On the sensitivity and robustness of Austrian-wide water balance components & groundwater recharge: A regional-scale evaluation of objective functions in calibration

Regional water resource management requires adaptation measures to cope with a changing environment and climate. Large- and regional-scale hydrological models can provide information on water cycle components and their future projections for designing these measures. Here, the plausible and consistent spatial distribution of the simulated water balance components is essential, not only when communicating the results to the relevant stakeholders.

Spatially consistent simulation results depend on the spatially distributed parameters estimated for the hydrological model. This task, however, remains a challenging step in the model set-up, especially for larger modeling domains with varying hydrometeorological conditions like Austria, ranging from high alpine areas with high rainfall sums, snow, and glaciers to low-lying areas, with semi-arid conditions exhibiting negative climatic water balances.

The main objective of this study is to analyze the impact of different objective functions (OFs) on the simulated water balance components of a regional hydrological model. The distributed rainfall-runoff model COSERO is set up for the area of Austria (89 000 km²) with a monthly temporal resolution and a target spatial resolution of 1 x 1 km². Initially estimated spatially distributed model parameters are optimized using different OFs, e.g., Nash-Sutcliffe efficiency (NSE), log-transformed NSE, Kling-Gupta efficiency (KGE), and combinations thereof.

The resulting simulations are analyzed based on the spatial distribution of simulated runoff, evapotranspiration, and groundwater recharge across the study area. Furthermore, time-series analysis of the water cycle components is performed, and selected statistical characteristics are derived.