Evaluating Hydrostatic vs. Non-Hydrostatic Dynamics in RegCM5: A 50-Year Simulation of Near-Surface Wind Speed over the Yellow River Basin

Accurate simulation of near-surface wind speeds over complex terrain is essential for wind energy assessment, ecological conservation, and hydro-climatic change studies. This study presents a 50-year (1971–2020) evaluation of the RegCM5 regional climate model over the Yellow River Basin, China. Three dynamical configurations are systematically compared: the traditional hydrostatic core (Model 1), the MM5-like non-hydrostatic core (Model 2), and the newly introduced Moloch non-hydrostatic core (Model 3). Driven by ERA5 reanalysis at a 25-km resolution, the simulations are validated against high-density observations (CN05.1). Results indicate that while all configurations capture the spatial climatology, they exhibit a systematic positive wind speed bias (+ 1.3 to + 2.8 m s⁻¹), particularly in the rugged Upper Reaches due to sub-grid topographic smoothing.. However, Model 3 (Moloch) consistently outperforms the other configurations, achieving the lowest bias magnitude and the highest distributional fidelity (Perkins Skill Score). Regarding long-term trends, all models reproduce the interannual variability driven by large-scale circulation but fail to capture the magnitude of the observed "terrestrial stilling", highlighting a critical need to incorporate time-varying land use forcing in future multi-decadal experiments. Crucially, the study reveals a decisive disparity in computational efficiency. While the traditional non-hydrostatic core (Model 2) is computationally expensive due to strict CFL stability constraints (dt = 25 s), the Moloch core (Model 3) maintains numerical stability with a significantly larger time step (dt = 200 s). This capability results in a 58% reduction in total runtime compared to Model 2 and a 17% reduction compared to the hydrostatic baseline, without compromising physical accuracy. Consequently, the Moloch core emerges as the optimal configuration for long-term regional climate downscaling over complex terrains, effectively bridging the gap between high physical fidelity and numerical efficiency.