Beruflich Dokumente
Kultur Dokumente
PHYSIOGNOMIC
LANDSCAPE
45
ABSTRACT
De Veer, A.A. and Burrough, P.A., 1978. Physiognomic
Netherlands. Landscape Plann., 5: 45-62.
Although physiognomic landscape studies have been used for 10 years by planners in
The Netherlands, there have been few comparisons of the various methodologies. Also,
publication of the methods in English has been rare. A brief description of the most
important visible aspects (physiognomy) of Dutch landscapes is followed by a review of
techniques used from 1966-1976
for classifying and mapping landscapes for planning
purposes, Three main approaches have been recognized - compartment; breadth of
view; and grid cell. The findings of a recent study comparing the results and techniques
of different Dutch landscape classification methods are summarized and discussed in the
light of the stated requirements of users of landscape data. As no single extant technique
satisfies all the users requirements, future developments in landscape mapping should
adopt a more flexible approach. Recent experience in using computer-assisted mapping
methods in soil science and geology suggests that these techniques can be adapted profitably to classify and map landscapes from elementary data according to users needs.
INTRODUCTION
46
OF DUTCH LANDSCAPES
Dutch landscapes are the product of a long and frequently bitter struggle
between people and their environment
(Lambert,
1971). Two broad landscape regions can be differentiated.
The flat, northwestern
half of the country, lying mainly below sea level,
consists of Holocene peat and clay deposits. A dune belt along the North Sea
coast protects these areas, but gaps in the dunes and high river-water levels
inland call for an extensive system of protecting dykes and drainage. In these
areas, the predominant
grassland and arable land of the polders contrasts
strongly with the dense urban concentrations
of the main cities that have
developed along the major water courses. As the main parcel boundaries on
the polders are ditches rather than hedges, the non-urban landscapes in the
northwest are very open.
The southeastern
half of the country is formed mainly from Pleistocene
glacial ridges and wind-blown coversands. The low to undulating relief and
the variations in soil fertility have induced a highly differentiated
land-use
system that includes mixed farming and heath and forest areas. In earlier
times the fields were surrounded by dense hedges to keep cattle and wild
animals from the arable land, causing very closed landscape types to develop.
The present landscapes of heath and forests owe much to the old agricultural
system. Removing sods from the heaths for mixing with farmyard manure
and overgrazing by sheep allowed severe erosion of the poor coversand soil.
This was first checked by oak plantation around small settlements,
and in the
nineteenth and twentieth centuries by concerted programmes of tree planting,
predominantly
pines.
The Dutch landscapes are still evolving. The open landscapes in the northwest were extended greatly by the recent IJsselmeer polder reclamations
and
were also significantly
affected by the major floods of 1953 when 1650 km*
were inundated (Huson and Verstappen,
196.8).
Many landscape changes have resulted from the pressures of a larger, more
mobile and more demanding populatiol~. Owing to increased population
after the second World War, much agricultural land has been changed into
urban or sub-urban sprawl; increased mobility has required new roads and
motorways.
Both have reduced considerably
the area of nonurban land, which
also has been broken into smaller units. In addition, the high-rise developments on the edge of urban areas, particularly in the large open landscapes in
the west of the country, have increased the visible impact of urbanization
(Nicolai, 1971; Ministry of Housing and Physical Planning, 1976).
At present, two major laws govern land use reallocation.
The older law
47
concerning reallottment
(ruilverkaveling)
allows agricultural economic improvements to be made by revising the patterns of drainage, field size, land
tenure and settlement, all of which have profound effects on the visible and
ecological qualities of the landscape. The second, the Physical Planning Bill,
enacted in 1962 (Brussaard, undated), regulates land use designation at local,
regional and national levels. One of the reasons for its introduction
was the
increasing demands on a~icultur~
lands to serve multiple uses such as nature
conservation or recreation and their consequent effects on the landscape.
For developments
planned under either law, planners are requesting more
landscape data to guide their decisions.
LANDSCAPE
AND LANDSCAPE
MAPPING
IN RELATION
TO PLANNING
mapping
in The Netherlands
and abroad
48
mapping (e.g., along the lines of some methods described by Linton, 1968;
Jacobs and Way, 1969; Kiemstedt,
1969; Steinitz et al., 1969; Dunn, 1974).
