Improvement of drag coefficient parameterization of WAVEWATCH-III using remotely sensed products during tropical cyclones

Abstract

In this study, the parameterization of the drag coefficient is improved using remotely sensed wind and wave products during tropical cyclones (TCs) and implemented for wave simulation using a third-generation numeric model, namely, WAVEWATCH-III (WW3). The significant wave height and peak wavelength from SWIM(Surface Waves Investigation and Monitoring) measured wave spectra are collocated with wind speeds derived from the Haiyang-2 (HY-2) constellation, Soil Moisture Active–Passive (SMAP) radiometer, and Advanced Microwave Scanning Radiometer-2 (AMSR2) in 2019 − 2023. Cyclonic winds at a spatial resolution of 0.25° and intervals of 6 h are obtained by fusion remotely sensed winds from the HY-2 scatterometers and radiometers from SMAP and AMSR2. The parameterization of the drag coefficient during TCs is improved using a matching dataset for 100 TCs consisting of fusion winds and SWIM-measured wave spectra. The significant wave height (SWH) is simulated by WW3 using the optimized parameterization in the several input/dissipation source terms, i.e., ST2, ST3, ST4 and ST6. It is found that the accuracy of simulated SWH using optimized parameterization in switch ST2, ST3 and ST4 is worse than that using optimized parameterization in switch ST6. Moreover, taking two parameterizations in switch ST6 of WW3, validation against the measurements from altimeters and the SWH calculated by SWIM-measured wave spectrum during other 20 TCs yields a root mean square error (RMSE) of 0.60 m, and a correlation coefficient (COR) of 0.85 by optimized parameterization, which are better than the 0.68 m RMSE, and 0.80 COR obtained using existing parameterization. Furthermore, the variation in the bias (altimeter minus WW3) shows that the larger bias (> 2 m) at wind speed > 20 m/s or SWH > 6 m is significantly improved using the optimized parameterization formula and fusion remotely sensed winds.