into
5. Rotten, disintegrated rock, preferably of a granite nature or derived
from dark-coloured igneous or volcanic material.
Cocoa-growing year, distinguished from cocoa-marketing year begins when
moist air (the Equatorial Maritime air mass) sweeps in from the coast at the end
of April or the beginning of May. From May to July and in September and
October, rainfall averages at least 150mm per month in nearly all cocoa-growing
areas. The “little dry season” in August when the climate remains humid but
rainfall is much less is characteristic of Southwestern Nigeria. If it is unduly long
or dry, the cocoa crop may suffer severely. The crop may also suffer severely if
the rains between July and October when the pods are ripening are unduly heavy.
The northern limit of cocoa belt is marked partly by the line where rainfall
approaches the lower limit of 1100mm/ year and partly, in the area of Ondo
provinces where the rainfall is well above that limit. In this zone the influence of
the dry Harmattan wind in the months from December to March is very strong
and cocoa, which cannot stand long drought, seems unable to survive.
Temperatures are lowest and relative humidity highest in the rainy months (the
growing season) and plant growth is then very rapid. In the dry season,
temperatures are high during the day and relative humidity falls very low during
the afternoons. Conditions are then unfavourable to growth. The period during
which even some of the hardiest crops cannot be successfully planted without
constant watering is rather over three months in most of the main cocoa belt.
Even at the highest elevations in the cocoa-growing areas temperature never fall
so low as to be adverse to growth. Farmers therefore face replanting problems in
old cocoa farms and high seedling failure in new plantings in marginal soils.
This situation is affecting the future planting plans of cocoa farmers in Oyo,
Ogun and Ondo States of Southwestern Nigeria. Although contiguous land
masses suitable for cocoa growing exist outside the main cocoa belt (particularly
in Ikom and Baissa of South-eastern Nigeria), the exceptional productivity
arising from the fertile soil conditions has been eroded by the high incidence of
Phytophthora (black pod) disease due to high rainfall. The disease is more
prevalent in the high rainfall areas of the Southeastern state of Cross River than
in the relatively drier cocoa areas of Southwestern Nigeria. This study is aimed
at assessing the impact of soil quality on Cocoa yield in the study areas in
Southwestern Nigeria. The specific objectives include:
1. To evaluate the soils of some areas in South-western Nigeria for cocoa
2. To identify factors affecting cocoa yield
3. To determine the biophysical and socio-economic constraints to quality
cocoa production
2 Methodology
2.1 Data set
The analysis presented in this paper is based on primary data, which were
collected by surveys of farm households, extensionists, farm mangers, etc. The
Cocoa farm household samples (classified by farm size) were selected from three
locations having similar agro-ecological features, namely Ibadan, Ife and Akure.
The primary data from the socio-economic survey were used to elicit information
on farm–farmers’ characteristics vis-à-vis constraints to cocoa. Soil sampling
and analysis were carried out to examine the role of soil properties in crop yield.
Socio-economic and soil data were thereafter integrated to study their influence
on crop yield.
2.2 Socioeconomic survey
A survey was carried out among cocoa growers in South-western Nigeria in the
year 2002. Three locations having similar agro-ecological features were selected,
namely Ibadan, Ife and Akure. Thirty cocoa farmers were randomly selected and
interviewed on their farms using standardised questionnaires. Farmers
interviewed in Ibadan belonged to three different groups of cooperative societies
or farmers’ organisation. The farmers interviewed at Ife and Akure belonged to
the respective cooperative multipurpose unions (CPMU) in these towns. These
cooperatives are profit -making organisations, however, they offer more than just
marketing services for their members. Access to the farmers and their farmlands
was obtained under the auspices of the Association of Nigerian Cooperative
Exporters Limited (ANCE) and ongoing Project for Improvement of Cocoa
Marketing and Trade in Nigeria (ICMT) sponsored by the International Cocoa
Organisation (ICCO). While cocoa farmers in Ife and Akure are fully integrated
into the ICMT project, the cocoa farmers in Ibadan are non-participants. The
selection along the line of participation or non-participation on the project
moreover enabled the evaluation of the project’s overall influence on cocoa
yield. The interview comprised of qualitative and quantitative components,
covering aspects of agronomic activities, resource quality and availability,
problems perceived and objectives set both at individual and cooperative levels.
Additional information was collected in key-persons interviews from cocoa
exporters, cocoa processors, cocoa researchers and representatives of
governmental and non-governmental institutions. Official statistics and other
secondary data served as background information.
2.3 Soil sampling and analyses
Biophysical data was obtained by analysis of soil samples taken from cocoa
farms. Soils were analysed for chemical analyses. The soils were analysed for
basic cations (determined in 1N NH4OAc), total N (Kjedahl method), available P
(Bray P method) organic C (Walkey-Black wet oxidation method) and pH (0.1M
CaCl2).

3.1 Soil properties, land use history and crop yield
The topography of the land at all study farms ranges from flat to gentle slope
(2-8%). Soil analyses (Tables 1and 2) showed that the soils are fairly high in
total N (mean = 0.20%), organic C (mean = 1.89%), available P (mean = 3.19
ppm), but low in Na. The pH is close to neutral (mean 6.5). In terms of
variability, pH is the least variable (cv = 7%) while available P is the most
variable property (cv = 69%). Statistics of soil properties at the sampled
locations are presented in Tables 5 and 6. Soil pH is the least variable across
locations (CV=3-9%). Highest CV (100%) among Ibadan soils occur for EA,
while the highest CV of 119% occur for P for Ife soils while Ca and K manifests
the most variable soil property amongst Akure soils (C=55%). This suggests that
cations are the most variable properties among Akure soils. Test of mean
differences show that none of the properties are significant across locations
(p>0.05). This is because the soils have developed from similar soil parent
materials derived from Precambrian rocks.
Table 1: General Statistics of soil properties (N=16). (EA= Exchangeable

Acidity, ECEC= Effective cation exchange capacity.)
Ca Mg Na K EA a ECEC a Total Availabl Orga- pH

(cmol (cmol (cmol (cmol (cmol (cmol N -e P nic C
/kg) /kg) /kg) /kg) /kg) /kg) (%) (ppm) (%)
Mean 4.34 1.11 0.07 0.38 0.02 5.90 0.20 3.19 1.89 6.46
Std 2.08 0.48 0.02 0.21 0.02 2.55 0.05 2.19 0.61 0.44
Cv (%) 48 43 29 25 100 43 27 69 32 7
min 1.49 0.30 0.04 0.06 0.00 2.15 0.14 0.99 1.16 5.70
max 7.51 1.22 0.10 0.76 0.07 9.77 0.33 7.05 3.33 7.30
3.2 Relationship between land use history and crop yield
Yields obtained differ depending on the variety of cocoa planted, age of the trees
and farm maintenance. Higher yield was observed on farms planted
predominantly with Amazon cocoa in comparison with those planted with
Amelonado. Older farms have lesser yield than younger farms and farms are
better maintained in regards to cultural operations were more productive.
Table 18 presents the yield recorded on the farms during the 2001/02 cocoa
seasons.
Table 2: Location statistics of soil properties (N=16).
Location No of Statistics Ca Mg Na K EA ECEC N P Org.C pH

samples
Ibadan 8 Mean 4.46 1.09 0.07 0.33 0.02 5.96 0.19 3.48 1.84 6.45
Std. dev 2.06 0.44 0.02 0.23 0.02 2.39 0.05 2.53 0.57 0.49
CV (%) 46 40 29 70 100 40 26 73 31 8
Ife 4 Mean 3.89 0.97 0.06 0.42 0.02 5.36 0.21 5.5 2.11 6.56
Std. dev 2.37 0.47 0.01 0.12 0.01 2.74 0.08 6.56 0.88 0.57
CV (%) 61 48 17 29 50 51 38 119 42 9
Akure 4 Mean 4.73 1.31 0.08 0.44 0.02 6.58 0.2 2.33 1.79 6.38
Std. dev 2.61 0.64 0.02 0.24 0.01 3.44 0.04 0.93 0.45 0.22
CV (%) 55 49 25 55 50 52 20 40 25 3
Table 3: Yield statistics (kg/ha) on surveyed farms in the 2001/ 02 season.
Location No of farmers Mean Std. Deviation CV (%)