A more fundamental reason, however, may be the predominantly
flat, agricultural character of the Dutch landscape, which contrasts so strongly with
the urbanized land. Differences
in landscape character depend strongly on
variations in space, and space is circumscribed
by elements such as lines
of trees, dykes or buildings. These elements form dominant criteria for landscape classification.
In other countries, differences in relief, outlook and landform play much greater roles in the visible landscape than in The Netherlanc!s.
Consequently,
classifications
adopted in other countries frequently make
more use of ecological or integrated survey techniques than is possible in
The Netherlands (e.g., Klink, 1966; Haase, 1967; Steinitz et al., 1969; Leser,
1976). Most Dutch physiognomic
landscape mapping methods developed so
far stop short of actual aesthetic evaluation and concentrate
first on understanding the spatial character and content of the landscapes.
of landscape mapping
All the Dutch landscape mapping methods discussed in this paper concentrate on describing and classifying the physiognomy.
There have been other,
national, surveys covering various other aspects of the environment;
for
example, soil and vegetation relations (Kalkhoven et al., 1976), or maps
showing the role of people in landscape building by depicting parcellation,
historical sites and land-use patterns (Ministerie van CRM, 1977). Both these
surveys have adopted a standard approach to mapping the whole country.
However, in physiognomic
landscape survey, several techniques have been
developed by geographers and landscape architects to meet a wide range of
planning problems on different scales.
The first attempt to classify the Dutch landscape in physiognomic
terms
was a small scale map (1:500
000) called The landscape of The Netherlands
and bordering regions (Rijksplanologische
Dienst, 1966). This map divided
the agricultural land into abstract classes defined as complexes of open spaces.
Open landscapes were defined as those where both the length and breadth of
spaces with unrestricted
vision exceeded 1500 m. For half-open landscapes,
the critical values were 1500 m and 500 m, respectively;
closed landscapes
comprised spaces where both dimensions were less than 500 m. One aim of
the map was to show that the Rand&ad Holland - the urbanized horseshoe
in the west of the country through Utrecht, Amsterdam, Haarlem, Den Haag,
Rotterdam
and Dordrecht - has an open green heart that should be preserved.
This map was the forerunner of several different approaches to physiognomic landscape mapping (De Veer, 1977). However, after several years
there was more diversity than concensus about the meaning of the visible
landscape and how it should be mapped, so more research was recommended
(Rijksplanologische
Dienst, 1974). The Netherlands Soil Survey Institute was
asked to compare the various Dutch landscape mapping methodologies
and
49
Author(s)
Type of method
(1) Compartment
Compartment
Compartment
Compartment
Compartment
Breadth of view
Grid and breadth of view
Grid and compartment
Special purpose
.- .-...~_.
-__
methods
The compartment
approach has been used most frequently.
A space is
defined as an area of the earths surface bordered by line or mass elements
higher than the eye level of a standing observer within which all points are
50
(_r-,=
tra~lsition
twn compartnlents
l---J
(De Veer et
\I\
al., 1977a).
Fig.2 (Centre) Complex of spatial types (a, b, c) observable in areas of undulating relief
(De Veer et al., 1977a).
Fig.3. (Bottom)
Compartment
51
tended and refined its principles into Method la, listed in Table I. Their aim
was to display differences in the size of space over large parts of the country.
In Method la, spaces were classified according to size (length and breadth
greater or less than critical values of 500 or 1500 m), relief and land use.
Finally, landscape types were translated into recreation suitability.
Vrij (Method lb; 1976) used the compartment
concept with critical values
for length and breadth of 500 and 1400 m in order to construct a basic
pattern of natural landscape units. The delineation of these relies more on
map and aerial photo interpretation
than on recognition of bounding elements
in the field. The landscape units formed the basis for several interpretative
maps. One of these was based on the degree of visual differentiation
within
the units estimated from the types of bordering elements, relief and land use.
Vrijs maps have possible applications in local and regional planning.
