Grid size effects on a distributed water quantity-quality model in a hilly watershed

Abstract

Effects of different grid sizes (50 m, 100 m, 150 m, … , 500 m) were analyzed for both watershed attributes and estimation by a distributed model for flow rate and water quality. Model was evaluated in terms of estimation differences and on the values of some parameters at different grid sizes. Grid resizing showed significant changes in the physical attributes related with topography, but attributes such as land cover and geology were not much affected. Model showed increasing flow rates estimation with increasing grid size when the same parameter values were used. Such increasing pattern was observed for root mean square error, magnitude of peak runoff points, and time for the peak recession. Observed differences did not show such variation patterns in case of water quality estimation at different grid sizes. The observed differences in the estimated value by the model could have been derived mainly from less number of grids required to travel, increased proportion of channel grids and smaller outlet height at coarser grid sizes. These estimated differences could be adjusted by changing values of some of the model parameters for each grid size within a narrow range.