Ibadan 10 245 86 35
Ife 10 347 166 48
Akure 10 434 189 44
Table 3 shows that average yield per hectare is highest in Akure and lowest in
Ibadan. However Ife has the highest variability in yield. Mean yield among the
sampled locations and coefficient of variation may largely reflect variability in
management practices. Test of mean differences in yield at the locations is
presented in Table 4.
Table 4: Least Significant Difference (LSD) test of yields.
Location (1) Location (2) Difference in mean yield Significant

(Location 1-Location2) in Probability
kg/ha
Ife Ibadan 102 0.15
Akure Ibadan 189 0.01**
Akure Ife 87 0.22
Table 4 shows that yield differences in Ibadan and Akure are highly
significant (p<0.01), while yield differences at other pairs of locations are not
significant (p>0.05). Ife and Akure farmers participate in the on-going ICMT
project of the International Cocoa Organisation (ICCO) and therefore benefit
from the expertise knowledge received at the farm management training sessions.
Statistical relationship between land use history and crop yield were studied
in terms of recorded yield, and three other variables: age of farm, proportion of
dormant trees, and proportion of trees replaced. Figure 2 shows that yield of
cocoa generally increases with age of farm up to 35 years, and then declines. Age
of trees explains 29% of variability in cocoa yield. This phenomenon may be
related to the physiology of the crop, crop varieties and the fact that the input
management strategy in terms of fertilizer application is generally low. Yield
decreases with the proportion of dormant trees on farmers’ field (Figure 3).
Highest yields (650-750 kg/ha) were obtained in fields with no dormant trees. As
dormancy is a function of several factors such as age, disease and pest attack etc,
the need for proper control of diseases and succession planting cannot be over-
emphasised. Dormant trees explain 81% of variability in cocoa yield. Figure 4
shows that yield of cocoa is strongly related to the proportion of replaced trees.
This management option is particularly appealing to farmers given the fact that
the application of farm inputs like fertilizers and pesticides is generally low.
Proportion of replaced trees explains 89% of variability in cocoa yield.
Figure 5 shows that the tendency of cocoa yield to be low at extreme values
of soil organic C (that is, organic C below 1.5% and above 2.5%). The same
observation holds for total N where yields are observed to be low for N values
below 0.2% and above 0.25% (Figure 6). This result suggests that accumulation
of organic materials in the soil is not necessarily favourable to the growth of
cocoa. Figure 7 also suggests that pH below 6.0 and above 6.5 adversely affects
yield. Yield is maximum at about a pH of 6.3 and attempt should be made to
maintain soil pH around this level.
3 2
800 y = -0 .0 7 x + 4 .7 5 x - 6 1 .1 0 x - 7 5 .0 2
2
R = 0 .2 9
600
yield (kg/ha)
400
200
0
20 25 30 35 40 45 50
A g e (y e a r s )
Figure 2: Relationship between yield and age of farm.
800
2
y = 9 .9 6 x - 1 3 8 .3 6 x + 6 3 2 .1
2
600 R = 0 .8 1
yield (kg/ha)
400
200
0
0 2 4 6 8 10 12
d o r m a n t tr e e s (% )
Figure 3: Relationship between yield and proportion of dormant trees.
2
y = 0 .0 7 x + 7 .0 0 x + 1 7 1 .2 6
800 2
R = 0 .8 9
600
yield (kg/ha)
400
200
0
0 10 20 30 40 50 60
r e p la c e d tr e e s (% )
Figure 4: Relationship between yield and proportion of replaced trees.
2
y = -1 2 7 .1 5 x + 4 7 2 .8 8 x - 2 3 .9 3
800 2
R = 0 .1 0
600
yield (kg/ha)
400
200
0
0 .5 1 1 .5 2 2 .5 3 3 .5
o r g a n ic C (% )
Figure 5: Relationship between yield and soil organic carbon.
800 2
y = -2 4 0 1 6 x + 1 0 4 2 1 x - 6 8 7 .6 6
2
R = 0 .1 3
600
yield (kg/ha)
400
200
0
0 .1 0 .1 5 0 .2 0 .2 5 0 .3 0 .3 5
to ta l N (% )
Figure 6: Relationship between yield and soil total nitrogen source.
3 2
y = 4 0 0 .2 7 x - 8 2 1 5 .8 x + 5 5 8 3 7 x - 1 2 5 3 0 8
800 2
R = 0 .2 2
600
yield (kg/ha)
400
200
0
5 5 .5 6 6 .5 7 7 .5 8
pH
Figure 7: Relationship between yield and soil acidity.
Thus proper monitoring of acidity in the soils is crucial. Relationships

between some soil variables are presented in Figures 8 and 9. Figure 8 shows the
tendency of pH to increase with the organic C content of the soil. This has the
potential to decrease yield (see Figure 5). The strong relationship between
organic C and total N (figure 9) suggests that management of soil organic C is

crucial to maintaining adequate levels of N in the soil.
8 3 2
y = 0 .3 1 x - 1 .8 4 x + 3 .8 5 x + 3 .6 9
2
R = 0 .4 4
7
pH
5
0 .5 1 1.5 2 2.5 3 3.5
o rg a nic C (% )
Figure 8: Relationship between pH and organic carbon.
0 .4 y = 0 .0 8 x + 0 .0 4
2
R = 0 .9 3
0 .3
total N (%)
0 .2
0 .1
0
0 .5 1 1 .5 2 2 .5 3 3 .5
o r g a n ic C (% )
Figure 9: Relationship between total N and Organic C.
3.3 Effects of biophysical and management variables on crop yield
Linear multiple regression was used to relate crop yield to biophysical and socio-
economic data. Owing to high level of multicollinearity, independent variables
were reduced to six. Among the variables in the model, two (Organic C and Age
of farm) are negatively related to cocoa yield, whereas other variables are
positively related to cocoa yield. However, soil variables are not significant to
the model (p>0.05), whereas three management variables (plant density,
proportion of dormant plants replaced and crop variety) are significant (p<0.1).
All the variables explain 97% of the variability of yield and the model can be
used to predict yield at 99% confidence level. Figure 10 further shows the
relative importance of the variables as measured by the standardized β
coefficients. Proportion of dormant plants replaced is the most significant
management variable affecting yield, followed by crop variety (that is, F3
Amazon).
0.6
0.5
st andar di z ed bet a
0.4
0.3
0.2
0.1
-0.1
Organic C Age of Plant Proportion Crop ECEC

farm density Replaced Variety
Figure 10: Relative importance of variables in the multiple regression model.
4 Conclusions and recommendation

