From Global to Regional: The Added Value of High-Resolution Dynamical Downscaling for Precipitation in Southwest Asia's Complex Terrain

Southwest Asia (SWA) is a climatically sensitive region where water resources are determined by the complex interactions between the Indian Summer Monsoon and Mediterranean winter systems. Coarse-resolution Global Climate Models (GCMs) have difficulties in capturing the arid-to-semi-arid hydroclimate of the region, which is characterized by high variability and orographically intensified precipitation. This study evaluates the added value of dynamical downscaling in representing mean and extreme precipitation in SWA. We use the Weather Research and Forecasting (WRF) model at a resolution of 10 km, driven by boundary conditions from the Community Earth System Model (CESM v1.2.2). For evaluating the models, we compare the native CESM (~2° resolution), the downscaled WRF simulation, and the ERA5 reanalysis for the common period 1950-2002. Our analysis reveals two outcomes for regional downscaling. First, the downscaled WRF simulation significantly improves the representation of the annual cycle, closely agreeing with ERA5, while the original CESM overestimates precipitation during summer. This overestimation can also be seen in the extreme precipitation values of CESM, especially in the south part of our region. Second, in areas of complex orography, like the Zagros Mountains, WRF tends to exaggerate precipitation compared to ERA5. Spatial differences between WRF and ERA5 precipitation in these complex regions can be attributed to the higher resolution of WRF. The extreme precipitation pattern generally agrees between WRF and ERA5 even though we observe the aforementioned spatial differences. The findings point out that dynamical downscaling can accurate simulate  topographically forced precipitation,  reducing large-scale GCM biases. This offers an important baseline for improved representation of precipitation in complex mountainous regions with low observational data availability, such as SWA.