Wide-swath synthetic aperture radar (SAR) observes large scale (450 km) ocean surface wind field in all-weather conditions. With the advance of understanding of the conversion of radar backscatter to sea surface wind using the recent developed geophysical model function CMOD-5, it is possible to derive the sea surface wind at very high hurricane wind conditions.

Improved capability in hurricane forecast such as track, wind speed, landfall, precipitation amount and location as well as storm evolution, is still a challenging problem. By taking advantage of the recent development in the SAR wind retrieval at high wind speed condition and the Weather Research and Forecasting (WRF) model, we demonstrate the impact of the SAR-derived wind product on hurricane model initialization and simulation. To achieve the objective, we choose different hurricane cases, e.g., Hurricane Katrina and Rita in 2005. First, we examine SAR-derived hurricane winds with error estimates; Second, the SAR-derived winds are used to improve the initial fields of WRF simulations with the WRF variational data assimilation (WRF-Var) system and investigate their impact on 36-60 hours hurricane forecasting; Finally, we examine the impact of the improved initializations on the forecasting of hurricane track, precipitation fields, strong wind zone, as well as other dynamical structures. Comparisons with other dataset such as wind analyses from Atlantic Oceanographic and Meteorological Laboratory (AOML) are also discussed.