Smits approach (Method lc; 1976), already familiar to British landscape
scientists through the review given by Land Use Consultants (1971), is a
variation of the compartment
model. His criterion for classifying space was
based on whether a circle with a diameter of more than 2000 m, between
500 and 2000 m, or less than 500 m could be placed over a map of the
terrain without covering lines of trees or built up elements. The spaces were
subdivided on the basis of the degree of their enclosure by elements higher
than eye level. The units were also characterized
in terms of relief, land use,
settlement type, parcellation, tree species and presence of surface water.
Smit is the only worker to attempt a detailed landscape inventory of the
whole of The Netherlands.
De Veer et al. (Method Id; 1977a) have classified compartments
according
to their area (common class limits 25 and 100 ha). Where compartment
edges
varied greatly in transparency
this attribute was also used for classification,
Other attributes used have been relief, land use and structure (parcellation
and forest path pattern), and type of trees and building elements. The latter
were used for edge classification in space and for content classification in
mass. Maps made according to Method Id have been applied in local and
regional planning (e.g., extension of built-up areas and reallottment
schemes).
Fig.6 is an example of a landscape classification map produced by Method Id.
(2) Breadth of view method
The exponents of this method have been Van der Ham et al. (Method 2;
1970). Its main tenet is that the effectiveness with which an element limits a
view depends on its distance from the observer. Objects farther than 1200 m
away are normally considered not to delimit the space perceived, because the
fine details can no longer be discerned by the naked eye. In practice, the
critical distance varies from approximately
1000 to 1300 m depending on
the colour, contrast and extent of the objects seen, and the meteorological
conditions. Objects further than the critical distance are called extraocular.
A further, more arbitrary, critical distance of 500 m has been used to divide
52
5
rl
r2
73
i6
<500m
500~1200m
U200m
Fig.4. Classification
(1) Superocular
(2) Extraocular
(3) Ocular
(4) Intraocular
(5) Exocular
(6) Inocular
SP = arc of single
of observation
points according
to Van
angle of view to r3 (a) > 180.
angle of view to r3 (a) > 60 SP; > 100
angle of view to r2 (a) > 60 SP or MP.
angle of view within r3 (b) > 300.
angle of view to r3 (a) < 60 SP; < 100
rl (b) < 240.
-angle
of view to r3 (a) < 60 SP; < 100
rl (b) > 240.
panorama;
MP = summed arcs of multiple
-
(modified).
MP.
53
land lots, the relief (of which the higher elements are considered as spacedelimiting elements) and the presence of visible influences of hills, urbanization or main roads. Masses such as woods, urban or industrial areas are mapped
separately.
Maps made according to Method 2 have been published in different reports
of the National Physical Planning Agency (Rijksplanologische
Dienst). Interpretation systems have been developed to evaluate landscape as a natural
resource and to estimate its suitability for recreation and dwelling.
(3) Grid landscape surveys
In Method 3a (Kerkstra, 1974), square grid cells of 125 X 125 m were
classified according to several principles (e.g., visual complexity, variation in
complexity,
and seasonal variation). The classification of visual complexity
is based on measurements taken mainly from topographic and other maps of
the amount of mass, number of buildings and length of different tree-line
elements per cell (Fig.8). Cell data are combined for further analysis including mapping zones of breadth of view. This method has been applied to
local planning problems, for example, for suggesting new residential areas.
Koster and De Veer (Method 3b; 1972) have used much larger cells of
1000 X 1000 m. Every landscape element within the cell is assigned a point
value depending on its height and extent. The degree of differentiation
(i.e.,
complexity),
natural or artificial character and/or spatial type are determined
from the elements present in the cell and from the values assigned. The method
was used in a study for predicting alterations in landscape physiognomy
resulting from industrial activities along the Belgian-Dutch border.
Coeterier and Dijkstra (1976) have also used a grid approach to measure
changes in landscape caused by reallottment
schemes. Their survey covered
two aspects - first, a measure of the changes in hedgerow density, and second,
public reaction to the landscape alterations.
(4) Special purpose
method
54
THE SUITABILITY
OF DUTCH PHYSIOGNOMIC
LANDSCAPE
MAPPING
NIQUES FOR A NATIONAL
LANDSCAPE
MAPPING PROGRAMME
TECH-
Comparison
Fig.5. Topography
of part of the test area near Rhenen (Provinces of Utrecht and Gelderland, The Netherlands).