Seventy percent of the cocoa farmers in the study areas are more than 60 years of
age and over 50% of the farms are older than 45 years. More than half of the
farmers are illiterates who keep no management records. Furthermore, yields on
over 50% percent of the farms are below 300 kg/ha. These figures pointedly
show a relationship between old age of farmers / farms and poor yield on the one
hand and poor or no education on the other. The world market is increasingly
flooded with average or low quality cocoa from new hybrid varieties. As long as
the market is flooded with relatively large amounts of cocoa, buyers are not
forced to pay a premium for cocoa of whatever origin. To the contrary, they can
enforce large discounts for lower quality. They are able to select and mix
according to their needs, all at relatively low cost. On the other hand, a small
producer experiencing severe problems to expand production, or quantity, will
probably find it more attractive to concentrate on improving quality. Once the
market tightens up, the potential quality premium can even be a valuable asset
and provide a competitive advantage. Achieving this improvement in quality will
however depend, in the first place, on the physical potential of the product.
Farmers who apply chemical inputs, must consider the aspect of profit making
and quality of cocoa due to chemical residues in the product, coupled with the
problem of environmental pollution through chemicals. Farmers should be
encouraged to practise conservation and rehabilitation agriculture. This involves
all steps which negate the processes of degeneration on cocoa farms, for example
through consequent replacement of dormant trees, control of tree population for
effective ground cover, properly managed canopy, integrated management
systems that utilise resistant cultivars and application practices that emphasise
proper timing and effective use of fertilizers at lower application rates. This
would help not only to improve yield but also has the advantages of profitability,
product quality and environmental protection. Given the situation in the
international market in and the natural limitations of the cocoa farmers in
Nigeria, the development of a market niche for quality was suggested. Further
research will need to find out whether new breed varieties have the physical
potential to achieve the said improvement in quality. The traditional Amelonado
cocoa has been proved to be superior in every physical aspect (shell content, nib
yield, fat yield and moisture content) when compared to beans derived from
other varieties. The Amazon variety which is now very common in Nigeria, in
addition to the suitable cocoa flavour characteristics, are more resistant to
diseases in the environment and give better yields.
References
[1] Rehm, S. & Epsig, G., The Cultivated Plants of the Tropics and
Subtropics, Verlag - Josef Magraf: Germany, 1991
[2] Amusan, O.A., Amusan, F.O., Braimoh, K.A. & Oguntunde, P.G., Quality
Management Practices in Cocoa Production in South - Western Nigeria:
Deutscher Tropentag 2005, ID–29 Online. www.tropentag.de
[3] Adegbola, M.O.K., Rehabilitating cocoa swollen shoot virus infected
Cocoa Farms- the Offa-Igbo experience. Proc. 5th Int. cocoa Res. Conf.
Ibadan, Nigeria 1975: 182-8, 1977
[4] Adenikinju, S.A., E.B. Esan and A.A. Adeyemi, Nursery Techniques,
Propagation and Management of Cacao, Kola, Coffee, Cashew and Tea
Progress in Tree Crop research (Second Edition) A Commemorative Book
to mark the 25th Anniversary of CRIN, 1989
[5] Ajayi and Okoruwa Managing Uncertainties and Risks in Cocoa
Production and Marketing in Nigeria, 1996
[6] Anon., Technical Consultation on Agricultural Methods and Techniques
for Cocoa. Report of a meeting held in Itabuna, Bahia, Brazil. Cocoa
Producers Alliance: Lagos, 1977
[7] Are, L., Methods of rehabilitating cocoa farms in Nigeria. Proc. 2nd Int.
Cocoa Res. Conf. Salvador, Brazil 1967: 383-7, 1969a
[8] Atanda, O.A. T. Quarcoo and M.O. Osundolire, Nursery techniques for
tree crops in Nigeria. In Progress in tree Crop Research in Nigeria. A
commemorative book, CRIN, 96-104, 1977
[9] Baiden, E. and A. Asare-Nyako (1986) Report of the Cacao Research
Institute of Ghana 1982/83–1983/84, 205, 1986
[10] Odegbaro, O.A., Prospects of rehabilitating Amelonado cocoa with
improved cocoa varieties in Nigeria without complete planting. Proc. 5th
Int. Cocoa Res. Conf. Ibadan, Nigeria 1975: 259-64, 1977
[11] Cocoa/Chocolate and Confectionary Alliance. Cocoa beans: Chocolate
Manufacturers quality requirements. The Cocoa, Chocolate and
Confectionary Alliance, 11 Green Street, London, 3rd Ed. Jan. 1984. 19pp,
1984
[12] Prior, C., Cocoa replanting practices and the possible dangers from root-
rots. Cocoa Industry Board of Papua New Guinea Newsletter 1, 2, 1981
Section 12
Vulnerability studies
Influence of ground water extraction in the

seismic hazard of Mexico City
J. Avilés1, L. E. Pérez-Rocha2 & H. R. Aguilar3
1
Instituto Mexicano de Tecnología del Agua, Mexico
2
Instituto de Investigaciones Electricas, Mexico
3
Centro de Investigación Sísmica, FJBS, Mexico
Abstract
The influence of ground water extraction in the seismic hazard of Mexico City is
examined. Available information on settlements of the land surface is used to
evaluate the subsidence effects on the predominant ground period. Microzoning
maps, as a function of this relevant site parameter, are constructed for the present
and future geotechnical conditions. Based on these maps, site-specific design
spectra throughout the whole city are determined by applying seismic code
provisions. It is found that the regional subsidence will generally be beneficial
for structures with fundamental period longer that the current period of the site,
but detrimental if the structure period is shorter than the site period.
Keywords: design spectra, ground water extraction, microzoning map, regional
subsidence, seismic hazard.
1 Introduction
Extensive ground-water exploitation has been identified as the main cause of
land subsidence in Mexico City. The phenomenon has caused large settlements
in the past, up to 10 m at some areas of the city. Subsidence of the city was
studied for the first time by Carrillo [1], showing a clear dependence with the
amount and rate of water extraction from artesian wells. Since then, many efforts
have been made to understand the phenomenon and mitigate its consequences. A
review of these efforts can be found in Ovando-Shelley et al. [2]. These authors
have studied the consolidation process in the central part of the city using the
well-known Terzaghi’s consolidation theory. They showed that exploitation of
the aquifers under the lacustrine clays is reflected not only on the gradual
doi:10.2495/GEO060451
reduction of their thickness, but also on the modification of their mechanical

properties.
In Mexico City, ground motion amplification during earthquakes is the most
important factor associated to the subsoil characteristics. The sedimentary basin
measures approximately 30×70 km and the thickness of compressible clay
deposits may exceed 70 m. With shear wave velocities ranging from around 50
to 100 m/s, the predominant ground period can be as long as 5 s. In view of these
geotechnical conditions, the resonant spectral acceleration for sites with
dominant period around 2 s may reach up to one g when intense subduction
earthquakes occur as far as 300 km. This was observed during the great 1985
Michoacan earthquake. Any variation in the present geotechnical conditions,
however, will result in a change in the seismic hazard in the future.
In this work, an empirical method to predict the effects of regional
subsidence on the seismic hazard in Mexico City for the coming years is
presented. Using available information regarding precise leveling of benchmarks
and geotechnical soil profiles, Aguilar et al. [3] have found correlations of the
subsidence rate and the thickness of sediments with the predominant ground
period. With the use of these data in a year-by-year incremental procedure, the
evolution of the microzoning map and the corresponding design spectra specified
by the building code, both in terms of the site period, is predicted for an exposure
period of 50 years. This would allow the designer to evaluate the effects of
regional subsidence during the life span of a structure.
2 Available information
The subsidence of Mexico City is studied here by using an empirical method
based on extrapolating available data for predicting future trends. With spectral
amplification functions for about 100 instrumented soft sites at Mexico City,
complemented with around 500 microtremor measurements, predominant ground
periods Ts were computed for a rectangular grid of 80x80 points covering most
part of the city, using the interpolation technique devised by Pérez-
Rocha et al [4]. This resulted in the microzoning map displayed in fig. 1, which
will be referred hereafter as the 2000 version. The isoperiod curves for Ts = 0.5
and 1 s roughly mark the separations between both the firm and transition zones
as well as the transition and soft zones, respectively.
For about 360 sites, the thickness of sediments H s is also known. Fig. 2 left-
hand side shows the correlation of this parameter with the site period obtained by
Aguilar et al. [3]. This relationship is expressed by
H s (m) = 31[Ts ( s ) − 0.5]1 2 for Ts ≥ 1 s (1)
Note that, for a given site, the shear wave velocity Vs is defined indirectly by
eqn. (1) since, according to the one-dimensional wave propagation theory, the
expression Ts = 4 H s Vs applies.
Figure 1: Curves of predominant ground period (s) in Mexico City.
For monitoring subsidence in Mexico City, a network of more than 2200

benchmarks distributed along the city has been installed. This network was
releveled in 1983 for the last time. By collecting elevation measurements at land
surface during the time period 1983-1998, Aguilar et al. [3] estimated the
amount of subsidence at soft sites. Then, the mean subsidence rate was
correlated with the site period by the expression
υ s (cm year ) = 2 [Ts ( s )]1.9 for Ts ≥ 1 s (2)
Fig. 2 right-hand side illustrates that the subsidence rate declines as the site
period shortens which in turn occurs when the depth of sediments shrinks (see
Fig. 2 left-hand side). This is consistent with the consolidation process in the
lakebed zone of the city. It should be mentioned that the observed subsidence
include not only the effect of ground-water extraction, but also the consolidation
from the own weight of buildings.
Figure 2: Left-hand side: thickness of compressible clay deposits in Mexico

