Changing Aerodynamic Roughness in WRF Reduces Bias and Improves Accuracy in Near‐Surface Wind Simulations

Abstract Accurate simulation of near‐surface wind speed is crucial for applications in renewable energy, dust emission, and air quality modeling. However, the Weather Research and Forecasting (WRF) model systematically over‐estimates 10 m wind speed due to simplified representations of surface roughness and momentum exchange. Here we incorporate into WRF a physically based changing aerodynamic roughness (CAR) scheme using satellite albedo, to dynamically adjust near‐surface wind friction and momentum exchange. Simulations during 2023 over China show that CAR substantially improves near‐surface wind speed estimates, reducing the normalized mean bias from 14.6% to 2.2%. The improvements are pronounced over cropland, shrubland, grassland, and bare land. Limited improvements occur in closed canopies (forests and urban areas) because CAR poorly represents skimming flow. Nevertheless, bridging satellite observations with boundary‐layer dynamics, the CAR scheme provides a more realistic representation of surface–atmosphere momentum exchange, offering enhanced reliability for climate, air‐quality, and wind‐resource studies.