Simulation of Complex Mountain Wind Field during Icing Period Based on WRF-LES and High-Resolution Terrain

Under modern climate change, accurate wind field prediction is crucial for wind energy assessment and wind farm planning. Guizhou’s Meihuashan area has abundant wind resources, but its complex terrain affects the accuracy of traditional wind speed prediction models. This study evaluates the performance of WRF model’s boundary layer parameterization schemes (MYJ, BL, MRF) through wind speed simulation experiments. A comparison of wind speed predictions and measured data from February 1st to 4th, 2024, shows the MYJ scheme is optimal with a deviation of $4.12 \mathrm{~m} / \mathrm{s}$. To refine wind field simulation, large - eddy simulation (LES) technology is introduced as a micro - scale simulation tool, with the MYJ scheme as the basis for subsequent experiments. The experimental setup includes multi - layer nested grids to enhance spatial resolution and Gaussian smoothing to address numerical instability from terrain changes. Compared to WRF’s default terrain data, 90 m resolution terrain data greatly improves geographic information detail and wind speed forecast evaluation indicators, reducing simulation deviation to $2.74 \mathrm{~m} / \mathrm{s}$. This study successfully identifies the optimal wind speed prediction model configuration for Meihuashan wind farm, confirms the positive role of LES and high - precision terrain data in improving prediction accuracy, and offers technical support for complex mountain meteorological forecasting and weather models in early warning of power disasters like icing.