Improved Analyses and Forecasts of Rapidly Intensifying Tropical Cyclones by Assimilation of Rapid Scan Satellite Winds Using an Ensemble Filter

Accurate analyses and forecasts of rapidly intensifying tropical cyclones have been challenging to produce, partly due to the lack of appropriate observations as well as appropriate data assimilation techniques. The recently developed rapid scan satellite wind vectors from CIMSS provide wind information at hourly intervals and may be valuable to improve the analyses and forecasts of tropical cyclones. The recent emerging ensemble data assimilation techniques provide flow dependent multivariate forecast error covariances which can be useful for the assimilation of the satellite wind observations, especially of the upper level outflow of tropical cyclones. This study examines the impact of assimilating the rapid scan wind observations with an ensemble filter. The wind observations are assimilated in NCAR's WRF/DART ensemble system for two tropical cyclone cases from 2008: Hurricane Ike and Typhoon Sinlaku. Analyses and forecasts of the rapid intensification of the storm are found to be improved in each case. The impact of the wind data on the analysis of the upper-level outflow of the storms is examined.