Proceedings of the Open source GIS - GRASS users conference 2002 - Trento, Italy, 11-13 September 2002 Potential Erosion Map For Bagmati Basin Using GRASS GIS Jha Raghunath Department of Civil Engineering, Institute of Engineering, Pulchowk, Kathmandu, Nepal Tel: + 977-1-545368, e-mail: <rnjha@ioe.edu.np> 1. Introduction Nepal varies from latitude 26 0 22 to 30 0 27 and longitude 80 0 04 to 88 0 12 and it occupies large part of central Himalayas and its foothills. Total area is 147,181 sq. km and about 80% of the land is mountainous with rugged topography. Elevation varies from 70 meter to 8848 meter above mean sea level and increase from south to north. Nepal is divided into five physiographic regions from south to north namely the high Himal, high mountains, middle mountains, Siwalik and Terai. The most vulnerable area is middle mountains and Siwalik in both terms of natural and man made induced erosion causing sever damage of lives and properties. RUSLE is the most extensively used empirical soil erosion model. It is the present state of art in soil erosion modeling. It has been modified to give sediment yield as well. Basically RUSLE, which lumps enter-rill and rill erosion together, is a regression equation. The Revised Universal Soil loss Equation (RUSLE) [1] like its predecessor the Universal Soil loss Equation (USLE) [2] is an erosion prediction model designed to predict the long-term average annual soil loss from specific filed slopes in specified land use and management systems (i.e. crops, rangeland, recreational areas). GRASS GIS software has been adopted to prepare a potential erosion map in the Bagmati basin using Revised Universal Soil Loss Equation (RUSLE). 1.1 RUSLE Formulation The RUSLE quantifies the soil erosion as the product of six factors representing rainfall and runoff erosivity (R), soil erodibility (K), slope length (L), slope steepness (S), cover and management practices (C), and supporting conservation practices (P). The equation is thus: A= R*K*L*S*C*P (1) where A is the computed spatial and temporal average soil loss per unit of area. Only two of the six factors in the equation, the rainfall and runoff erosivity factor (R) and the soil erodibility factor (K), have units. The average annual soil loss per unit of area is expressed in the units selected for K and the period selected for R. 2 Project Description The area to be investigated is within the middle mountain region of Nepal as shown in Figure 1, in the watershed belonging to river Bagmati. The area is chosen because of its bio- climatic diversity due to elevation differences from valley floors to mountain summits, and related land use changes having influence on soil erosion, which is considered typical for the Middle Mountains of Nepal (Figure 2). The Bagmati Watershed covers an area of 3500 sq. km. and drains out of Nepal across the Indian State Bihar to reach the Ganges. It, partially or wholly, spreads over 8 districts of the kingdom, Kathmandu, Lalitpur, Bhaktapur, Makwanpur, Kavre, Sindhuli, Rautahat and Sarlahi. For
2 Potential Erosion Map For Bagmati Basin Using GRASS GIS technical purposes it has been classified into three main areas: the Upper, Middle and the Lower Bagmati Watershed Areas. The Upper Bagmati Watershed Area covers the whole of the Kathmandu valley including its source at Shivapuri. From the Chovar gorge, the river flows into the Middle Bagmati Watershed Area across the Mahabharat and Siwalik ranges. The catchment area of upper and middle Bagmati Basin is 2807 km 2. Figure 1: Location of Project Area in Nepal The Bagmati watershed area is mainly composed of three parts: Kathmandu Valley, the Middle Bagmati watershed area and the Lower watershed area. The terrain of the upper and middle BWA is rugged and comprised of several steep mountains except Kathamndu valley as shown in Figure 3. The Middle Bagmati watershed area (MBWA) consists of the two physiographic regions of Nepal. Those regions are: (i) the Siwalik region (450-815 m above MSL, in the southern part of the MBWA) (ii) the middle mountain region (Highest pt. 2625m, in the Northern part of MBWA). The Kathmandu Valley is an almost circular, tectonic basin, some 26 km in E-W direction and 20 km in N-S direction. The valley bottom is composed of the small ridges and elevated flat land and low basins drained by Bagmati river and its tributaries. The average altitude is 1350 m above msl. The surrounding mountains, whose height range 1500 and 3000 m, contain 4 major passes. The climate of the Bagmati watershed can be subdivided into three altitude/climate zones. These are: Subtropical sub humid zone below 1000 m. The southern most parts of the Bagmati Watershed area including the Siwaliks region lie in this zone Warm temperate humid zone between 1000-2000 m. A large part (more than 60 %) of the BWA lies in warm temperate humid zone between 1000-2000m altitudes. Cool temperate humid zone between 2000-3000 m. Only a small portion (about 5%) of the Bagmati watershed falls above 2000m.
