Journal of Environmental Management (2002) 66, 239247 doi:10.1006/jema.2002.0574, available online at http://www.idealibrary.

com on

Erosion risk analysis by GIS in environmental impact assessments: pru a case studySeyhan Ko Dam construction
S . S ahin* and E. Kurum
Ankara University Faculty of Agriculture, Department of Landscape Architecture, 06110 Ankara, Turkey
Received 11 July 2001; accepted 8 March 2002

Environmental Impact Assessment (EIA) is a systematically constructed procedure whereby environmental impacts caused by proposed projects are examined. Geographical Information Systems (GIS) are crucially efcient tools for impact assessment and their use is likely to dramatically increase in the near future. GIS have been applied to a wide range of different impact assessment projects and dams among them have been taken as the case work in this article. EIA Regulation in force in Turkey requires the analysis of steering natural processes that can be adversely affected by the proposed project, particularly in the section of the analysis of the areas with higher landscape value. At this point, the true potential value of GIS lies in its ability to analyze spatial data with accuracy. This study is an attempt to analyze by GIS the pru areas with higher landscape value in the impact assessment of dam constructions in the case of Seyhan-Ko Hydroelectric Dam project proposal. A method needs to be dened before the overlapping step by GIS to analyze the pru Hydroelectric Dam project proposal of the present work, areas with higher landscape value. In the case of Seyhan-Ko considering the geological conditions and the steep slopes of the area and the type of the project, the most important natural process is erosion. Therefore, the areas of higher erosion risk were considered as the Areas with Higher Landscape Value from the conservation demands points of view. # 2002 Published by Elsevier Science Ltd

Keywords: GIS, EIA, erosion risk, dams.

Introduction
The use of GIS for impact assessment
GIS have been applied to a wide range of different impact assessment projects. The most common GIS applications for the impact evaluation are the roads, pipelines, housing developments, coast and ood protection works, dams, tourism related projects, ports and power lines. GIS is also widely used by environmental consultancies for all impact assesso, 1998). ment stages (Joa

* Corresponding author. Email: sahin@agri.ankara.edu.tr

03014797/02/$ see front matter

Many of the GIS applications for impact assessment use basic GIS functions such as measurement of lengths and areas, map production, buffering, o, 1998). and the classic overlay operation (Joa Recognizing the spatial nature of many environmental impacts, overlay mapping in the manner of ecosystem analysis was pioneered by McHarg (1969) who is both a landscape architect and a city planner. Compared with the cumbersome manual process of overlaying transparencies, the overlay analysis is made much more powerful through the use of GIS (Smith, 1993). Arithmetic and logical overlay operations are part of all GIS software packages. Arithmetic overlay includes such operations as addition, subtractions, division, and multiplication of each value in a data layer by the value in the
# 2002 Published by Elsevier Science Ltd

240

S . S ahin and E. Kurum

corresponding location in a second data layer. A logical overlay involves nding those areas where a specied set of conditions do or do not occur (Aronoff, 1991). Those abilities of overlay mapping by GIS technology make it an efcient tool for the assessment of incremental changes in an environment; the site selection for a project proposal; the comparison of the alternatives; the examining and visually displaying the spatial nature of the impacts in an EIA work.

