distributed hydrological modelland usesuperficial geologyspatial datasets

A distributed hydrological model was applied to a 69 km(2) experimental catchment, Eddleston Water, Scotland, UK. The impact on model outputs of applying progressively simpler representations of spatial variability in land use and superficial geology was assessed. Alternative representations of the spatial distribution of superficial geology and land use produced differences in model outputs. These differences were generally small with the exception of the maximum absolute error (E-max). Inter-model differences were most sensitive to the largest precipitation events. Although variations in superficial geology dominated over those for land use, exceptions were seen in two sub-catchments. These were connected with particularly large variations in land use and/or the small spatial extent of superficial geology. Lower resolution spatial data produced superior model performance in the majority of sub-catchments. This has implications for modelling other catchments, especially in situations where the high-resolution data employed herein are not available.