City’s lakebed zone. Right-hand side: subsidence rate in Mexico
City’s lakebed zone.
3 Predicting procedure
Pumping operations in Mexico City have varied over the years, making difficult
the prediction of subsidence effects. Nevertheless, assuming the present pumping
conditions will be maintained in the near future, useful estimations may be made
as to changes expected in the existing microzoning map and the corresponding
design spectra. With the information that has been presented, a year-by-year
incremental procedure can be implemented for predicting the evolution of site
period, as follows:
● For a given site, the initial dominant period Ts1 is taken from fig. 1.
● By application of eqn. (1), calculate the initial thickness of sediments H 1s
corresponding to Ts1 .
Then,
i
● Calculate the subsidence rate υ s by use of eqn. (2).
● Calculate the amount of compaction as ∆H = υ is ∆t , which for ∆t = 1 year

represents the annual subsidence.
● Calculate the change in thickness of sediments as H si +1 = H si − ∆H .
● From eqn. (1), calculate the change in site period as Tsi +1 = ( H si +1 31) 2 + 0.5 .
● Replace i by i+1 and repeat the process for successive time steps until the
target exposure period is reached.
Let us now illustrate the application of the proposed procedure for estimating
the changes expected in the microzoning map of the city for the next 50 years.
Fig. 3 exhibits the modified isoperiod curves by the effects of regional
subsidence. The main variations are observed at the airport, Xochimilco and
Tlahuac areas, where long site periods undergo drastic reductions. In contrast,
the border between the transition and soft zones experiences little modification.
The evolution of site period can be appreciated in fig. 4 for sites with current
values of Ts = 1 to 5 s.
Figure 3: Modified isoperiod curves by the effects of regional subsidence for a

time period running from 2000 to 2050.
4 Evolution of seismic hazard

The Valley of Mexico is affected by earthquakes having different causes. They
have been divided in four groups, namely: 1) local earthquakes, 2) continental
plate earthquakes, 3) intermediate depth earthquakes and 4) subduction
earthquakes. For the contribution of all these events, the expected Fourier
amplitude spectrum at firm ground for a 125-year return period has been
estimated with the use of a standard probabilistic approach. This result is shown
in fig. 5.
Figure 4: Evolution of site period in Mexico City’s lakebed zone for a time
period running from the 2000 to 2050 year.
The site effects due to local soil conditions have been explicitly considered in
the Mexican Building Code [5] after the great 1985 Michoacan earthquake. The
resulting site-specific design spectra are defined by the following expressions:
1 + 3Te Ta
Sa g = c, if Te < Ta , (3)
4
S a g = c , if Ta ≤ Te ≤ Tb , (4)
r
T 
S a g =  b  c , if Te > Tb . (5)
 Te 
where S a is the spectral acceleration expressed as a fraction of gravity, r = 1 for
the lakebed zone of Mexico City, Ta and Tb are the lower and upper periods of
the flat part of the spectrum, respectively, and c is the seismic coefficient. The
flat part of the spectrum is specified by the limiting periods
Ta = max (0.35Ts ,0.64 s ), for Ts > 1 s , (6)
Tb = 1.2Ts . (7)
while the seismic coefficient is specified as
4Ts
c= (8)
4 + Ts2
These spectra are intended to cover not only the resonant peak response
associated with the first soil period, but that with the second period as well.
Figure 5: Fourier amplitude spectrum at firm ground in Mexico City for a 125-
year return period, including the effect of earthquakes from different
origins.
To illustrate the evolution of code design spectra, a characteristic site with

current value of Ts = 3.5 s was selected. The shortened site period resulting for
an exposure period of 50 years is seen from fig. 4. A comparison between the
2000 and 2050 design spectra is made in fig. 6. Also shown are the
corresponding response spectra obtained by using as input motion the Fourier
amplitude spectrum of fig. 5. The structural response was computed by applying
the random vibration theory [6] to a simplified model consisting of a one-story
structure placed on a uniform stratum under vertically incident shear waves,
corrected empirically to account for the bedrock flexibility [7]. Although the
representation is not perfect, the design spectrum intends to reproduce the
general trends observed in the response spectrum, for both present and future
geotechnical conditions. It is seen that regional subsidence will have either
favorable or unfavorable effects on the seismic safety of existing buildings,
depending primarily on the period ratio of structure and site.
The peak structural response can be conveniently represented by means of
spectral contours of acceleration expressed in terms of the structure and site
periods. In this way, the maximum response of any building at any location may
be readily estimated. We computed the 2000 and 2050 spectral contours, the
comparison of which is shown in fig. 7. They are constructed from site-specific
design spectra specified by the building code throughout the whole city. Each
site spectrum can be recovered by drawing a section along the site period of
interest.
Figure 6: Comparison between the 2000 (solid line) and 2050 (dashed line)
spectra for a site with presently dominant period Ts = 3.5 s; design
spectra (thick line) versus response spectra (thin line).
Figure 7: Comparison between the 2000 (thick line) and 2050 (thin line)
spectral contours of acceleration expressed as a fraction of gravity.
Looking at the current site spectra, notice that the plateau width is an
increasing function of the site period. This is to cover the influence not only of
the first mode of vibration of the soil, but of the second mode as well (see fig. 6).
In fact, the peak response associated to the latter may be as large as that
associated to the former for long site periods, say, Ts > 3.5 s. Also, it is apparent
that the plateau height increases with the site period for Ts < 2 s, but decreases
for Ts > 2 s. Thus, the most vulnerable buildings would be those with roughly 20
stories, assuming a fundamental period of 0.1 s per story. Nevertheless, this is
true for the time being but not for the coming years.
The difference between the 2000 and 2050 spectral contours reflects the
effects of regional subsidence. The 2050 spectral contours tend to rotate counter-
clockwise and elongate with respect to those for the 2000 year. Note that sites
with Ts > 2 s will migrate in the future to the region in which spectral ordinates
reach their peak values.
Figure 8: Subsidence impact on seismic base-shear coefficient: ratio of the

