GRID-SIZE EFFECTS ON SURFACE RUNOFF MODELING

The effects of grid-cell size on surface runoff modeling using a distributed hydrologic model are examined. The raster-based hydrologic CASC2D model is applied to two watersheds in northern Mississippi. Observed streamflow for Goodwin Creek (21 km(2)) is used in calibrating the model's infiltration, runoff, and routing parameters. The results are extended to Hickahala-Senatobia (560 km(2)), which has similar physical characteristics. The applicability of CASC2D in simulating rainfall-runoff processes is tested using square grid-cell sizes ranging from 127 to 914 m. Event-based simulation results indicate that coarser grid-cell resolutions can be used in rainfall-runoff simulations as long as parameters are appropriately calibrated. It is demonstrated that the primary effect of increasing grid-cell size on simulation parameters is to require an increase in overland and channel roughness coefficients. The concept of a watershed time-to-equilibrium is used in evaluating the runoff response of overland and channel cells. At different grid-cell sizes, raster maps generated within a GRASS GIS environment provide information regarding the spatial distribution of drainage area, time-to-equilibrium, and equilibrium discharge on Goodwin Creek and Hickahala-Senatobia. It is shown that at increasing grid-cell size, the statistical distribution of drainage areas associated with overland cells is significantly increased, while the statistical distribution of drainage areas associated with channel cells is hardly affected by grid-cell size. It is also demonstrated that flow on overland cells is more sensitive to changes in grid-cell size than is channel flow. Since channel flow dominates the shape of the hydrograph in large watersheds, the results indicate that coarse grid sizes can be used for rainfall-runoff simulations on large watersheds. The results also indicate that coarser grid sizes will be more appropriate when simulating events of high intensity or of long duration.