The EURO-CORDEX initiative stands as a pivotal international effort dedicated to refining climate projections for Europe through high-resolution regional climate models (RCMs). This study delves into an evaluation of the ERA-Interim-driven EURO-CORDEX RCM ensemble, scrutinizing the period from 1989 to 2008 to ascertain how effectively these models capture the fundamental spatiotemporal patterns of Europe’s climate. Our analysis focuses primarily on near-surface air temperature and precipitation, utilizing the E-OBS dataset as a benchmark for observational comparison.
Within our ensemble, we examined 17 simulations derived from seven distinct models, operating at grid resolutions of 12 km (nine experiments) and 50 km (eight experiments). To rigorously assess model performance across the European continent, we employed a suite of performance metrics, calculated from monthly and seasonal mean values across eight defined subdomains. Furthermore, we contextualized our findings by comparing them against results from the ERA40-driven ENSEMBLES simulations, providing a broader perspective on the evolution of climate modeling capabilities.
Our investigation affirms the capacity of RCMs to adeptly represent core characteristics of European climate, encompassing its spatial and temporal variability. However, it is crucial to acknowledge that our analysis also pinpoints significant shortcomings in simulation accuracy for specific metrics, geographic regions, and seasons. Temperature biases, when averaged seasonally and regionally, generally remain below 1.5 °C. In contrast, precipitation biases typically fluctuate within a ±40% range. Certain recurring bias patterns emerge, notably a prevalent cold and wet bias across much of Europe throughout the year, contrasted by a warm and dry summer bias in southern and southeastern Europe. These patterns suggest systematic model tendencies that warrant further investigation.
Interestingly, when considering seasonal mean values averaged over extensive European subdomains, we found no conclusive evidence that finer spatial resolution (12 km versus 50 km) consistently translates to enhanced model performance. The bias ranges observed within the EURO-CORDEX ensemble largely mirror those documented in the ENSEMBLES simulations. Nevertheless, we identified tangible progress in certain aspects of model performance, such as a reduced warm summer bias in southern Europe, indicating ongoing refinement in model capabilities relevant to the climate of the 2014 Euro era and beyond.
A noteworthy observation is that the temperature bias variability across different configurations within a single model can rival the variability observed across different models entirely. This highlights the substantial impact of specific choices in physical parameterizations and experimental design on the resultant model performance. By leveraging a set of easily reproducible metrics, this study provides a quantitative measure of the current precision achievable with RCMs in regional climate simulations over Europe. This, in turn, establishes a valuable benchmark for evaluating future advancements in model development, ensuring continued progress in our ability to project and understand European climate dynamics in the years following 2014 euro and into the future.