Erosion risk mapping

Investigations on the evaluation of water erosion and soil loss started in the beginning of 1900s, but more signicant studies were conducted after the 1940s. Initially soil loss was predicted in accordance with empirical equations. After the 1950s parametric models have been developed. This parametric models rely on statistical relationships between soil-loss and various parameters derived from larger sets of data such as rainfall, soil erodibility, slopelength, slope gradient, crop management, etc. The most common and widely used parametric model for the prediction of soil losses is the well known `Universal Soil Loss EquationUSLE' (MAPA/ ICONA, 1983; MOPU, 1985), developed in the USA as an aid for conservation considering agricultural activities. In general, the methods to measure soil loss can be divided into two: Quantitative (such as USLE, MUSLE) and Qualitative Methods (such as ICONA, CORINE). Quantitative methods usually involve the measurement and quantication of various components. Qualitative methods, on the other hand, rely more on the judgments and responses of the analyst. However it is important that this professional judgment takes place in the context of a systematic and structured approach (NRA, 1993). In practice, a more accurate prediction of erosion risks can be achieved by the integration of the qualitative and quantitative methods considering long-run land conservation strategies, management and monitoring (Gardi et al., 1996). PAP/RAC (Priority Actions Program/Regional Activity Center) of the MAP (Mediterranean Action Plan)UNEP in corporation with FAO prepared the Guidelines for Erosion and Desertication Control Management. This Guidelines, dealing with management related issues of erosion/desertication control are a logical and thematic follow up of the guidelines on mapping and measurements by PAP/RAC in 1997 (UNEP/MAP/PAP 2000).

The areas with higher landscape value: problem description and target formulation
The analysis of the areas with higher landscape value is a requirement of the current EIA regulation in Turkey within the content of General Format of EIA Report (Section IV.2.14). This practice is among the responsibilities of landscape architects, however there is still room for improvement for a widely accepted common approach. From the view point of the landscape architecture discipline, in determining of `the areas with higher landscape value', the ecological, cultural and visual characteristics of the area in question are examined with an holistic approach considering Conservation Demand and Development Demands. Human welfare and the quality of life in many ways depend directly and indirectly on the availability of environmental goods and services, thereby on the natural process and ecosystem that provide them (Vellinga et al., 1994). This means that the conservation demand of the area, and accordingly the interrelations among the components of the bio-physical environment should be analyzed, and the section of determining the areas with higher landscape value of the EIA process presents an opportunity for this purpose. In practice, however, the assessment of mentioned natural processes, what is also called ecological impact assessment in literature, is not executed properly in formal EIA studies in Turkey. The more important purposes of this paper are to disclose this deciency of EIA studies, to present the way in which this problem can be overcome, and to conduct a case study to prove the crucial outputs of the analysis of the areas with higher landscape value. Another equally important purpose is to show indispensability of GIS in such cases. At the present article, erosion risk analysis is recommended to determine the value of natural landscape taking into account the characteristics of the environment and the nature of the dam construction as a development demand.

Determining the areas with higher landscape value: a case

Dams may have a series of environmental consequences that can or can not be anticipated. Some of these are dealt with in greater detail elsewhere, such as subsidence, earthquake triggering, the transmission and expansion in the range of organisms, the build-up of soil salinity, changes in ground-levels creating slope instability, logging

GIS Environmental Impact Assessments

241

Map 1.

ksu Brook. Study area: Dam lake catchment over Seyhan-Go

and sediment load reduction of the river downstream (Goudie, 1995). Otherwise dams can be effected by the physical environment where they are built. The life span of a dam is very strongly related with the soil erosion caused by rainfall runoff. Hence one of the steering natural processes that may have an impact on proposed project and which in turn may be affected by the project itself is `erosion process'. Vegetation cover and accordingly fauna, river sedimentation, dam life span can be affected by this process as well. Thus, the `erosion process' is considered in the analysis of `the areas with higher landscape value' at present work. In this paper this process is analyzed in the case area.

the dam lake was determined (Map 1). Impact areas related to the proposed project were divided into two regions within the entire catchment of the dam lake. First degree impact areas are the slopes, which feed directly the dam lake by rainfall run-off. Second degree impact areas are the rest of the catchment in ksu Brook are loaded which the tributaries of the Go by sediments at rst, than those tributaries carry whole sediment loads into dam lake.

Method
In this study ICONA erosion risk method developed by the Directorate General for the Conservation of Nature (DGCONA, previously ICONA) in Spain was used and adopted to the study area. The method integrated with the Guidelines for Mapping and Measurements of Erosion prepared by UNEP/MAP/ PAP (2000), particularly in order to use a standardized mapping legend (Annex IV of the Guidelines) and to recommend the follow up steps of erosion risk analysis.