Jha Raghunath 3 Figure 2: Project Area Details 3 Calculation of RUSLE Parameters 3.1 Rainfall Erosivity Factor(R) This is the rainfall erosivity index that represents the energy that initiates the sheet and rill erosion [3]. It is a function of kinetic energy (E). In this research the rainfall factor was computed from the following empirical relationship, as described in [4]. R = 0.1059a. b. c. + 52.0 ----- (2) Where R = rainfall factor (Tons/ha/cm/hour); a = average annual rainfall (cm); b= maximum 24 hours rainfall with a return period of 2 years (cm/24hours); and c = one hour maximum rainfall with recurrence interval of 2 years. In Bagmati watershed there are six daily rainfall stations and two high resolution rainfall stations; After analyzing the existing 10 years data it was observed that a = 130.4 cm, b = 12.7cm/day, c = 4.5cm/hour and R is calculated as 841.0 Tons/ha.
4 Potential Erosion Map For Bagmati Basin Using GRASS GIS Figure 3: DEM and River Network of Upper and Middle Bagmati River Basin 3.2 Soil erodibility factor (K) Soils of Bagmati Watershed area can be divided into 3 major groups: Major soil types of the Siwalik range are Udipsammente and Dystrochrepts. The texture of these soils is sandy in nature that varies from sandy clay to loam. Soil type of Rhodostalfs is commonly found in the gently undulating slopes. These soils are prone to severe soil erosion. The soil in the south face on low altitude of Mahabharat range is Dystrochrepts and Hapluclafs. These soils are mostly cultivated. Major soil types of mountainous lands are Haplumbbrepts and Dystrochrepts. K is the soil erodibility index referred as mean annual soil loss per unit of R for a standard condition of bare soil, recently tilled up-and-down slope with no conservation practice and on slope of 5 0 and 22m length. Soil erodibility is related to the integrated effect of rainfall, runoff, and infiltration. From different available literature the following K value for different crop was established. SN Soil Type Erodibility Factor 1 Sand/Cobbly 0.1 2 Sandy 0.12 3 Variable 0.12 4 Loamy 0.23 5 Loamy/Boulder 0.2 6 Loamy skeletal 0.23 7 Fragmental/sand 0.10 Table 1: The adopted value of K for different soils A digital soil map was collected from department of the Land Resources in Arc/info format. The Arc/info format map is converted in GRASS GIS function using r.in.arc function. The soil samples were reclassified using r.reclass function into 7 classes and values of kfactor (erodibility factor) from Table 1 were assigned using again by r.relcass function.
Jha Raghunath 5 3.3 Slope length and steepness factor Slope length factor in raster GIS The concept of replacement of slope length with up slope drainage area per unit of contour length applied. With this substitution each of the grid cells within the DEM is considered as a slope segment having uniform slope. To incorporate the impact of flow convergence, the hillslope length factor was replaced by upslope contributing area A. The modified equation for computation of the LS factor in GIS in finite difference form for erosion in a grid cell representing a hillslope segment was derived by [5]. A simpler, continuous form of equation for computation of the LS factor at a point r=(x, y) on a hillslope is LS(r) = (m+1) [ A(r) / a 0 ] m [ sin b(r) / b 0 ] n (3) where A[m] is upslope contributing area per unit contour width, b [deg] is the slope, m and n are parameters, and a 0 = 22.1m = 72.6ft is the length and b 0 = 0.09 = 9% = 5.16deg is the slope of the standard USLE plot. It has been shown that the values of m=0.6, n=1.3 give results consistent with the RUSLE LS factor for slope lengths <100m and slope angles <14 deg for slopes with negligible tangential curvature. Exponents m and n can be calibrated if the data are available for a specific prevailing type of flow and soil conditions. The DEM, which was created by contour interval of 20 m, was collected by Department of Survey in Arc/info format in 50m-grid resolution. The imported map was resampled into 100m-grid resolution using r.reclass command. Two methods can be used to calculate Slope length factor in GRASS GIS. The first method is as follows: Using r.flow commands calculate accumulated flow (flowacc) using r.slope.aspect command calculate slope of the catchment (slope) using r.mapcalc calculate lsfactor using following expression lsfactor = 1.6*exp(flowacc*resolution/22.1,0.6)*exp(sin(slope/0.09,1.3) The second algorithm is simple. In this method the command r.wastershed is used, and this command creates LS factor directly. The calculated LS factor should be divided by 100 to convert as real LS factor. The method is as follows: using r.watershed calculate LS factor using r.mapcalc calculate o lsfactor = LS / 100.0 3.4 Cover management factor C is the crop management factor, which represents the ratio of soil loss under a given crop to that from bare soil. For better result C factor should be calculated from the equations which is as follows: SLR = PLU*CC*SC*SR*SM.. (4) Feature Feature Soil ID C. Factor Type Category Cultivation Agriculture 1 0.42 Vegetation Forest 2 0.01 Shrub 3 0.1 Other Urban Area 4 0.05 Swamp 5 0.01 Sand 6 0.35 Snow 7 0.01 Barren Land 8 0.7 Water Body 9 0.0 Table 2: The adopted value of C for different Landuse.
