Automated Approach for Rainfall-Runoff Model Generation

Abstract

Manually constructing hydrological model descriptions for urban areas tends to be laborious due to the detailed mosaic land cover and the required high-resolution model setup. Here, the performance of a novel automated subcatchment generator with a detailed DEM-based surface flow routing is assessed against observations and manually constructed models. In general, the auto-generated models perform well against observations and comparably to manually constructed models regardless of the detail of land cover information input. The introduced inter-subcatchment connections may require previously acquired model parameters to be re-calibrated. This is due to the calibrated parameters in manually constructed models, even with high-resolution landuse, partly compensating for missing flow routes due to the larger scale used in subcatchment description.