WRF-Based Wind Assimilated With SAR Measurements: With Application of Wave Simulation by WW3 During Tropical Cyclone

The purpose of this study is to assimilate wind derived from synthetic aperture radar (SAR) image with Weather Research and Forecasting (WRF) model during tropical cyclone (TC) and then the TC wind is implemented for wave simulation by WAVEWATCH-III (WW3) model. CyclObs winds derived from Sentinel-1 (S-1) and RADARSAT-2 (R-2) images during 15 TCs are provided by the French Research Institute for Exploitation of the Seas (IFRMER), in which the observations from the Stepped-Frequency Microwave Radiometer (SFMR) are available. Wind speeds from WRF model are compared with SFMR measurements yields a root mean squared-error (RMSE) of 3.48 m/s and a correlation (Cor) of 0.89. Two kinds of wind fields are taken in hindcasting TC wave by WW3 model, i.e., WRF and European Centre for Medium-Range Weather Forecasts reanalysis (ERA-5) and the WW3-simulated significant wave heights (SWHs) are validated against the measurements from Haiyang-2 (HY-2) altimeters and National Data Buoy Center (NDBC) buoys. The statistical result using WRF assimilation with HY-2 measurement yielded a RMSE of 0.78 m with a Cor of 0.91 and a bias of −0.36, which is better than the result by ERA-5, i.e., a RMSE of 0.82 m with a Cor of 0.82 and a bias of −0.67. The significant deceasing of bias is also observed in the comparisons with NDBC buoys. Our work confirms that SAR-derived wind can improve the effectiveness of TC wind simulation by WRF model and improve the under-estimation of simulated SWH forced by ERA-5 wind.