Home Articles Curve Number Estimation for Watershed using Digital Image of IRS-1D LISS-III

Curve Number Estimation for Watershed using Digital Image of IRS-1D LISS-III

V.K.Pandey, S.N.Panda
Department of Agricultural and Food Engineering,
Indian Institute of Technology
Kharagpur-721302 (WB), India.

S.Sudhakar
National Remote Sensing Agency,
Department of Space, Govt. of India
Hyderabad-500037 (AP), India.

Introduction
Watershed is all the land and water area, which contributes runoff to a common point. In India the need of accurate information on basin runoff and silt yield has felt during the past two decades along with the acceleration of the watershed management for conservation and development of soil and water resources. The hydrologic behaviors of watershed play an important role in water resources planning and management (Shin et al., 2002). Advances in computational power and the growing availability of spatial data have made it possible to accurately describe watershed characteristics for modeling of watershed hydrology. Recent studies (Sanware et al., 1988, Prasad et al., 1992, Sharda et al., 1993, Schumann et al., 2000, Saxena et al., 2000) reveled that Remote Sensing (RS) and Geographic Information System GIS techniques are of grate use in characterization and prioritization of watershed areas. Land use/land cover is the category in which RS has made its largest impact and comes closest to maximizing the capability of this technology. The degree of various categories of agricultural, forest and other land covers can be determined accurately through RS (Garbrecht, et al. 2001). The large amount of spatially detailed information derived from digital images, ground surveys, digital terrain models and handled within a GIS, offers new opportunities for watershed parameterization. One of the options for use of RS and GIS is to improve the estimation of watershed parameters like Curve Number for a drainage basin with widely used SCS model from its land use data and digitized soil map.

In general land use/land cover accuracy is directly related to the spatial resolution of the sensors. Satellite data helps in deriving CNs for large drainage basins (Still and Shih, 1985; Kumar, 1997). The use of IRS-1D LISS-III digital image of 23.5 m resolution may further provides appreciable accuracy of classification. In the present investigation an attempt has been made to establish the SCS Curve Number from Indian remote sensing digital database for Banikdih watershed.

Materials and methods
The Banikdih watershed that is a part of catchment area of Gowai River lies in Bokaro district of Jharkhand and some part in Purulia district of West Bengal states in Eastern India. It is located between 86o 16′ 0″ & 86o 19′ 0″ E longitude and 23o 24′ 5” & 23o 30′ 0″ N latitude (Fig. 1). The Survey of India topographic maps No. 73-I/7, 73-I/6, and 73-I/3 covers the entire watershed. The average annual rainfall in the area is 1250 mm more than 80% of which is occurred during the monsoon months (June to September). The daily mean temperature ranges from a maximum of 44.0 oC to a minimum of 4.0 oC. The daily mean relative humidity varies from a minimum of 40% in the month of April to a maximum of 95% in the month of the July, during which the area receives its major amount of rainfall. Region falls within sub-tropical climate with alternate dry and wet periods.

Figure 1 Location and sub-watershed map
Topographic maps (1:25000 scale) were collected from Survey of India (SOI) office, Kolkata. Soil resource data and related maps (1:25000 scale) were collected from All India Soil and Land Use Survey (AISLUS), Department of Agricultural and Cooperation, Govt. of India, Kolkata. Digital data of Indian Remote Sensing Satellite IRS-1D LISS-III, Path-106 and Row-55, of 23.5m spatial resolutions pertaining to 23rd October, 2000 was obtained from National Remote Sensing Agency (NRSA), Govt. of India, Hyderabad.

Landuse/landcover classification map
The digital image (IRS-1D LISS-III) of the watershed was registered with the original satellite scene and the mask of the image was prepared. The enhancements and Histogram Equalization were applied to further processing of image. False color composite (FCC) of the scene was prepared and multispectral classification of image data was carried out applying supervised classification using ERDAS IMAGINE-8.4 software. Maximum Likelihood Classifier (MLC) algorithm was used for classification of the land use. MLC is based on the estimated Gaussian probability density functions for each of the reference classes. Overall accuracy of classification and Kappa Coefficient was found to be 86.950 % and 0.88 respectively. Maximum likelihood report for land use classification for watershed and classified output of the image is presented in Table 1 and Fig. 2 respectively.

Figure 2 Land use map of watershed
Table 1 Maximum likelihood report for land use classification for watershed

Land use/land cover classes No of pixels Area (ha) % Image
Water body 4338 229.89 2.598
Lowland paddy 68127 3610.26 40.682
Upland paddy 36125 1914.37 21.572
Fallow land 25781 1366.13 15.395
Upland crops * 9165 485.77 5.473
Settlement 4146 219.65 2.475
Mixed open forest 7712 408.72 4.605
Waste land 12053 638.63 7.179
Null 17 8873.42 0.01
* Non-paddy crops

Soils
The soil map of 1:25000 scale was traced, scanned and exported to ERDAS IMAGINE. Image to image registration was performed using the registered topographic maps. The scanned map was loaded in ARC/INFO and boundaries of different soil textures were digitized carefully and the polygons representing various soils were assigned and flood filled with different colors for identification (Fig. 3). Different gray level values were assigned to different soil texture while preparing the maps. Four hydrologic soil groups, A, B, C, and D, were considered for the basic classification of soils of the watershed. The soils of group A are of low runoff potential, high infiltration rate, high rate of water transmission, the soils of group B are of moderate infiltration rate, moderately well drained to well drained, the soils of group C are of moderately fine to moderately coarse textures, moderate rate of water transmission and the soils of group D are of slow infiltration and high runoff potential.