The project area

pru In the present work proposed Seyhan-Ko Hydroelectric Power Station to be constructed ksu Brook is taken as the case study. In over Go the analysis of natural processes that form a landscape, the natural boundaries should be taken as the study area limit, for this reason, the catchment of

242

S . S ahin and E. Kurum

QUALITATIVE EROSION ANALYSIS

VEGETATION

SLOPE

GEOLOGY

SOIL PROTECTION

ERODIBILITY

EROSION RISK

Figure 1.

Method: MOPU (1985), MAPA/ICONA (1983), and MAPA/ICONA (1991).

The Directorate-General for the Conservation of Nature (DGCONA) executed a project called n en el LUCDEME (Lucha Contra la Deserticacio neo/Combat with Desertication in Mediterra Mediterranean) at the South-East of Spain between the years of 19811984, and developed a method in order to dene erosion risk (MOPU, 1985). Figure 1 illustrates the steps of the method. For the elaboration of the present method, conventional maps (at 1/25 000 scale) of topography (for slope analysis), vegetation cover (for soil protection grades), geology (for erodibility analysis) were transferred to the computer media by the AutoCAD R.14 software. The rectication of the transferred data was performed by ERDAS Imagine 8.3. Slope analysis was made by LANDCADD R.12 software. Then all the data were transferred into a GIS engine that is ArcCAD R.14. This software provided the ability to create, manipulate, analyze and display topologically correct geographic data in digital form.

environmental parameters. In this case study, for the denition of soil protection grades the followulica n de Hidra ing table developed by IFIE-Seccio Torrencial del Antiguo Instituto Forestal de Investigaciones y Experiencias in 1968 (MAPA/ ICONA, 1983) was used (Table 1). The vegetation cover map to be used in this study was re-coded from the `Vegetation Cover Map of Turkey' by OB-Directorate General of Forestry in accordance with the following classes dened from Table 1:      Bare areas; Dense woodland; Loose woodland; Degraded woodland; Agricultural areas.

Slope
In accordance with MAPA/ICONA (1983), following slope gradients were used: 012, 1218, 1824, 2435, 3560 and b60%.

Vegetation cover and soil protection grades

Vegetation cover is the variable controlling erosional activity that is most affected by human manipulation and is therefore an important component of any predictive model (Trimble, 1990). Each vegetation cover type has its own soil protection grade, which depends upon land-use differences, current soil management practices, and other

In the next step, the geographical data (slope condition and vegetation cover) of the site were interpreted by the indices given Table 1 in order to produce site specic composite indices. `The soil protection grades' being qualitative values were then assigned to these composite indices in the following fashion: 10: very high (VH), 0908: high (H), 0706: moderate (M), 0403: low (L), 0200: very low (VL). Accordingly, The map of `Soil Protection Grades' was produced with the aid of GIS by the superimposition of soil condition and vegetation cover data in accordance with these qualitative values shown in Table 2.

Geology and erodibility

Physical and chemical characteristics of bedrocks have signicant inuence over erosion process. Geological structure map to be used in this study was re-coded from the `Geological Resources Map of Turkey' by MTA, Directorate General of Mineral

GIS Environmental Impact Assessments Table 1. Soil protection indices by vegetation cover by IFIE (MAPA/ICONA 1983). Statement Dense woodland cover (07 density) Woodland cover with less than 07 density and non-degraded bushes and herbaceous plant cover Woodland cover with less than 07 density and degraded bushes and herbaceous plant cover Non-degraded bush cover Degraded bush cover Well-conserved pasture Degraded pasture Agriculture Agriculture without conservation practices Agriculture with conservation practices Bare-land Slope for any slope gradient for any slope gradient 3 2 1 for any slope gradient 3 2 1 530% b30% for any slope gradient 3 2 1 1 and 2 3 3 2 1

243

Vegetation type Forest

Protection index 10 10 04 08 10 10 02 02 08 09 06 03 00 05 09 10 03 00 05 09

1. Slope inferior than the gradient of erosion initiation. 2. Slope between the gradient of erosion initiation and total dragging. 3. Slope superior than the gradient of total dragging.