2050 to 2000 spectral contours given in fig. 7.
To have a whole scenario for the changes expected in the seismic hazard of
the city, the ratio α = S a ( 2050 ) S a ( 2000 ) was computed. Fig. 8 shows a
general view of the subsidence impact on base-shear coefficient. Both
detrimental ( α > 1) and beneficial ( α < 1) effects are observed, depending on the
period ratio of structure and site. For any structure and site configuration, the
subsidence impact may be assessed directly from this figure, entering with the
corresponding structure and site periods. It can be seen that, in general, regional
subsidence will affect the structural seismic safety adversely for Te < Ts and
positively for Te > Ts .
5 Conclusions
In this prospective study, the effects of regional subsidence on the seismic hazard
for Mexico City were examined. For an exposure period of 50 years, changes
expected in the existing microzoning map and the corresponding design spectra
specified by the building code were assessed. In view of complexity for making
an accurate prediction, only approximations were established about the
magnitude of subsidence effects. After some numerical evaluations, the
following main conclusions can be drawn:
a) Long site periods will undergo large reductions (e.g., from Ts = 5 to
≈ 3.25 s), whereas short site periods will undergo small reductions (e.g.,
from Ts = 1 to ≈ 0.95 s).
b) Sites with dominant period Ts > 2 s will migrate in the future to the region in
which spectral ordinates reach their peak values.
c) In general, regional subsidence will affect the structural seismic safety
adversely ( α > 1) for Te < Ts and positively ( α < 1) for Te > Ts .
The assessment of subsidence effects was made in an environment of high
uncertainty, where data may change rapidly from time to time due to local
pumping conditions. Although the results are location specific, the predicting
procedure may be applicable to other geographic locations facing similar
problems of land subsidence due to ground water extraction.
References
[1] Carrillo N., Influence of artesian wells in the sinking of Mexico City, Proc.
2nd Int. Conf. on Soil Mechanics and Foundation Engineering, Rotterdam,
Holland, 1948.
[2] Ovando-Shelley E., Romo M. P., Contreras N. and Giralt A., Effects on soil
properties of future settlements in downtown Mexico City due to ground
water extraction, Geofísica Internacional, 42, pp. 185-204, 2003.
[3] Aguilar H. R., Galicia M., Pérez-Rocha L. E., Avilés J., Vieitez L. and
Salazar M., Effect of regional subsidence on dynamic soil properties, Proc.
12th Panam. Conf. on Soil Mechanics and Geotechnical Engineering,
Boston, USA, 2003.
[4] Pérez-Rocha L. E., Ordaz M. and Sánchez-Sesma F. J., Spatial interpolation
of seismic data: the case of the Valley of Mexico, Proc. 10th Panam. Conf.
on Soil Mechanics and Foundation Engineering, Guadalajara, Mexico,
1995.
[5] MBC, Complementary Technical Norms for Earthquake Resistant Design,
Mexico Building Code, Federal District Government, 1987.
[6] Boore D. M. and Joyner W. B., A note on the use of random vibration
theory to predict peak amplitudes of transient signals, Bulletin of the
Seismological Society of America, 74, pp. 2035-2039, 1984.
[7] Avilés J. and Pérez-Rocha L. E., Site effects and soil-structure interaction in
the Valley of Mexico, Soil Dynamics and Earthquake Engineering, 17, pp.
29-39, 1998.
Predicting favourable areas for

landsliding through GIS modelling in
Aparados da Serra (Brazil)
A. J. Strieder1, S. A. Buffon1, T. F. P. de Quadros2
& H. R. Oliveira3
1
Lab. Modelagem Geológica e Ambiental, MODELAGE-UFRGS, Brazil
2
Fund. Estadual de Proteção Ambiental, FEPAM – SEMA-RS, Brazil
3
Transportadora Gasoduto Bolívia-Brasil S.A., TBG-Sul
Abstract
This paper presents the results of GIS modelling for predicting areas where
natural landslides can occur in the Aparados da Serra region (Southern Brazil).
The scarp is mainly developed upon basalts and dacites (Serra Geral Fm.) and
sandstones (Botucatu Fm.) of the Paraná Basin. The 1000 m high scarp has been
geomorphologically evolving since the break-up of the Gondwana
Supercontinent and the opening of the South Atlantic Ocean. Geologic,
geomorphologic and geotechnical mapping were conducted in order to acquire
field data and to define a conceptual geomorphologic-geotechnical model for
GIS data modelling. These mappings were aided with ASTER image processing
and aerial photograph analysis. ASTER images do permit a high resolution and
accurate DEM. The prediction of areas able to develop natural landslides along
the scarp was based upon Factor of Safety (FS) algorithms. The GIS modelling
results for Aparados da Serra region were grouped into four categories. The areas
with the lowermost FS number were defined as highly susceptible to developing
natural landslides. Field investigations upon some existing landslide structures in
the region do confirm such predictions. However, large scale FS determination
through GIS modelling in order to predict susceptible areas for landslides must
be interpreted in a relative manner. FS modelling can, then, be applied to
landslide susceptibility mapping in areas of poor historical records, since
supported by adequate geological and geotechnical investigations.
Keywords: landslides, susceptibility maps, GIS modelling, factor of safety.
doi:10.2495/GEO060461
1 Introduction
Landslides are a natural phenomenon which include a wide variety of material,
such as debris, soils, rocks, organic matter, and constructions, moving down-
slope [1]. Landslides describe many types of down-slope mass movement,
ranging from rapidly moving catastrophic rock avalanches and debris flows, to
slowly moving earth slides.
Landslides concerned here relate to slow rotational and translational slides of
debris and soils [2]. The gravity is the main driving force for these slides, but a
number of factors can influence them. The increase in water content, for
example, contributes to loading of the debris and soil materials in the slope. In
the same way, the shear strength of the slope material is decreased while the pore
water pressure is increased. This seems to be the second main driving force for
landslides in tropical countries, such as Aparados da Serra region in Southern
Brazil.
Landslide susceptibility maps are produced by a number of methods by
means of GIS. Quantitative methods may include inventory, slope angle, terrain
aspect, geology, vegetation, land use maps overlaid through the named landslide
susceptibility matrix technique [3]. Multivariate analysis of factors controlling
landslides are also developed [4]. GIS development enhanced landslide
susceptibility analysis in different ways (e.g. [5,6,7]). It seems that an inventory
of historical data is a key factor in landslide susceptibility map construction.
However, how can a landslide susceptibility map be assessed in areas displaying
poor landslides records?
This paper aims to present a landslide susceptibility map for Aparados da
Serra area (Southern Brazil, fig. 1) based on a rational deterministic approach [8]
through GIS modelling. The prediction of areas able to develop natural
landslides along the Aparados da Serra scarp was based upon Factor of Safety
(FS) algorithms. The factor of safety (FS) was determined for high resolution
areas (15 m) according the discussion by Wu and Abdel-Latif [9]. The results
were tested looking at some existing, shallow landslides at Aparados da Serra
scarp.
2 Geological and geotechnical investigations: a landslide

model for Aparados da Serra (Southern Brazil)
The Aparados da Serra region (Southern Brazil) shows a 1000 m high scarp
developed since the break-up of the Gondwana Supercontinent and the opening
of the South Atlantic Ocean. It mainly exposes Botucatu Fm. (sandstones) at the
base, and basalts and dacites of Serra Geral Fm. up-ward. The base of scarps has
thick Tertiary talus deposits, cemented by iron oxides and hydroxides.
ASTER image processing through AsterDTM tool of ENVI software (RSI) do
permit a high resolution (15 m) and accurate digital terrain model. The
coordinate system was set to SAD69/96. ASTER and LANDSAT images
processing, DTM, aerial photographs analysis and fieldworks were performed to
produce the geologic, geomorphologic and geotechnical maps. These data were
all introduced in ArcGIS software (ESRI) for additional modeling.
Figure 1: Regional map locating Aparados da Serra region, Southern Brazil.
Geological mapping could distinguish 29 volcanic flows composing the Serra

Geral Fm. 15 of these flows are made of basic volcanic rocks (basalts), while 14
of them are acid volcanic rocks (dacites). The limit of basic and acid volcanic
rocks is set at 800 m of altitude. The lower contact of Serra Geral volcanic rocks
is close to 245 m of altitude.
Geomorphologic mapping defined a step-platform structure for Aparados da
Serra scarp (Fig. 2). Under these structural conditions, recent weathering and
mass wasting processes accumulated debris and soils as triangular prisms of
colluviums over the platforms. The soils are mainly composed of clays, and
some of them show latteritic features.
Geotechnical and geomorphologic mapping recognized a vertical
differentiation for colluviums prisms. In the higher scarps (> 800 m of altitude),
the step height is close to or greater than platform length (Fig. 2). On the other
hand, in the middle and lower part of the Aparados da Serra scarp (< 800 m of
altitude), the platform length is much greater than step height. It was also noticed
that soils have low thickness (< 5 m) in the upper scarp, while they show up to
30-40 m in some places of the lower scarp. These differentiations explain why
some block falls are reported in the upper scarp, and why some shallow,
translational slides are reported in the lower scarp.
The shallow (1-3 m depth), translational slides in the Aparados da Serra

region (Southern, Brazil) shows clear elliptical shapes. They wide vary from 8 to
25 m, and the length is greater than 25-30 m. Laboratory geomechanic essays of
colluviums in the Aparados da Serra region were done by Silveira [10] (Table 1).
Figure 2: Geologic-geomorphologic structure of the Aparados da Serra region