Figure 3 Soil map of watershed
The AVSWAT (Arc View SWAT) interface model (Neitsch, 2001) that is linked with raster based GIS to facilitate the input of the spatial data such as land use map, soil map and digital elevation model was used for delineation of watershed. The land use /land cover and soil maps were loaded in AVSWAT to clip these layers. The area of the map grids that falls outside the watershed boundary was removed and map grids were reclassified and resampled at the DEM map grid resolution. Finally areas of different land use class and soil combinations were obtained for individual watershed.

Soil Conservation Service (SCS) Model
The SCS runoff equation (SCS, 1972) that is an empirical model was developed to provide a consistent basis for estimating the amounts of runoff under varying land use and soil types (Neitsch, 2001).

where Qsurf is the accumulated runoff or rainfall excess (mm), Rday is the rainfall depth for the day (mm), Ia is the initial abstractions which includes surface storage, interception and infiltration prior to runoff (mm), and S is the retention parameter (mm). The retention parameter varies spatially due to changes in soils, land use, management and slope and temporally due to changes in soil water content. The retention parameter is defined as

where CN is the curve number for the day. The initial abstractions, Ia, is commonly approximated as 0.2S and equation 1 becomes

The SCS curve number is a function of the soil’s permeability, land use and antecedent soil water conditions. The daily curve number value adjusted for moisture content is calculated by rearranging equation 2 and inserting the retention parameter calculated for that moisture content:

where S is the retention parameter calculated for the moisture content of the soil on that day.

For Indian conditions this equation has been modified (Hand book of hydrology, 1972) as follows

where Q is actual direct runoff (cm), P is the total storm rainfall (cm), S is the potential maximum retention (cm). The weighted value of CN for individual watershed was determined as by the following equation

where CN is curve number if individual land use and soil group, A is the area of respective combination and TWA is the total area of watershed. The average condition regarding runoff potential (AMC II) was considered in the present study for determination of CN values. The CN values for AMC II can be converted in to CN values for AMC AMC III and I by using available conversion factors (Suresh, 1997).

Results and Discussions
The study watershed was delineated in to eight sub watersheds considering topographical parameters derived from Digital Elevation Model. On the basis of classification of digital image (IRS-1D LISS-III pertaining to 23rd October, 2000) of study watershed, total eight land use classes namely, water body (2.60%), lowland paddy (40.68%), upland paddy (21.57%), fallow land (15.40%), upland crop (5.47%), settlement (2.48%), mixed open forest (4.61%) and wasteland (7.20%) were identified. Among the land uses classes, Paddy (rice) was found to be the dominant land use. The area covered by four soil textural classes namely Sandy loam (58.98%), Loamy sand (11.55%), Sandy clay loam (22.54%) and Clay loam (6.92%) was determined for each sub watershed. The maximum area of watershed comes under sandy loam textural category. Further the land use and soil classes were distributed sub watershed wise and presented in Table 2. The Sub watershed wise individual combination of land use and soil and its distribution presented in Fig. 4 reveled that the maximum area of the watershed comes under paddy(rice)/sandy loam category. The USDA curve number table (Tripathi, 1999) modified for Indian conditions (AMC II and I = 0.3S) was used for the determination of the curve number for individual sub watersheds based on the hydrological soil groups and land use classes of respective areas. The CN value of individual sub watershed comes to be 78.64, 74.58, 76.34, 77.79, 74.52, 78.39, 78.87 and 76.84. The weighted curve number for the whole watershed was found to be 76.99.

Figure 4 Area coverage of different land use and soil combinations under each sub watershed
Table 2 Sub watershed wise distribution of land use and Soil