Table 2. Soil protection grades and Erodibility (Adopted from MAPA/ICONA 1983, MAPA/ICONA/1991), Atucha et al., 1993; Gardi et al., 1996. Slope 012 Type of vegetation cover Soil protection grades Bare-lands Dense woodland Loose woodland Degraded busy areas Agricultural areas Material Erodibility Compacted siliceous rocks Slightly consolidated rocks 1218 1824 2435 3560 b60

V VH VH H M

V VH H M V

V VH M V V

V VH M V V

V VH V V V

V VH V V V

EN EL

EL EM

EL EM

EM ES

ES EV

ES EV

Soil protection grades: VH, Very High; H, High; M, Moderate; L, Low; VL, Very Low. Erodibility classes: EN, erodibility from none to low; EL, erodibility from low to moderate; EM, erodibility from moderate to severe; ES, erodibility from severe to very severe; EV, erodibility from very severe to total dragging.

Research and Exploration in accordance with the classication of erodibility by MAPA/ICONA 1983.  Igneous rocks;  Well cemented calcareous rocks;  Compacted siliceous rocks;

 Slightly consolidated rocks;  Soft formations;  Alluvial deposits. After this interpretation, it was observed that there exist only the following two geological classes

244

S . S ahin and E. Kurum

in the study area:  well cemented calcareous rocks;  siliceous rocks. The inherent characteristics of these re-coded geological structure were then interpreted with slope gradients to come up erodibility classes shown in Table 2 and superimposed to produce the map of `Erodibility'.

Table 3, and Map 2 was produced as a result. Table 3 enumerates the degrees of erosion risks in accordance with Annex IV on Mapping Legend of the Guidelines for Erosion and Desertication Control Management of UNEP/MAP/PAP (2000): These numbers denote the following: 1: very severe; 2: severe; 3: moderate; 4: slight; 5: very slight. As it is seen on Map 2, the signicant coverage of the study area including the surroundings of the dam lake presents very severe risk of erosion. When analyzed by conventional indices, such as
Table 3. Criteria for erosion risk (Adopted from MAPA/ ICONA 1983, MAPA/ICONA/1991), Atucha et al., 1993; Gardi et al., 1996. Erodibility VH EN EL EM EV EE 1 1 1 2 2 Soil protection grades H 1 1 2 3 3 M 1 2 3 3 4 L 2 3 4 5 5 VL 2 4 4 5 5

Result
The `Erosion Risk Map' which is one the main goals of this article was produced by the superimposition of two critical attributes of the site `Soil Protection Map' (produced by the superimposition of the map of vegetation cover and slope) and `Erodibility Map' (produced by the superimposition of the maps of geology and slope). The superimposition of the two attributes was achieved according to the criteria presented in the

Map 2.

Erosion risk.

GIS Environmental Impact Assessments

Moderate 0.11% Severe 0.10% Slight 6.75%

245

Ver y slight 10.78%

Erosion risk degree Very severe Severe Moderate Slight Very slight

Very severe 82.26%

Figure 2.

The area portion of erosion risk degrees.

percentages (Figure 2), the picture of the distribution reveals a clear pattern of gravity as far as the erosion risk is concerned. The areas at severe and very severe level of erosion risk excluding rocky areas are those in high landscape value from conservation point of view. For the sustainability of natural landscape, such activities that can accelerate erosion should not be permitted in those areas. Therefore, to improve the conditions of land in order to provide resistance against the potential erosion problem revegetation practice needs to be improved in particular on the rst degree impact areas around the dam lake to be able to both extend the dam life-span as well as to decrease the potential risk of rainfall erosion. As it is stated before in the section of project area denition, rst degree impact areas are the slopes that feed directly to the dam lake by rainfall surface runoff. The conservation value of these areas is very high. The more rational and effective impact mitigation measures can be achieved by the combination of erosion risk map with existing land-use types and socio-economic structure. The further steps of ICONA method provide an opportunity for this purpose with its erosive landscape classication.

