Changing environmental conditions might have severe impacts on future flood frequencies and water availability. In order to develop mitigation strategies by carrying out impact studies, lumped hydrological models might not suffice anymore and the need for flexibility in spatial resolution arises. Recent developments already allow to change the grid size of hydrological models depending on the needs and to use different spatial resolutions in parallel for different applications. These developments, however, still face some challenges such as the following: how to avoid strong or sudden changes in simulation results when level of spatial details is changed and how to limit the problem of overparameterization/equifinality while increasing the spatial resolution. We developed a methodology to partially meet these challenges by building parsimonious distributed conceptual hydrological models with a flexible spatial resolution. The methodology starts with calibrated lumped conceptual models. Parameters are then spatially disaggregated based on spatially variable catchment properties, introducing only few additional calibration parameters. The approach is demonstrated for three different conceptual rainfall‐runoff models for two medium‐sized Belgian catchments. The resulting spatial models are found to be accurate with respect to the observations and consistent with the lumped conceptual model results, when evaluated for peak, low, and cumulative total runoff and subflows at downstream and internal gauging stations. A climate change impact analysis further shows consistency of lumped and spatial model results for changing conditions.

No abstract

No abstract