(Southern Brazil).
Table 1: Geomechanic properties of colluviums in the Aparados da Serra

region (Southern Brazil).
Geomechanic parameters Mean Minimum Maximum

ρw (water density) 9,806 9,806 9,806
ρs (natural soil density) 18 15,5 20,4
ρw/ρs ratio 0,544778 0,632645 0,480686
ρt (saturated soil density) 28,15 27,7 28,6
ρ (dry soil density) 11,7 10,5 12,9
Φ peak (peak shear strength) 29,5 28 31
Φ residual (residual shear strength) 16 10 18
β (slope) 20 2 70
tanΦpeak/tanβ ratio 1,554448 15,22617 0,218695
tanΦres/tanβ ratio 0,787826 5,049344 0,118261
Triaxial cohesion 18 12,5 25
Undeformed direct shear cohesion 27 25 28
3 Predicting favourable areas for landslide through FS

Predicting favourable areas for landslides through FS determination in GIS
environment requires an evaluation on existing landslide dimensions. In this
way, the reported landslide in the Aparados da Serra region covers 1-2 pixels in
wide and 3-5 pixels in length. The dimensions of reported landslides enable
classification procedure through filtering in ENVI software.
The main variables to be considered in mathematical modelling are:
- W (mean wide of landslide structures);
- L (mean length of landslide structures);
- D (colluviums thickness);
- H (water table height);
- φ (effective shear strength),
- β (slope).
These characteristics enable some simplification in FS computing through GIS
modelling. Mean W and L for reported landslide structures are greater than pixel
resolution (15x15 m). Colluviums thickness can be computed from DTM pixel
and lower volcanic flow platform. The computed colluviums thickness map
shows lateral and longitudinal slope variations. The deterministic FS model can,
in this way, consider that forces acting parallel to slope are in equilibrium and
that the slope is infinite. These conditions can be applied chiefly to middle and
lower scarps of the Aparados da Serra region.
The FS determination can take the equation presented by Haneberg [8]:
FS = {1 – [(ρw / ρs) . (H / D)]} . (tan φ / tan ß) (1)
However, equation 1 applies to low cohesion, unconsolidated materials. A more
realistic FS determination must then consider cohesion of colluviums deposited
over the volcanic basement. The equation presented by Wu and Abdel-Latif [9]
takes cohesion into account and was developed for planar failure and infinite
slope:
FSp = {cf + c’ + [ρ(D – H) + (ρt – ρw)H] cos2ß tan φ}
/ {[ρ(D – H) + ρtH] sin ß cos ß} (2)
where c’ is cohesion and cf is cohesion increase for forested areas.
The FS is clearly dependent on variations of physical properties. But, water
table is known to influence slope failure. The water table is not parallel to
groundsurface and varies according thickness, composition and permeability of
soils. It is also dependent on precipitation: it is well known that slope failure
occur during on immediately after raining periods, when water table is high
enough to decrease colluviums shear strength. Taking into account that water
table (piezometers) close to a monitored landslide in Aparados da Serra
(Southern Brazil) is 1-2 m below the groundsurface during dry periods. Then, it
is possible to consider H = D (conservative approach) and determine FS for
extreme situations.
Cohesion increase due to vegetation cover was not considered in this FS
determination. The dense Atlantic forest cover surely contributes to cohesion and
to FS value. The influence of Atlantic forest in cohesion increase for colluviums
is being investigated. It is interesting to note, in this respect, that shallow
landslides reported in Aparados da Serra region occur in deforested areas. Then,

to partially overcome this uncertainty, it was taken into account undeformed
direct shear cohesion in FS determination.
4 Landslide susceptibility map for Aparados da Serra region

The FS is the ratio between resisting to driving forces acting in slope failure. It is
usually said that FS > 1 represents stable conditions, while FS < 1 represents
unstable conditions. A FS ~1.25 or somewhat higher is generally accepted for
slope stability. However, this determination is valid for a specific slope, where
all the geomechanic parameters are known, controlled and modeled.
Table 2 shows that FS is very sensitive to slope angle and colluviums
thickness. The colluviums in the Aparados da Serra region show small variations
in cohesion, shear strength, and densities when regional analysis is performed.
However, FS rapidly decrease to values lower than 1 when slope angle goes to
25 (D = 5), and to 17 (D = 10). It is also realized that FS decreased 1/3 as
colluviums thickness increased by 2.
Table 2: Sensitivity analysis of Aparados da Serra colluviums considering FS

equation presented by Wu and Abdel-Latif [9]. FS was calculated for
different slope angles and two colluviums thickness (D). Water table
was fixed to be 1 m below the groundsurface.
D=5 D = 10
β H=4 H=9
FSp5 FSp10
2 12,02876 8,528288
4 6,018449 4,264291
6 4,016854 2,843039
8 3,01748 2,132492
10 2,419042 1,706247
12 2,021122 1,422175
16 1,526263 1,067379
20 1,232421 0,854974
24 1,039572 0,713977
28 0,905004 0,614038
32 0,807534 0,540058
40 0,682551 0,44025
50 0,611113 0,371193
60 0,618595 0,348311
The determination of FS in large regions using a GIS environment must

overcome large scale parameters variability. This introduces uncertainties in GIS
modelling. The FS values determined in GIS environment varies from 0.2 to up
to 200. Taking into account the sensibility of FS equation (2) for geomechanic
physical parameters, the predicted map of FS was computed for two different
cases: i) residual shear strength (φ = 16o), that is the worst condition attained
during and after raining periods, and ii) peak shear strength (φ = 29o). The
predicted FS maps are presented in figure 3A and 3B, respectively. Based on the
locations of some recorded shallow landslides in the Aparados da Serra region
(Southern Brazil), it was possible to define five classes of susceptibility based in
FS values (Table 3).
Figure 3: Landslide susceptibility map for Aparados da Serra region (Southern

Brazil), based in FS equation presented by Wu & Abdel-Latif [9]. A)
Predicted FS classes using residual shear strength (φ = 16o). B)
Predicted FS classes using peak shear strength (φ = 29o).
Table 3: Susceptibility classes for landslides in the Aparados da Serra region

(Southern Brazil), based in FS equation presented by Wu & Abdel-
Latif [9]. Susceptibility classes defined for residual shear strength (φ
= 16o) and for peak shear strength (φ = 29o).
Predicted FS Susceptibility FSp