Scale 1:50 000. The area displayed in Figs 6, 7 and 8 is marked.
55
the methods in the field, a suitable test area of 60 km* was chosen near the
town of Rhenen in the Provinces of Gelderland and Utrecht (Fig.5). This
area was chosen for its wide variety of landscapes which include low forested
hills, large open areas, the flood plain of the Neder Rijn, smaller, complex
field patterns and urban influences.
All except Method la were mapped at a common field scale of 1:25 000;
the map for Method la was taken from the published map of the Geldersche
Vallei that covers the study area. Methods lb, lc, Id and 2 were mapped
over the whole study area, but Methods 3a and b (grid mapping techniques),
were mapped only over the 6 km* tract shown in Fig.5. Method 4 was confined to the flood plain of the Rhine. Figs 6, 7 and 8 show details of Methods
Id, 2 and 3a over the 6 km2 strip; all maps are reproduced in full in De Veer
et al. (197713).
The eight methods were compared on 19 points ranging from the number
of areas already mapped according to the method, the survey scales employed,
to the costs and procedures of data gathering, the flexibility of the approach,
the legibility of the maps and their declared application possibilities. Table II
presents some of these data concerning the extent of use, map scales, mode
of presentation and costs. Several methods, lc, Id, 2 and 4 presented the
landscape typology in a single map; others, notably Methods lb and 3a, used
several maps to present all their classifications. A wide range of scales has
been used, both between and within methods. Survey costs increased with
Fig.6.
Spaces
Landscape
classification
according
edge transparency
to Method
Size of space
Large
(>lOO
Open-edge
> 7 5% transparent
Half-open-edge
Enclosed-edge
Associations
25-75%
transparent
not occur
in this area.
ha)
Medium
(25-100
ha)
Small
(< 25 ha)
2*
3*
6*
of space-mass
Mass: transparent
Mass: transparent
and non-transparent
Mass: non-transparent
*Does
Mass
10*
11
12
Transparent
Transparent
and non-transparent
Non-transparent
13*
14
15*
56
Fig.7.
Landscape
classification
according
to Method
2 (Van
Space
Classification
observation
Map
code
Very
of
points
Summed
wide view
Wide view
View bounded,
far outlooks
Bounded
Closely
Other
fSB
landscape
but with
view
view
7 = periphery
of objects
in viewing-circle
(degrees)
1200 m
Extraocular
500-l
200 m
Ocular
500 m
Intraocular
>180
<180
<180
>60 SBf or
100-180
<180
<180
5-60
5-100
.240
<120
SBf
<5
bounded
types:
angles
5-60
5-100
SBf
>300
<60
<120
>240
<60
>300
<5
area;
8 = forest.
not occur
in this area.
Fig.8. Landscape
classification
according
to Method 3a (Kerkstra,
1974).
Visual
is ranked from 1 (small) to 5 (large). (Type 5 does not occur in this area.)
complexity
map scale, the ratio of field to office work and with the amount of data
collected. Method Id, which requires a complete field survey of the landscape
elements, was most expensive while the map produced for the whole country
by analysis of topographic and other maps (Method lc) was cheapest.
Table III compares the methods in terms of the dominant landscape
attributes used for classification. Although Method lc appears to have the
most comprehensive
cover of concrete landscape attributes, its complex and
*s 1
= Dfl4.20.
--~
la
lb
lc
Id
2
3a
3b
4
-_-.
Author(s)
Method
TABLE II
>2
4
1
>16
13
1
5
1
No. of
areas
mapped
1260
3850
37 000
1826
3211
100
1270
380
Area mapped
in published
studies
(km)
~~
100-250
25-250
600
5-100
25-250
60
25-100
50
---.-
Scales
(25 = 1:25 000
etc.)
GlO
1
1
>I
5
2
1
No. of
maps
per area
?
41-52
30-41
70-204
77-95
120
68-l 02
51
58
TABLE
III
Dominant
---__~
Method
Concrete
Tree
vegetation
Buildings
urbanization
Water
Relief
USC!
