From GCMs to RCMs: Consistency Assessment of Surface Water and Energy Fluxes

Understanding the interaction between energy and water fluxes at the atmosphere-surface interface is essential for advancing knowledge of climate dynamics and variability. This study evaluates the surface water (including soil moisture) and energy flux components, alongside air temperature, in global and regional climate model (GCM and RCM) simulations  applied in the NUKLEUS project (funded by the German Federal Ministry of Education and Research, BMBF). The analyses aim to enhance the understanding of the consistency of the process representation between RCMs and their driving global climate models.

In this study, biases are assessed for winter and summer between RCMs and their driving GCMs, using ERA5 as a reference to explore model behavior over both land and ocean. Historical and future climatologies under the SSP3-7.0 scenario are examined as well to explore climate change signal patterns. The work addresses two main research questions: (1) How consistent are biases between GCMs and RCMs? (2) Do GCMs and RCMs exhibit similar climate change signal patterns?

The study evaluates nine simulations: three GCMs from the CMIP6 framework (EC-Earth3-Veg, MIROC6, and MPI-ESM1-2-HR) and their corresponding downscaled simulations with two RCMs (COSMO-CLM6 and ICON-CLM, version 2.6.5). Analyses focus on the European domain at 12 km grid resolution, with detailed assessments of the Mediterranean and Mid-Europe regions, which represent different climate conditions.

Our results reveal partly low consistency in biases and climate change signal patterns across GCMs and RCMs, with significant variations both within and between models. These discrepancies are often seasonal- and variable-specific. Notably, the Mediterranean region exhibits stronger biases compared to Mid-Europe, particularly over oceanic areas. Contrary to expectations of bias propagation from GCMs to RCMs, the study identifies cases where biases are amplified or newly introduced in RCMs. This is particularly evident in MIROC6-driven simulations, where the RCMs react in a plausible way to strong biases of the sea surface temperature, inherited from the GCM.

In summary, this study provides a clearer understanding of biases in RCMs and their underlying causes. Furthermore, the study highlights the importance of assessing surface fluxes across different domains to fully capture the complexity of model performance and enhance the accuracy of future climate extreme projections.