class
Areas highly favourable for landslides 1 < 0.5
Areas favourable for landslides 2 > 0.5 < 1.0
Areas that can develop landslides 3 > 1.0 < 1.5
Areas not favourable for landslides 4 > 1.5 < 2.5
Stable areas under natural conditions 5 > 2.5
5 Discussion
The Factor of Safety determination using GIS modelling can be a useful
technique for assessing landslide susceptibility map in areas of poor historical
records on hillslopes of tropical regions, such as Aparados da Serra (Southern
Brazil). The methods based on intense data records seek on statistical and
multivariate capabilities for an adequate susceptibility map assess in Aparados
da Serra region. In this way, careful geological, geotechnical and
geomorphologic investigations in existing records were applied to assess the
main causes of shallow, translational and rotational landslides in Aparados da
Serra region. The geological and geotechnical model for Aparados da Serra
region, the accurate and high resolution DTM (15x15 m), and the GIS
capabilities do enable FS determination.
The FS determination, however, introduces uncertainties in GIS modelling
due to spatial variability of parameters. The value determined for each cell
(15x15 m pixel) must be evaluated in a relative way. Then, a careful examination
of existing records surely helps on establishing the limits between different
classes. It is to be kept in mind that absolute FS value is an important parameter
for construction decision in a given, specific investigation site, where a large
number of geotechnical soil properties are well controlled.
Landslide susceptibility map derived form FS determination in Aparados da
Serra (Southern Brazil) helps, in this way, in showing “where” the areas more
favourable to develop translational and rotational slides are. Ongoing
investigation on Atlantic forest contribution on soils and colluviums cohesion
can put some clues on questions regarding “when” and “how big” these
landslides can be.
Acknowledgements
The authors want to thank FINEP, PETROBRAS and TBG (Proc. No. 0682/01)
for research funds that supported this investigation.
References
[1] Cruden, D.M., A simple definition of a landslide, Bull. IAEG, 43, pp. 27-
29, 1991.
[2] Cruden, D.M. & Varnes, D.J., Landslide Types and Processes. Landslides
Investigation and Mitigation, A.K. Turner & R.L. Shuster Eds..
Transportation Research Board, Special Report 247, Washington DC,
p.36-75, 1996.
[3] DeGraff, J.V. & Romesburg, H.C., Regional landslide-susceptibility
assessment for wildland management: a matrix approach. Thresholds in
Geomorphology, D.R. Coates & J. Vitek Eds., George Alien & Unwin
Publ., pp. 401-414, 1980.
[4] Carrara, A., Multivariate models for landslide hazard evaluation,
Mathematical Geology, 15(3), 403–426 pp., 1983
[5] Carrara, A., Cardinali, M., Guzzetti, F., & Reichenbach, P., GIS
technology in mapping landslide hazard, Geographical Information
Systems in Assessing Natural Hazards, A. Carrara & F. Guzzetti Eds,
Kluwer Academic Publishers, pp. 135–175, 1995.
[6] Santacana, N., Baeza, B., Corominas, J., Paz, A. & Marturia, J., A GIS-
based multivariate statistical analysis for shallow landslide susceptibility
mapping in La Pobla de Lillet area (Eastern Pyrenees, Spain), Natural
Hazards 30: 281–295, 2003.
[7] Tangestani, M.H., Landslide susceptibility mapping using the fuzzy
gamma approach in a GIS, Kakan catchment area, southwest Iran,
Australian Journal of Earth Sciences, 51(3), pp. 439-450, 2004.
[8] Haneberg, W.C., Deterministic and probabilistic approaches to geologic
harzad assessment, Environmental & Engineering Geoscience, 6(3), pp.
209-226, 2000.
[9] Wu, T.H. & Abdel-Latif, M.A., Prediction and mapping of landslide
hazard, Can. Geotech. J., 37, pp. 781-795, 2000.
[10] Silveira, R.M., Propriedades geotécnicas dos solos coluvionares do
gasoduto Bolívia-Brasil em Timbé do Sul (SC), Dissertação de Mestrado
em Engenharia Civil, Escola de Engenharia/UFRGS, Porto Alegre, 131
pp, 2003
Author Index
Abad M. P................................ 317 Frey T. ....................................... 69
Aguilar H. R. ........................... 457 Fyfe W. S..................................... 1
Aguilar M. ............................... 317
Alcalá L. .................................. 317 García-Hidalgo J. F.................. 317
Amusan F. O............................ 443 García-Quintana A................... 317
Amusan O. A. .......................... 443 Gecchele G. ............................. 349
Aunap R................................... 305 Giles J. R. A............................. 187
Avilés J. ................................... 457 Giménez R. .............................. 317
Godfrey A. E. .......................... 317
Badiani B. ................................ 341 González-Martín J. A. ............. 317
Bahena A. ................................ 263 Guida D. .................................. 283
Barragán P. de L. G. ................ 401 Gutiérrez A. ............................. 263
Barrera I................................... 317
Bautista M. R. .......................... 135 Hall P. J. .................................. 151
Beneš V.................................... 361 Hernandez M. V. ..................... 401
Boukalová Z. ........................... 361 Holguin C. ............................... 135
Bridge D. McC......................... 187 Hsiao L. S. ............................... 105
Brissette F. ............................... 273 Huang C. Y.............................. 105
Broderick M. A.......................... 15 Huang C.-P. ............................. 143
Buehler Y. A............................ 391
Buffon S. A.............................. 467 Itten K. I................................... 391
Capra A.................................... 175 Jafari M.................................... 431

Carrère V. ................................ 381 Jiménez J. ................................ 135
Cebrián M. ............................... 317
Cecioni A..................................... 7 Kaikis M. ................................. 253
Champagne P. .................. 115, 125 Kask P........................................ 69
Chang J.-H. .............................. 143 Kauer K. .................................. 413
Cheng S. F. .............................. 105 Kellenberger T. W. .................. 391
Chopra M................................. 215 Kelly J...................................... 115
Clerico M................................... 33 Khalekuzzaman M................... 125
Cremonini R. ........................... 293 Khalili M. ................................ 273
Czarnecki A. ............................ 241 Kõiv M....................................... 93
Kõlli R. .................................... 413
D’Alessandro W. ..................... 369 Koodhathinkal B...................... 215
de Quadros T. F. P. .................. 467 Kořán P.................................... 361
Durning B. ................................. 15 Köster T. .................................. 413
Kriipsalu M................................ 93
Emmanouloudis D. .................. 253
Lafragua J. ............................... 263
Fernández de Villalta M. C...... 317 Lahdenperä J............................ 423
Férnández S. ............................ 263 Leconte R................................. 273
Frey J. ........................................ 69 Lehtinen K. ................................ 25
Lehto K. ................................... 423 Rabuffetti D. ............................ 293