Sub watershed Waste land Water body Upland crops Mixed open forest Settlement Fallow land Paddy (Rice) Sandy loam Sandy clay loam
SWS1 Area, ha 68.71 28.02 42.22 37.82 44.02 156.91 668.81 577.22 351.59
%, HWS 0.77 0.32 0.48 0.43 0.50 1.77 7.54 6.50 3.96
%, SWS 6.57 2.68 4.03 3.61 4.21 14.99 63.91 55.16 33.60
SWS2 172.80 20.92 100.81 32.81 8.28 113.61 341.36 694.87 78.89
%, HWS 1.95 0.24 1.14 0.37 0.09 1.28 3.85 7.83 0.89
%, SWS 21.86 2.65 12.75 4.15 1.05 14.37 43.18 87.89 9.98
SWS3 88.14 32.26 73.79 31.20 25.27 243.34 594.32 712.70 124.69
%, HWS 0.99 0.36 0.83 0.35 0.28 2.74 6.70 8.03 1.41
%, SWS 8.10 2.96 6.78 2.87 2.32 22.36 54.61 65.49 11.46
SWS4 49.74 29.38 34.89 73.60 13.20 129.75 859.16 633.33 372.02
%, HWS 0.56 0.33 0.39 0.83 0.15 1.46 9.68 7.14 4.19
%, SWS 4.18 2.47 2.93 6.19 1.11 10.91 72.22 53.23 31.27
SWS5 84.96 30.78 76.43 55.25 29.19 193.23 626.73 761.43 57.58
%, HWS 0.96 0.35 0.86 0.62 0.33 2.18 7.06 8.58 0.65
%, SWS 7.75 2.81 6.97 5.04 2.66 17.62 57.15 69.44 5.25
SWS6 49.79 30.80 37.99 97.31 47.41 140.01 1098.62 690.33 537.09
%, HWS 0.56 0.35 0.43 1.10 0.53 1.58 12.38 7.78 6.05
%, SWS 3.32 2.05 2.53 6.48 3.16 9.32 73.15 45.96 35.76
SWS7 19.07 14.55 28.84 54.28 33.09 144.52 819.40 593.55 308.28
%, HWS 0.21 0.16 0.32 0.61 0.37 1.63 9.23 6.69 3.47
%, SWS 1.71 1.31 2.59 4.87 2.97 12.98 73.57 53.29 27.68
SWS8 105.42 43.19 90.79 26.45 19.19 244.75 516.26 571.38 170.60
%, HWS 1.19 0.49 1.02 0.30 0.22 2.76 5.82 6.44 1.92
%, SWS 10.07 4.13 8.67 2.53 1.83 23.38 49.31 54.58 16.30
SWS = Sub Watershed, HWS = Whole watershed,

Conclusions
In the present study the methodology for determination of Curve Number for Banikdih watershed using an integrated approach of remote sensing, GIS and SCS model has been described. This approach may be applied in other Indian watersheds for planning of conservation measures and developing effective management scenarios.

Acknowledgement
The facilities and technical supports provided by the Regional Remote Sensing Service Centre (RRSSC), Department of Space, Govt. of India, Kharagpur are gratefully acknowledged.

References

  • Garbrecht, J., Ogden, F.L., DeBarry, P.A. and Maidment, D.A. (2001). GIS and distributed watershed models. I: Data Coverages and Sources. Journal of Hydrologic Engineering, Nov./Dec., pp 506-512.
  • Hand book of Hydrology (1972). Soil Conservation Department, Ministry of Agriculture, New Delhi.
  • Kumar, P., Tiwari, K.N. and Pal D.K. (1997). Establishing SCS runoff curve number form IRS digital database. J. Indian society of remote sensing, Vol 19, No 4, pp 246-251.
  • Neitsch, S.L., Arnold, J.G., Kiniry, J.R., & Williams, J.R. (2001). Soil and Water Assessment Tool, theoretical documentation and user’s manual, Version 2000. Soil and Water Research Laboratory, Agricultural Research Service, Grassland, 808 East Black Land Road, Temple, Texas.
  • Prasad, K.S.S., Gopi, S. and Rao, S.R. (1992). Demarcation of priority watersheds in Mahboob Nagar district, A.P. using remote sensing techniques. In Remote sensing application and GIS, recent trends, Tata Mcgrow-Hill Publishing Co. Ltd., New Delhi, pp 180-186.
  • Saxena, R.K., Verma, K.S., Chary, G.R., Shristava, R. and Barthwal, A.K. (2000). IRS-1C data application in watershed characterization and management, Int. J. of Remote Sensing, 21(17), pp 3197-3208.
  • Sanware, P.G. Singh, C.P. and Karale, R.L. (1988). Remote sensing application for prioritization of sub watersheds using sediment yield and runoff indices in the catchment of Masano Barrage (Sahibi), UNDP/FAO project tech. Report – 13 of Remote Sensing Centre, AIS LUS, Govt of India, New Delhi.
  • Schumann, A.H., Funke, R. and Schultz, G.A. (2000) Application of geographical information system for conceptual rainfall runoff modeling, J. of Hydrology 240 (2000) 45-61
  • Sharda D. Ravi Kumar M.V. Venkatatatnam, L and Malleswara Rao, (1993). Watershed prioritization for soil conservation – A GIS approach, Geo Carto International (1): 27-34.
  • Shin-Min, C., Ting-Kuei, T. and Stephan, J.N. (2002). Hydrologic regionalization of watersheds. I: Methodology Development. J. of Water Resources Planning and Management. Jan./Feb. 2002/3.
  • Soil Conservation Service (1972). Section 4: Hydrology In National Engineering Handbook. SCS.
  • Suresh, R. (1997). Soil and Water Conservation Engineering. Standard Publishers Distr., Delhi, pp 48-51.
  • Still, D.A. and Shih, S.F. (1985). Using Landsat data to classify land use for assessing the basin wise runoff index. Water Resource Bulletin, 21:931-940.
  • Tripathi, M.P. (1999). Hydrological Modeling of Small Watershed, Unpublished Ph.D. Thesis, Department of Agricultural and Food Engineering, Indian Institute of Technology, Kharagpur.