Improving consistency of WRF-UCM-based typhoon wind field simulations with field observations: A sensitivity analysis of model parameters

Accurately simulating near-surface wind fields in coastal urban areas during typhoon events is essential for wind-resistant structural design and safe operation of cities. To identify suitable mesoscale model parameters that enhance the typhoon simulation accuracy, this study conducts a series of numerical experiments using the Weather Research and Forecasting (WRF) model coupled with the Urban Canopy Model (UCM). Thirteen different model configurations are tested across six historical typhoon events, systematically evaluating the sensitivity of simulation results to physical parameterization schemes, vertical grid resolutions, and nudging strategies. The simulated typhoon track, intensity, spatial scales, and near-ground wind fields are quantitatively evaluated against datasets and field observations. The results show that the BouLac planetary boundary layer combined with the revised MM5 Monin-Obukhov surface layer performs best overall. Vertical grid resolutions (VGRs) have limited impact, though finer near-ground layering is preferable to meet the requirements of the Building Effect Parameterization (BEP) UCM. Nudging improves accuracy but requires careful application: the grid nudging enhances near-ground wind field reproductions but introduces abnormal result oscillations and suppresses turbulence, while spectral nudging applied to all layers yields more balanced and stable performance. These findings provide practical guidance for modeling typhoon wind fields in urban environments. © 2026 Elsevier B.V.