Comparative analysis of conceptual rainfall-runoff modeling in Chattisgarh, India

Abstract

Rainfall-runoff modeling is indispensable tools for hydrological research. Runoff estimation isquite influential for watershed development as its value has vital importance for most of the hydrological analysis.The major artifact for the hydrologist is Runoff estimation in the developing countries where the most of the watersheds are ungauged. There are different models for estimating runoff such as SWAT, HEC-HMS, HEC-RAS, MIKE SHE, MIKE NAM 11 etc.The proposed research work focused on the use of Rainfall Runoff Library (RRL) which incorporates five models such as Australian Water Balance Model (AWBM), Tank model, Soil Moisture and Accounting Model (SMAR), Sacramento model and Simhyd model. It has 8 calibration optimizer and 10 objective functions.This study comparedthe AWBM model and Tank model for its relevance in terms of performance and efficiency in Tandula catchment of Chhattisgarh, India.This paper undertakes 20 years daily rainfall, evapotranspiration and runoff data for simulation from 1995 to 2014. Runoff is estimated for 20 year period. The calibration has been carried out from 1995 to 2004 and validation from 2005 to 2014. The training and testing values for AWBM model based on Nash Sutcliffe efficiency and genetic algorithm optimizer were found 0.784 and 0.769 respectively and for Tank model, the values were found 0.837 and 0.791 respectivelyduring calibration and validation. Through the analysis it is revealed that Tank model performed better than AWBM Model in term of Nash Sutcliffe efficiency. This analysis shows a good relation between simulated and observed runoff. This can be used for predicting the runoff for extended period.