6 Potential Erosion Map For Bagmati Basin Using GRASS GIS RUSLE looks at the impact of cropping and management on several sub factors. It looks at the impacts from previous cropping and management (prior land use, PLU), the protection offered the soil surface by vegetative canopy (canopy cover, CC), the reduction in erosion due to surface cover and surface roughness (surface cover, SC; surface roughness, SR), and in some cases the impact of low soil moisture (SM) on reduction of runoff from low-intensity rainfall. C values were estimated for Nepalese condition by different literatures and presented in Table 2. Figure 4: Landuse Map of the Upper and Middle Bagmati River Basin A digital landuse map created from 1:25,000 scale map were collected from Department of survey in Arc/info format. The map was exported from Arc/info format and imported in GRASS GIS using g.in.e00 command. The vector map of landuse is converted to raster map using v.to.raster command. A simplified landuse map is shown in Figure 4, which is created using ps.map command. R.reclassify command was used to create cfactor using according to Table 2. Using R.report command the area of selected crops are calculated and presented in table 3. SN Name Area (Km2) Area % 1 Agriculture 860.1 30.7 2 Forest 1603.5 57.8 3 Shrub 179.9 6.4 4 Urban 22.73 0.8 5 Others 118.76 4.3 Table 3: Area of Different Crops in Upper and Lower Bagmati River Basin 3.5 Support practice factor The support practice factor "P" in RUSLE is the ratio of soil loss with a specific support practice to the corresponding loss with up and down slope tillage, which has a value of 1. The support practices principally affect erosion by modifying the flow pattern, grade, or direction of surface runoff. For cultivated land the support practice generally includes
Jha Raghunath 7 contouring, strip-cropping, terracing, and subsurface drainage. P factor is calculated multiplying the sub factors such as contouring, terracing and strip cropping. P factor for this study was assumed as 1.0. 4. Potential Soil Loss Calculation As stated before, the rain erosivity, soil erodibility and the factor slope as elements of the RUSLE equation can be considered as naturally occurring factors determining the sheet and rill erosion processes. Together, they can be considered as the erosion susceptibility or potential erosion or soil loss (tons/ha/year) for the area. GRASS GIS version 5.0 has been used to calculate the erosion from the Bagmati basin. As explain before the rfactor, lsfactor, kfactor, cfactor and pfactor are multiply to get the erosion from the catchment. This calculation is done in r.mapcalc function, which is very powerful function in GRASS GIS for map algebra. The erosion map was calculated by following syntax: mapcalc>erosion = rfactor* lsfactor* kfactor* cfactor * pfactor Figure 5: Potential Erosion Classes in the Bagmati River Basin The value range of erosion in erosion raster map was 0 to 731 t/ha (r.info command). The quantitative output of predicted soil loss was then collapsed into five ordinal classes as shown in table 4. Erosion Class Numeric Range Erosion Potential (t/ha/year) 1 0-1 Very Low 2 1-10 Low 3 10-30 Moderate 4 30-100 High 5 > 100 Very High Table 4: Derivation of the ordinal categories of soil erosion potential
8 Potential Erosion Map For Bagmati Basin Using GRASS GIS 5. Results and Discussion GRASS GIS has been applied to calculate the potential erosion in Bagmati River Basin. The practice factor in this study is assumed to be 1.0, this is why the result called the potential erosion in the catchment. Table 5 shows the area of different potential erosion classes in the catchment. This table shows that nearly 13.4% land of the catchment has very high erosion potential. These areas should be treated as the erosion hazards and the watershed management practices should be adopted in this area to prevent the erosion. Erosion Class Area (km 