Hydrologic model calibration using MODIS-ET data: The impact on predictions at gauged and ungauged locations

Increasing availability of satellite remote sensing data triggered the use of hydrologically relevant satellite-based fluxes and variables towards improved modelling. Importance of innovative and satellite data sources for a better understanding of hydrologic processes has been highlighted in the IAHS scientific assembly’s 23 unsolved problems in hydrology. In this context, the present study investigates the use of satellite ET dataset form MODIS in the physically based semi-distributed model Soil and Water Assessment Tool (SWAT). The study area is Nith river watershed, located in Southern Ontario, Canada. We compare the potential of MODIS ET in improving the performance of SWAT at gauged and ungauged locations of Nith River watershed. Streamflow calibrated SWAT model is used as a benchmark model. The benchmark model results are compared with the MODIS ET only calibrated model results to understand the importance of satellite data in hydrologic model calibration. It is found that the calibration of SWAT model only using MODIS ET data resulted better or similar results to that of streamflow-based calibration. The results show that SWAT model improvement is highly dependent on the input data quality such as the precipitation data and land use data used in the initial model set up. The SWAT model calibrated using MODIS ET improved the soil moisture accounting and crop yield estimation. From our results we conclude that the satellite datasets can be a potential solution to parameter estimation in ungauged basins across the world.