to seas and lakes is about 500 million tons per year an et al., 2000). In this end environmental (Dog management practices such as EIA have to take into consideration properly this severe problem in their contex. One of the main contributions of this paper is to propose a method how to carry out erosion risk analysis in an EIA work under the section of the analysis of the areas with higher landscape value toward internationally appreciated standards. There are some commonly known soil erosion risk estimation and mapping methods such as ICONA, CORINE,1 USLE, etc. In this present work ICONA method was used and integrated with the Guidelines for Erosion/Desertication Control Management which were prepared by UNEP/MAP/ PAP (2000) in corporation with FAO. All these methods have their own merits and the paper does not aim to make an argument in favor of the ICONA method. In the countries where the problem of soil erosion is very severe and the dearth of detailed and upgraded data have been signicant and the availability of data is very expensive and time consuming, the design of data processing within the framework of a method carries a crucial importance. In this sense the ICONA method presents an opportunity for rapid evaluation of potential erosion risk in large areas. Other methods can be easily incorporated with the results as the complimentary part of it taking into account mainly the areas where the erosion risk is severe.

Technological and practical aspects

The denition of the areas with higher landscape value can be considered as an Ecological Impact Assessment (EcIA) work of the EIA in respect to ecological concerns. Briey, EcIA is a formal process of identifying, quantifying and evaluating the potential impacts of dened actions on ecosystem (Treweek and Hankard, 1998). At the same time the effects of ecological phenomena over the proposed projects is the question of EcIA. Treweek (1999) states that GIS can be used as a tool in scooping or conceptualizing the EcIA, deriving suitable study limits and generating appropriate impact scenarios and mitigating strategies as well as simply handling

Discussion
The results of this study might be discussed in two aspects.

Methodological aspects
1

Erosion in Turkey is one of the most important ecological problems threatening natural resources. According to the sediment measurements made on 26 large basins, the amount of sediment transported

CORINE Method in erosion mapping analyses the following factors for the determination of potential and actual erosion risk; soil texture, soil depth, stoniness, Modied Fournier Indices and Bagnouls-Gaussen Draught Indices and slope for the potential erosion risk, and also vegetation cover for the actual erosion risk an et al., 2000; Genc (Dog ler et al., 2000).

246

S . S ahin and E. Kurum  zel, M. E. and Yldrm, H. an, O., Ku Dog cu kc akar, N., y (2000). Erosion Risk Mapping of Dalaman Basin Located in West Mediterranean Region Using Corine Method. In International Symposium on Desertication, Konya, Turkey. Gardi C., Pisa, P. R., Rossi, M., Kurum, E. and S ahin, E. (1996). Qualitative Analysis of Land Degradation by Erosion in Centonara River Basin, Bologna, Italy. In Proceedings of 1st International Conference on Land Degradation, pp. 204216, Adana, Turkey. Genc ler, G., Azdiken, S. and Altan, M. (2000). Preparation of Soil Erosion Risk Map of Middle Sakarya Valley by Using Remote Sensing (RS) and Geographical Information Systems (GIS). In Proceedings of 2nd International Conference on GIS for Earth Science Applications, on CD, I zmir, Turkey. Goudie, A. (1995). The Human Impact on the Natural Environment. Blackwell Publisher Ltd, 454pp, UK. o, E. M. (1998). Use of Geographical Information Joa Systems in Impact Assessmet. In Environmental Methods Review: Retooling Impact Assessment for the New Century (A. L. Portae and J. F. Fittipaldi, eds.), pp. 154170, AEPI, IAIA. MAPA/ICONA (1983). Paisajes Erosivos en el Sureste a para el Estudio de su ol: Ensayo de Methodolog Espan n y Cuanticacio n. Proyecto LUCDEME, Cualicacio 66p, Espan a. o MAPA/ICONA (1991). Metodologia para el Disen de Actuaciones Agrohidrologias en las Cuencas del Ambito Mediterraneo. Proyecto LUCDEME, pp. 131, Espan a. McHarg, I. (1969). Design with Nature. Doubeday: Garden City, NY. n de la a para la Evaluacio MOPU (1985). Methodolog n H n General del Medio Amdrica. Direccio Erosio biente, Espan a. NRA (1993). River Landscape Assessment: Methods and Procedures. Conservation Technical Handbook No. 2, National Rivers Authority, UK. ra Bilgi S ahin, S . and ve C abuk, A. (1998). Cog erlendirmesinde Sistemlerinin C evresel Etki Deg Kullanm (Use of Geographical Information Systems in Environmental Impact Assessments). Ula silabilir GIS Semineri, Saysal Grak, 16 Aralik, Ankara. Smith, L. G. (1993). Impact Assessment and Sustainable Resource Management. Longman Group, p. 21, UK. Stow, D. A. (1993). `The Role of Geographical Information Systems for Landscape Ecological Studies'. In Landscape Ecology and Geographical Systems (R. Haines-Young, D. R. Green and S. Cousins, eds.), pp. 1119, Taylor & Francis Ltd. Treweek, J. and Hankard, P. (1998). Ecological Impact Assessment. In Environmental Methods Review: Retooling Impact Assessment for the New Century, (A. L. Portae and J. F. Fittipaldi eds.), pp. 262272, AEPI, IAIA. Treweek, J. (1999). Ecological Impact Assessment. Blackwell Science Ltd., 352pp. Trimble, W. S. (1990). Geomorphic Effects of Vegetation Cover and Management: Some Time and Space Considerations in Predicting of Erosion and Sediment Yield. In Vegetation and Erosion (J. B. Thornes ed.), John Wiley and Sons Ltd, UK. UNEP/MAP/PAP (2000). Guidelines for erosion and desertication control management with particular

