ABSTRACT As a result of climate change, extreme weather phenomena are expected to become more frequent and severe, so adaptation in the future will be essential. Global and regional climate model simulations offer crucial insights into future climate change, forming the basis for planning and implementing adaptation and mitigation strategies. Therefore, evaluating these simulations is essential for generating reliable projections. This study assesses ten historical Euro‐CORDEX simulations using different combinations of Global and Regional Climate Models, validated against the E‐OBS dataset, which is derived from meteorological station measurements across Europe. All data are available at daily temporal resolution and interpolated to a 0.1° regular grid. The study analyses the period 1970–2005. For study areas, 14 plain regions were selected across Europe with an objective three‐step methodology based on elevation and geographical borders. Plain regions were defined as relatively flat areas below 200 m elevation. We calculated 11 climate indices; 8 of these are associated with extreme precipitation events, and for the evaluation we used four statistical metrics designed to assess the performance and capability of the model simulations. Based on these results and metrics, a model performance scoring system was established to rank and compare simulations and identify well‐performing models. This ranking methodology enhances the reliability of climate projections by highlighting models that perform best for specific regions or variables, thereby supporting more targeted and informed adaptation planning through reduced uncertainty and increased confidence in future outcomes. To test the sensitivity of model evaluation to the uncertainty originating from the selected reference data, an additional assessment was carried out in two specific regions using high‐resolution datasets (i.e., CARPATCLIM and ROCADA). Results indicate a clear overestimation of both non‐extreme and extreme precipitation in Euro‐ CORDEX simulations with distinct seasonal biases. A general overestimation is found across the domains in winter, while a substantial underestimation can be identified in the south‐eastern regions. Based on the model performance scores, different metrics yielded varying rankings of the models; thus, it is not possible to set an evident order of model simulations. The use of high‐resolution datasets confirmed the importance of incorporating additional observational data in areas with limited station coverage. Nevertheless, the models are generally capable of reproducing the patterns of historical precipitation conditions, and the ensemble is adequate for analysing the future climate.
Berényi et al. (Sat,) studied this question.