TitleUncertainties in the Hydrological Modelling Using Remote Sensing Data over the Himalayan Region.
AbstractHydrological modeling is the basis for water management practices that poses a big challenge for Himalayan River Basins due to complex topography and climatology. Lack of good quality input data and suitable modeling tool is another hurdle in a way of hydrological modeling for this region. Therefore, in this study we have highlighted the uncertainties in the hydrological modeling using remote sensed inputs and SWAT model setup over a Himalayan “Gandak River Basin”. Three widely used DEMs SRTM 30m, ASTER 30m and SRTM 90m have been evaluated for the possible error in delineation of river network and basin boundary digitally. Also, we have evaluated error in the amount and frequency of two satellite precipitation estimates: TRMM 3B42V7 and CMROPH. The error in the SWAT outputs due to change in modeling routine are assessed by considering sixteen set of DEM grid sizes ranging from 40 to 1000m and that has been resampled with three distinct methods nearest neighborhood (N), bilinear (B) and cubic convolution (C). After that, we have carried out the examination for error in the SWAT outputs due the change in the model setup and parameters. We have also accessed the suitability of TRMM 3B42V7 products to simulate the discharge for various rainfall intensity classes. The results from the delineation analysis shows that the SRTM 90m DEM does well for the study area however, TRMM 3B42V7 precipitation estimate shows better statistical scores than CMROPH. SWAT model setup with change in the spatial inputs (different set of DEM grid size) reveals that DEM grid size has significant impact on the model outputs at different spatial and temporal scale. Analysis for different model setups shows that the SWAT model with distributed elevation band setup does better to simulate streamflow and snowmelt for this region.