TheGSIcapability to assimilateTRMMandGPMhydrometeor retrievals inHWRF

Abstract

Hurricane forecasting skills may be improved by utilizing increased precipitation observations available from the Global Precipitation Measurement mission (GPM). This study adds to the Gridpoint Statistical Interpolation (GSI)capability to assimilate satellite‐retrieved hydrometeor profile data in the operational Hurricane Weather Research and Forecasting (HWRF)system. The newly developed Hurricane Goddard Profiling (GPROF)algorithm produces Tropical Rainfall Measuring Mission (TRMM)/GPMhydrometeor retrievals specifically for hurricanes. Two new observation operators are developed and implemented inGSIto assimilate HurricaneGPROFretrieved hydrometeors inHWRF. They are based on the assumption that all water vapour in excess of saturation with respect to ice or liquid is immediately condensed out. Two sets of single observation experiments that include assimilation of solid or liquid hydrometeor from HurricaneGPROFare performed. Results suggest that assimilating single retrieved solid or liquid hydrometeor information impacts the current set of control variables ofGSIby adjusting the environment that includes temperature, pressure and moisture fields toward saturation with respect to ice or liquid. These results are explained in a physically consistent manner, implying satisfactory observation operators and meaningful structure of background error covariance employed byGSI. Applied to two real hurricane cases,Leslie(2012) andGonzalo(2014), the assimilation of the HurricaneGPROFdata in the innermost domain ofHWRFshows a physically reasonable adjustment and an improvement of the analysis compared to observations. However, the impact of assimilating the HurricaneGPROFretrieved hydrometeors on the subsequentHWRFforecasts, measured by hurricane tracks, intensities, sizes, satellite‐retrieved rain rates, and corresponding infrared images, is inconclusive. Possible causes are discussed.