Sensitivity of Initial Abstraction Coefficient on Prediction of Rainfall-Runoff for Various Land Cover Classes of 'Ton Watershed' Using Remote Sensing & GIS Based 'Rinspe' Model

Abstract

Rainfall, Runoff, Infiltration are most important parameters in water balance equation of any hydrological problems. Among these parameters the real prediction of surface runoff is the crucial task. Many hydrologic models are available to estimate surface runoff from precipitation data. The development of a GIS based model that would make it possible to assess the surface runoff accurately and efficiently. On the basis of this requirement a model called  ‘Runoff, Infiltration and Nonpoint Source Pollution Estimation’ (RINSPE) was developed for estimating runoff, infiltration and nonpoint source pollution in the Ton river catchment based on the different land cover type distribution within the catchment. The Curve Number method for estimating direct runoff from rainstorms is now widely used in engineering design, post-event appraisals, and environmental impact estimation. In the present investigation the Initial Abstraction coefficient in the Curve Number method considered with 2 different scenarios by assuming ‘λ’ value as 0.2 for general conditions and 0.3 for Indian conditions to run RINSPE model. The results reveal that the surface runoff of 860.79 mm occurred with a loss of 62.93 mm in the form of infiltration for scenarios: 1. The most occupied area in the Ton watershed is forest which is having a runoff of 656 mm with 124mm infiltration and a very slight variation in prediction of runoff and infiltration for scenario: 2 for the same land cover classes. In the present study the results reveals that ‘λ’ value is 0.2 of I^a gave minute higher runoff values than ‘λ’ value with 0.3. This experimental investigation makes sure that, higher the initial abstraction lowers the runoff and vice versa.