Performance Assessment of CMIP6 Global Climate Models over the Imam Turki bin Abdullah Royal Reserve

الملخص

Reliable assessment of climate change impacts at regional scales requires a rigorous evaluation of Global Climate Models (GCMs) against high-quality reference datasets. This study evaluates the performance of five Coupled Model Intercomparison Project Phase 6 (CMIP6) GCMs in simulating daily maximum temperature (Tmax), minimum temperature (Tmin) and precipitation over the Imam Turki bin Abdullah Royal Reserve (ITBA) region in northern Saudi Arabia. Model simulations for the historical period (1985-2014) are assessed using in-situ station observations from five meteorological stations (Al Jouf, Gassim, Hail, Qaisumah, and Rafha) and ERA5 reanalysis data as reference. Model performance is evaluated through annual cycle analysis, distributional characteristics, Taylor diagrams and multiple statistical metrics, including Nash–Sutcliffe Efficiency, coefficient of determination, root mean square error–standard deviation ratio and percent bias.

Observed climate analysis (1985-2023) reveals statistically significant warming and drying trend over ITBA , with annual mean temperature increasing of ~0.68 ^oC/decade and rainfall decreasing by ~6.21 mm/decade indicating an emerging warming-drying climatic conditions over ITBA. Results show that ERA5 consistently reproduces observed seasonal cycles, variability, and distributional characteristics for both Tmax and Tmin and serves as a reliable benchmark. CMIP6 GCMs generally capture the seasonal evolution of temperature but exhibit model-dependent biases, particularly during summer when variability and extremes are largest. INM-CM5-0 and IPSL-CM6A-LR show the most consistent performance for Tmax, while MRI-ESM2-0 and CNRM-CM6-1 perform best for Tmin at several stations. In contrast, precipitation is poorly simulated by all CMIP6 GCMs, with low correlations, negative efficiency scores, large biases, and misrepresented variability, reflecting the difficulty of reproducing episodic rainfall in arid environments. Overall, the findings highlight strong skill for temperature simulations but substantial limitations for precipitation, underscoring the need for careful model selection and bias correction in regional climate impact assessments over arid regions.