the relevant data and making them accessible. GIS can therefore play an important part in managing the EcIA process. GIS holds much promise for supporting numerical modeling of spatially distributed ecosystem processes. There are a number of ways that GIS and ecosystem models can be integrated for ecological studies (Stow, 1993). This paper presents an example for its potential uses. The further step of the erosion risk analysis executed in this article as a part of EIA (or being the possible requisite of an other environmental management activity) should be integrated with the recommended procedures by UNEP/MAP/PAP (2000) for erosion control management as:  Integration of mapping outputs with socioeconomic and land-use features;  Denition of the negative impacts of the risk areas over proposed project and vice versa, identication of priority areas, formulation of remedial;  Formulation of the strategy and the program for management of erosion control, and implementation;  Environmental impact monitoring and auditing. This evaluation practice of the erosion risk levels and the higher landscape values for dam projects has to be started at more strategic levels before project level EIA, when the policies, programs and master plans for dam constructions are being developed over a stream. This study also shows that the environmentally determinative factors and processes may extend beyond of the legal and political boundaries. For the success of the overall project in any EIA practice the responsibility area of the developer should not be limited by the legal and political boundaries but include naturally effective areas. The neighboring parties and the shareholders need to be acknowledged, and be given responsibilities for the well-being of the project and the nature itself.

References
Aronoff, S. (1991). Geographical Information Systems: A Management Perspective. WDL Publications, Ottowa, 294pp, Canada. Atucha, J. L., Ben Dadj Ali, H., Echeverria, J. L., Kristensen, M. J., Rios, J., Rozpide, M. and S ahin, S . (1993). `Nuevas Orientaciones para el Uso Integrado de los Recursos Naturales en la Comarca del Moncayo', Volume I, pp. 3144, Instituto Agronomico Mediterraneo del Zaragoza-Espan a.

GIS Environmental Impact Assessments reference to Mediterranean coastal areas. Split, Priority Action Programme. Vellinga, P., R. S. de Groot, R. J. T. Klein Rudolf de Groot and Richard Klein (1994). An Ecologically Sustainable Biosphere. In The Environment: Towards a

247

Sustainable Future (Dutch Committee for Long Term Environmental Policy ed.), pp. 317346, Environment and Policy-series Vol. 1, Kluwer Academic Publ., Dordrech, The Netherlands.