Improving the accuracy of a grid-based distributed hydrological model using slope and river length corrections in a large river basin: case Mekong

Abstract

The effect of the grid cell size on a distributed hydrological model performance was investigated using five grid resolutions (1, 5, 10, 20 and 50 km) in a model for the Mekong Basin in Southeast Asia. Two main factors affecting computation results and related to topography and the grid cell size were identified: the terrain slope decreases and the computed drainage network gets shorter with the increasing grid cell size. The decrease of terrain slope induced a rise of the ground water table, increase in saturation overflow, slight increase in evapotranspiration and slower outflow from soil to river network. The shortening of the drainage network induced more peaked response to high discharge events. To compensate the effect of coarse grid cell size, new correction parameters were introduced to the model. The new parameterisation method improved goodness of fit to measured discharges with all coarse resolution model grids. However, almost the same level of improvement was obtained by modifying existing soil water conductivity parameters. Nevertheless, the proposed parameterisation decreases the dependence of soil water conductivity parameterisation from slope and improves the transferability of model parameterisation between different grid scales.