Scaling input data by GIS for hydrological modelling

An analysis of scaling effects is performed to evaluate whether data aggreg ation is a useful regionalization tool or whether it leads to an unacceptab le loss of information. One issue concerns the appropriate resolution of di gital elevation models (DEMs) used to derive geomorphological input paramet ers for hydrological models. In particular, the scale problem of watershed division by a channel network and smaller sub-basins is addressed. The inve stigation involved commercially available data sets with different horizont al and vertical resolutions and systematically aggregated DEMs. A stream ne twork and the contributing subareas were derived from a DEM with a distinct critical support area. By varying this threshold area various watershed co nfigurations were obtained. The sensitivity of surface runoff simulations t o all watershed configurations was studied with synthetic storms and by mea ns of an infiltration excess runoff model. The study revealed that elevation data with different resolutions diverge e normously in landscape representation and in the derived parameters such as slopes, flow directions and channel networks. Coarse DEMs show a smoother terrain and shorter flow paths than highly resolved data. The contributing threshold area controls the extent of the watershed configuration and there fore determines the drainage density. These topographic and geomorphologica l features were used to explain differences in the runoff simulation result s. When watershed configurations with a varying extent of the channel netwo rk were derived from a distinct DEM and then used to simulate surface runof f, the drainage densities of the configurations correlated with the simulat ed runoff volume. A distinct drainage density, however, did not necessarily lead to similar simulation results when different DEMs were used. Since th e hydrological model permits reinfiltration, the runoff volume depends dire ctly on the lengths of the overland flow. Therefore, the mean length of the overland flow paths might to a certain degree be considered as a scaling f actor. Copyright (C) 1999 John Wiley & Sons, Ltd.