Effect of Remotely Sensed Data on the Performance of a Distributed Hydrological Model: Case Study

Abstract

In this study, a distributed hydrological model (MIKE SHE) was employed in an integrated framework to use remote sensing data, and its advantages over using conventional data were analyzed. The command of 6 Main Canal of the Damodar Irrigation Project, West Bengal, India was chosen as the study area. IRS LISS III and RADARSAT SAR data were processed for generating soil and land use/cover maps, respectively. Thematic layers for topography, groundwater table, Thiessen polygons for rainfall and evapotranspiration, settlement area, and soil were prepared in GIS (ARC-INFO). An application program was developed to transform GIS processed ASCII data into the model input format (matrix data) and also model output to GIS. The model was subsequently calibrated and validated for monsoon season (June to October) of 1999 and 2000, respectively, considering the goodness of fit criteria between observed and simulated groundwater depths. Simulations were done with different combinations of remote sensing (RS) and conventional data setups to study the impact of using RS data on the model calibration and simulation results. The introduction of remotely sensed data shows improvements in the model calibration and eventually in its performance. Statistical results for the RS data integrated model show high Nash–Sutcliffe coefficient (0.85), lower root mean square error (0.31), and closer agreement between the mean simulated