An innovative approach of GSSHA model in flood analysis of large watersheds based on accuracy of DEM, size of grids, and stream density

Abstract

AbstractDistributed modeling approach may have much better performance and accuracy compared with lumped-parameter hydrologic models. The main goals of this research are: investigating the possibility of combining distributed hydrological models with an one-dimensional hydraulic model and simulating waterways in large watersheds with limited hydrological and hydraulic data. Then performing sensitivity analysis on different parameters in order to identify the parameters containing the major influences on results. In the current research, an innovative approach in Gridded Surface Subsurface Hydrologic Analysis (GSSHA) model, the cross-sections of all 414 waterways in the 3450 km2 Karvandar watershed, used for flow routing calculations, are uniquely extracted. Then, the effect of three essential factors are evaluated. These factors are accuracy of the digital topographic model, cell size of grid network, and density of streams, on the results of GSSHA model simulations. This watershed is located in southeastern Iran, has a dry climate with limited available hydrological data. Results showed that peak discharges obtained from the GSSHA model, developed based on a DEM with a spatial resolution of 12.5 m, are slightly (< 4%) lower than the corresponding values ​​in the GSSHA model with a 30 m DEM resolution. This fact confirms that the use of the topographic model with a lower spatial resolution has no substantial effects on the accuracy of simulation. Also, the peak discharges increased significantly (44% to 57%) by increasing the density of waterways in the GSSHA model. Furthermore, results showed that peak discharge obtained from three models with grid cell sizes of 100, 150, and 200 m (base model), are close together. Comparing with two models of coarser grids (250 and 300 m), significant differences observed, which indicated that the grids larger than 200 m could induce substantial errors in results.