++
++
++
++
0
-
0
+
+
++
0
-
+
+
+
c
+
0
+
++
0
+
0
+
c+
c
+
0
++
0
0
0
-
0
+
D
+*
Land
la
lb
In
Id
2
3a
3b
4
Abstract attributes
attributes
++ = extensively
structure
+
+
-
Complexity,
diversity
++
0
++
f
. -~-
space
(-mass)
++
0
+
++
++
++
++
f
59
Users were asked about the application and data content of landscape
maps. Five categories of application scored high:
- vulnerability
designation (e.g., visibility of a new building, road or power
line);
- suitability designation (e.g., for different recreation activities);
- public landscape preferences (e.g., as determined by a questionnaire using
colour photographs of selected landscapes);
- landscape evaluation (using parameters such as diversity, rareness, or
replacement possibilities) for conservation planning;
- landscape design (the creation of new or modification
of old landscapes).
Only landscape management scored low.
Many users stated that data about the distribution of single landscape elements (e.g., the location of all rows of transparent, high poplars) or specific
combinations
of elements (concrete or abstract typologies) were often required, as well as information
on different visible influences such as industrial
or suburban areas. However, landscape evaluations for specific purposes were
not necessarily required to be displayed on maps but could be presented in
tabular form.
The main conclusion from the questionnaire
was that users demands for
physio~omi~
landscape mapping vary enormously,
both in terms of mapping
scale and map content.
DISCUSSION
60
REFERENCES
Brussaard,
W., undated. The rules of physical planning in The Netherlands.
Ministry of
Housing and Physical Planning, The Hague, 36 pp.
Coeterier,
J.F., 1977. De betekenis
van de omgeving IV. Literatuurstudie
naar de waarneming en de waardering
van landschappen.
Rijksinstituut
voor Onderzoek
in de Bosen landschapsbouw
Lde Dorschkamp,
Wageningen,
102 pp.
Coeterier,
J.F. and Dijkstra,
H., 1976. Research
on the visual perception
and appreciation
01, and visual changes in a hedgerow landscape.
Landscape
Plann., 3: 421-452.
61
62
RijkspIanologische
Dienst, 1974. Ori~nteringsnota
ruimtelijke
ordening.
Den Haag, 124 pp.
Sehuurmans,
J. and Van Schie, J., 1968.
Landschapstypen.
Tijdschr.
K. Ned. Heidemaatsch.,
79: 101-110.
Smit, H.F., 1976. Uitgewerkte
schets voor een middelsehalige
classificatie
van Nederlandse
landschappen.
Zwolle, Van Hille Gaerthestraat
42. 65 pp. (stencil).
Steinitz,
C., Murray, T., Sinton,
D. and Way, D., 1969. A comparative
study of resource
analysis methods.
Harvard University,
Department
of Landscape
Architecture,
Cambridge, Mass., 382 pp.
Tomlinson,
R.F., Calkins, H.W. and Marble, D.F., 1976. Computer
handling of geographical data. UNESCO
Natural Resources
Research
XIII, Paris, 214 pp.
Van der Ham, R.J.I.M.,
Schut, G.F.E. and Iding, J.A.M.E.,
1970. Een voorstel voor een
nieuwe landsehaptypologie
naar visuele kenmerken.
Stedebouw
en Volkshuisvesting,
51: 421-438.
Van der Poel, A.J., 1976. From landscape care to research in landscape planning.
Landscape Plann., 3: 363-370.
Vrij, F.V., 1976. Landschapsbeleving
en ruimtelijke
planning.
Provinciale
Planologische
Dienst Noord-Brabant,
s Hertogenbosch,
70 pp.
Weir, M.J.C.,
1976. Landscape
evaluation
in Great Britain - a short review. Meded. Werkgemeenschap
Landschapsecologi~h
Onderzoek,
3: 15-20.
Zonnevetd,
IS., Tjallingii,
S.P. and Meester-Broertjes,
H.A. (Editors),
1975. Landschapstaal.
Meded. Werkgemeenschap
Landschapsecologisch
Onderzoek,
extra nummer,
38 pp.