Obtaining refined Euro-Mediterranean rainfall projections through regional assessment of CMIP6 General Circulation Models

Abstract

Despite the European Mediterranean Region (Euro-Med) being highly vulnerable to climate change the evaluation of General Circulation Models (GCMs) is limited in this region. This research aims to enhance Euro-Med rainfall projections by employing an optimized spatial assessment of 34 CMIP6 GCMs. Firstly, GCMs were interpolated to a common 0.1° grid, matching ERA Land, the reference dataset. Empirical Bayesian Kriging was identified as the most suitable technique, exhibiting the lowest error between raw and interpolated data. After that, annual rainfall interpolated data were compared against ERA5 Land. The performance of GCMs was assessed using five different goodness-of-fit metrics: Normalized Root Mean Squared Error, Nash–Sutcliffe Efficiency, Kling–Gupta Efficiency, Coefficient of Correlation and modified index of agreement. CanESM5, CAS-ESM2-0, UKES-M1-1-LL, HadGEM-GC31-LL, MPI-ESM1-2-HR, and GISS-E2-2-G were identified as the top performers. The Multi-Model Ensemble was created assembling these six GCMs through Random Forest (RFE), since it was found that it discriminates local under or overestimations in GCMs. RFE was used for short-term (2026-2050), medium-term (2051-2075), and long-term (2076-2100) projections under three emission scenarios: SSP1-2.6, SSP2-4.5, and SSP5-8.5. Results indicate a projected decrease in annual rainfall in southern areas, accompanied by more notable declines in summer and autumn. On the contrary, the Alps, particularly the Austrian Alps, show resilience to summer stress, with some regions even showing increases in rainfall. Additionally, there is a noticeable north-south pattern in spring and winter. Finally, anomalies were compared against those reported by IPCC, detecting a potentially drier summer, spring and annual yield, and a wetter winter and autumn. The study provides valuable insights into GCM performance in the Euro-Med, offering a robust foundation for improved future rainfall projections and a reduction in associated uncertainties, crucial for climate change adaptation strategies in the European Mediterranean Region.