2 ) % Area 1 797.8 28.4 2 846.1 30.1 3 428.2 15.3 4 358.2 12.8 5 376.6 13.4 Table 5: Area of different Erosion Classes in the Watershed Total potential erosion in the catchment in ton/ha/year is calculated (using r.info command) for each major landuses and presented in Table 6. This table shows that the average potential erosion in the agriculture land is highest (245 t/ha/year), however, the forest has minimum potential erosion (11.5 t/ha/year). The average erosion from the whole catchment is 91.5t/ha/year. The agriculture land generates 84% of the total erosion. If the erosion from agriculture land is minimized then it will minimize the total erosion. The LS factor for the land more than 10 o slope has higher values. The crop factor for agriculture is higher. Therefore, the agriculture land on more than 10 0 slope has higher erosion potential. These areas should be identified and agriculture should be converted to agro-forestry farming to minimise the erosion. Landuse Total Potential Erosion (10 6 Ton) Area Km 2 Potential Sediment Yield (ton/ha) Agriculture 21.10 860.1 245.3 Forest 1.85 1603.5 11.5 Shrub 2.10 179.9 116.7 Others 0.64 163.4 39.1 Total 25.69 2806.9 91.5 Table 6: Potential Sediment Yield in each Landuse % % % Erosion Agriculture Forest Shrubs Other Total Area Classes Area % Area Area Area Area % Km 2 Area Km 2 Area Km 2 Area Km 2 Area Km 2 Area 1 3.2 0.4 695.1 43.3 1.0 0.6 98.5 60.3 797.8 28.4 2 416.3 48.4 308.7 19.3 84.0 46.7 37.1 22.7 846.1 30.1 3 49.2 5.7 353.1 22 8.3 4.6 17.6 10.8 428.2 15.3 4 86.2 10 246.2 15.4 18.5 10.3 7.3 4.5 358.2 12.8 5 305.2 35.5 0.4 0 68.1 37.9 2.9 1.8 376.6 13.4 Total 860.1 1603.5 179.9 100 163.4 100 100 Table 7: Area of different Erosion Classes in different Landuses The erosion classes in different landuse has been also calculated using r.info command and presented in Table 6. This shows that 305 sqkm of agriculture land is prone to very high erosion potential.
Jha Raghunath 9 7. Conclusions In the under developed and developing countries, where the commercial softwares are not affordable by the universities, government agencies and other organizations, GRASS GIS is the best GIS software for any GIS application. The capability of GRASS GIS is similar and some times better than the costliest software of GIS. In this study the GRASS GIS has been applied in the Bagmati River Basin to prepare the erosion hazard map (Figure 5). This map could be used by the government organizations to fix the priority area for the integrated watershed management practice in the catchment. Monitoring of sediment yield has been started this year. Once the sufficient data of sediment is collected then sediment delivery ration can be calculated References [1] Renard, K.G., Foster, G.R., Wessis, G.A., and McCool, D.K. (Coordinators), 1997, Predicting soil erosion by water a guide to conservation planning with the revised universal equation (RUSLE), Agriculture Handbook 703, US Govt. Printing Office [2] Wichmeier W.H., and W. Smith, Predicting rainfall Erosion Losses Losses From Croplands East of Rocky Mountains. USDA Agriculture Research Service Handbook No. 282,1965 [3]. Renard, K.G., Freimund, J.R, Using monthly precipitation data to estimate the R factor in the revised USLE, Journal of Hydrology, 157(1994)287-306,Elsevier Science B.V., 1993 [4] Bergsma E., Provisional rain-erosivity map of the Netherlands. Assessment of Erosion, John Wiley and Sons, page 121-126, 1980 [5] Mitasova, H., J. Hofierka, M. Zlocha, L.R. Iverson, Modeling topographic potential for erosion and deposition using GIS. Int. Journal of Geographical Information Science, 10(5), 629-641. (reply to a comment to this paper appears in 1997 in Int. Journal of Geographical Information Science, Vol. 11, No. 6), 1996 [6] Adapting the RUSLE to model soil erosion potential in a mountainous tropical watershed By Andrew A. Millward, Janet E. Mersey, ELSEVIR, Catena Vol. 38, Page 109 129, 1999