Lewandowska-Czarnecka A. ... 241 Ranke U. .................................... 43
Li P. P. ..................................... 225 Reinhart D. .............................. 215
Liu C. Q. .................................. 225 Rodríguez-Borreguero J. M. .... 317
Longobardi A........................... 283 Roosaare J................................ 305
Lucía A. ................................... 317 Rubio A. H............................... 401
Luukkanen O. ............................ 55 Rubio H. O............................... 135
Ruiz López de la Cova R. ........ 317
Maher A................................... 431
Mander Ü..................... 69, 93, 305 Salandin A. .............................. 293
Martín-Duque J. F.................... 317 Sarala P...................................... 25
Martín-Loeches M. .................. 317 Sarmad A. ................................ 431
Michel F................................... 115 Saucedo T. R. .......................... 135
Mirzaeian M. ........................... 151 Scicolone B.............................. 175
Möller A. ................................... 43 Small D.................................... 391
Moriel G. ................................. 263 Soffredini G. .............................. 33
Morris B. L. ............................. 187 Solís A. .................................... 317
Stone Jr B. ............................... 195
Negreiros J................................. 79 Strieder A. J. ............................ 467
Nieminen J................................. 55
Nishihama Y. ........................... 231 Tamura R. ................................ 231
Noormets M. ............................ 413 Tena V. M................................ 401
Tokola T. ................................... 55
Oder M..................................... 413 Tõnutare T. .............................. 413
Oliveira H. R............................ 467 Treier K. .................................... 69
Oliver T...................................... 55 Tsouchlaraki A. ....................... 329
Ortega-Guerrero M. A. ............ 205
Uuemaa E. ............................... 305
Paal T....................................... 413
Painho M.................................... 79 Vajirkar M. .............................. 215
Pajuste K.................................... 69 Velez C. ................................... 401
Pérez-Rocha L. E. .................... 457 Villani P................................... 283
Pineda V. ..................................... 7
Pinedo A. A. ............................ 401 Wang B. L. .............................. 225
Pinedo A. C.............................. 401 Williams D. D.......................... 163
Pizzo G. ................................... 349 Wilson K. P. ............................ 163
Porto P. .................................... 175 Wood K. .................................. 135
Wu Y. Y................................... 225
Qiang Z. ................................... 143
Quijada E. ................................ 317 Yanai S. ................................... 231
Zhou P. ...................................... 55
Environmental Toxicology Modelling, Monitoring and
Edited by: A. G. KUNGOLOS, University of
Thessaly, Greece, C. A. BREBBIA, Wessex Management of Air
Institute of Technology, UK,
C. P. SAMARAS, TEI of West Macedonia,
Pollution XIV
Greece, V. POPOV, Wessex Institute of Edited by: C. A. BREBBIA, Wessex Institute
Technology, UK of Technology, UK
This book addresses the need for the Pollution is widespread throughout the world
exchange of scientific information among and elimination of the risks to human health
experts on issues related to environmental is of the utmost importance. This book
toxicology, toxicity assessment and examines the development of experimental
hazardous waste management. Publishing as well as computational techniques to achieve
papers from the First International a better understanding of air pollution
Conference on Environmental Toxicology, problems and seek their solution. The book
the text will be of interest to biologists, brings together work from researchers who
environmental engineers, chemists, are active in the study of air contaminants
environmental scientists, microbiologists, and contains papers from the Fourteenth
medical doctors and all academics, Annual International Conference on the
professionals, policy makers and Modelling, Monitoring and Management of
practitioners involved in the wide range of Air Pollution. Papers featuring case studies
disciplines associated with environmental particularly those discussing the evaluation
toxicology and hazardous waste management. of proposed emission techniques and strategies
The text encompasses themes such as: Acute are featured, as well as papers of a more
and Chronic Bioassays; Tests for Endocrine theoretical nature dealing with advanced
Disruptors and DNA Damage; Interactive mathematical and computational methods.
Effects of Chemicals; Bioaccumulation of
Chemicals; Assessment of Ecotoxicological ISBN: 1-84564-165-5 2006 apx 450pp
Properties of Hazardous Wastes; Hazardous apx £165.00/US$295.00/€247.50
Waste Management Techniques; Legislation
Regarding Environmental Effects of
Chemicals; Hazardous Waste Reduction and
Recycling Techniques; Biodegradation and
Bioremediation; Monitoring of Hazardous
Waste Environmental Effects; Laboratory
Techniques and Field Validation; Effluent
Toxicity, Microbiotests; On-line Toxicity
Monitoring; Forensic Toxicology;
Genotoxicity/Mutagenicity; Exposure
Pathways; Risk Assessment; Biotesting and
Environmental Control Strategy; Hot Spots WITPress
and Accidental Spills. Ashurst Lodge, Ashurst, Southampton,
ISBN: 1-84564-045-4 2006 apx 400pp SO40 7AA, UK.
apx £145.00/US$265.00/€217.50 Tel: 44 (0) 238 029 3223
Fax: 44 (0) 238 029 2853
E-Mail: witpress@witpress.com
Water Pollution VIII Water Resources
Modelling, Monitoring and Management III
Edited by: M. DA CONCEICAO CUNHA,
Management of Water Pollution ISEC, Portugal, C. A. BREBBIA, Wessex
Edited by: C. A. BREBBIA, Wessex Institute of Technology, UK
Institute of Technology, UK,
J. S. ANTUNES DO CARMO, University Today water resources are under extreme
of Coimbra, Portugal pressure all over the world. This book brings
together the work of researchers and
This book publishes the proceedings of the professionals currently involved in the
Eighth International Conference on development of technical and scientific skills
Modelling, Measuring and Prediction of that contribute to achieving water resources
Water Pollution. Water pollution is a subject sustainability. Over 65 contributions
of growing public concern. The scientific originally presented at the Third
community has responded very rapidly to International Conference on Water
the need for studies capable of relating the Resources Management are included.
pollutant discharge with changes in the water The wide variety of subjects covered are as
quality. The results of these studies are follows: Hydrological Modelling;
permitting industries to employ more Groundwater Flow and Remediation;
efficient methods of controlling and Reservoirs and Lakes; Coastal and Estuarial
treating waste loads, and water authorities Problems; Water Quality; Pollution
to enforce stricter regulations regarding this Control; River Basin Management; Water
matter. Bringing together papers from world Management and Planning; Wastewater
renowned experts in this field, the text Treatment and Management; Irrigation
encompasses themes such as: Groundwater Problems; Residential Water Management;
and Aquifer Contamination; Wastewater Sustainable Water Use; and Arid Region
Treatment; Re-use of Water; Lakes, Rivers Problems.
and Wetlands; Coastal Areas and Seas;
Biological Effects; Agricultural Pollution; Series: Progress in Water Resources
Oil Spills; Mathematical and Physical Vol 11
Modeling; Experimental and Laboratory
Work; Surveying Techniques, Monitoring ISBN: 1-84564-007-1 2005 712pp
and Remote Sensing; Remediation Studies; £249.00/US$398.00/€373.50
Health Risk; Social and Economic Issues;
Pollution Prevention; GIS and Remote
Sensing Applications; Environmental
Management and Decision Analysis; All prices correct at time of going to press
Environmental Impact Assessment. but subject to change.
WIT Press books are available through
ISBN: 1-84564-042-X 2006 apx 400pp your bookseller or direct from
apx £145.00/US$265.00/€217.50 the publisher.

GEO Environment and Landscape Evolution II PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

GEO Environment and Landscape Evolution II PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Geo-Environment and

WIT Press publishes leading books in Science and Technology.

WIT Transactions on Ecology and the Environment

D Almorza Gomar A M Amer

J-L Uso R van Duin

Geo-Environment & Landscape Evolution 2006

INTERNATIONAL SCIENTIFIC ADVISORY COMMITTEE

For USA, Canada and Mexico

Computational Mechanics Inc

British Library Cataloguing-in-Publication Data

A Catalogue record for this book is available

The texts of the papers in this volume were set

© WIT Press 2006

Printed in Great Britain by Cambridge Printing.

Section 1: Environmental planning and management

Ethics, geological risks, politics and society

Environmental impact assessment and environmental management

Geological factors in monitoring and planning nature trails at

The acoustical improvement plan as a process to re-establish

Estimation of the on-farm-costs of soil erosion in

Section 2: Environmental modelling and monitoring

Large scale soil erosion modeling for a mountainous watershed

SAKWeb© – Spatial Autocorrelation and Kriging Web Service

Section 3: Environmental pollution and remediation

After treatment of landfill leachate in peat filters

Studies on the distribution of heavy metal Cd in contaminated soils

Mitigation of alkaline mine drainage in a natural wetland system

A sequential aerated peat biofilter system for the treatment of

The sorption characteristics of PAHs onto soils in the presence of

Carbon dioxide sequestration in coal: implications for CO2 disposal

Section 4: Climatological processes

Response of dissolved organic carbon in a shallow groundwater

Section 5: Geo-environment in urban settings

The urban geo-science model: an essential tool to support planning

Developing design-oriented strategies to combat regional scale

Land subsidence evolution and controlling mechanisms near

Stability of slopes of municipal solid waste landfills with

Net photosynthetic O2 evolution and calcium precipitation in

Dynamics of suspended sediment concentration and the impact on

Potential of poplar plantation for enhancing Polish farm

Section 7: Hydrological studies

Modified quantitative estimation model of erosion and degradation

Efficient watershed modeling using a multi-site weather generator

Hydrological modelling for river basin management in a highly

Hydrological modelling of snow cover in the large upper Po river

Section 8: Landscape analysis

Spatial correlograms and landscape metrics as indicators of land

The role of geosciences and landscape in the management of

Exploring the effect of demographic elements on the evaluation

An approach to the landscape analysis

Evaluation and analysis criteria of the environmental risk factor

M3 (Monitoring, Management and Mapping) – tool for the solution

Gas hazard: an often neglected natural risk in volcanic areas

Section 10: Remote sensing

Identification of As-bearing minerals associated with mine wastes

Rapid mapping with remote sensing data during flooding 2005 in

Using spatial technology for analyzing disturbed areas and potential

Section 11: Soil and rock properties

The pedoecologial conditions of natural and opencast peat fields

Monitoring programme for underground rock characterization

Geotechnical evaluation of Stabilized Dredged Material (SDM)

The impact of soil quality on cocoa yield in Nigeria

Influence of ground water extraction in the seismic hazard of