Enhanced methods of data assimilation for simulated hurricane rain rates and wind speeds in a mesoscale model

One of the most enigmatic problems in hurricane forecasting is accurate intensity prediction. Methods of data assimilation in numerical weather prediction are most accurate when the ability to assess a hurricane's characteristics are enhanced by incorporating precise observations into the data assimilation schemes. Enhancing these methods of data assimilation, and incorporating more spatially and temporally dense observations of rain rates and wind speeds is shown to improve the intensity forecast in this study. Synthetic observations from the NASA Hurricane Imaging Radiometer (HIRAD), an airborne microwave remote sensor, aids the study the wind field of a hurricane, specifically observing surface wind speeds and rain rates, in what has traditionally been the most difficult areas for other instruments to study; the high wind and heavy rain regions. One portion of this study builds on past results obtained from utilizing the Krishnamurti technique of physical initialization of rainfall observations, and improves the method for mesoscale data assimilation of rain rates. Reliable data from the inner core regions of a hurricane at such a high resolution and wide swath as HIRAD provides, can be very valuable to mesoscale forecasting of hurricane intensity. The other portion of this study shows how the data assimilation technique of Ensemble Kalman Filtering (EnKF) in the Weather Research and Forecasting (WRF) model can be used to incorporate simulated HIRAD wind speed data into a mesoscale model forecast of hurricane intensity as well. The study makes use of an Observing System Simulation Experiment (OSSE) with a simulated HIRAD dataset sampled during Hurricane Karl (2010) and uses EnKF to assimilate the data and produce improved forecasts for the track and intensity of the hurricane. Comparisons to truth and error metrics are used to assess the data assimilation skill and forecast performance.