Evaluating Weather Research Forecasting and Parametric Wind Models to Enhance Wavewatch3 Hurricane Wave Simulations in the Gulf of Mexico

Hurricane wind fields are highly complex, both spatially and temporally, and are often inadequately represented in existing regional and global datasets. This study evaluated the effectiveness of the Advanced Research Weather Research and Forecasting model (WRF-ARW) and parametric wind models in reproducing hurricane-induced winds. We also examined how these models improve the accuracy of hurricane wave simulations using the Wavewatch3 model in the Gulf of Mexico, focusing on Hurricane Ida (2021). The parametric wind models assessed include the Asymmetric Holland 1980 (H80) and Asymmetric Holland 2010 (H10). Sensitivity analyses were conducted on initial and boundary conditions, as well as WRF model physics, to optimize simulation outcomes. Data sources for initial and boundary conditions included the National Centers for Environmental Prediction Global Forecast System final gridded analysis (FNL) 1-degree (6-hourly), FNL 0.25-degree (6-hourly), and ECMWF Reanalysis v5 (ERA5) with 0.25-degree (hourly) spatial and temporal resolutions. The impact of different physics suites, such as CONUS and Tropical configurations with varying microphysics, radiation parameterizations, and cumulus schemes, was also examined. Produced WRF wind fields, along with H10, H80, and ERA5 winds, were used to force the Wavewatch3 model. The results were compared against buoy and satellite observations and the Ocean Weather Inc. (OWI) wind product. The modeled winds showed better performance than ERA5 in reproducing Ida's wind fields, with H10 outperforming H80 based on buoy, satellite and OWI data. The wave simulations revealed that the H10, H80, and WRF winds better replicated the bulk wave properties than ERA5, although H10 and H80 tended to overestimate the low-frequency energy of the wave spectrum.