Three-dimensional variational data assimilation implemented in numerical modeling for Wuhan torrential rain in July 1998

On July 20～21, 1998, a torrential rain with 24-hour rainfall of 286 mm occurred in Wuhan, Hubei Province of China. In this study, a three-dimensional variational data assimilation system developed at the Fifth-Generation Pennsylvania State University/National Center for Atmospheric Research Mesoscale Model (MM5) along with conventional rawinsonde sounding and surface observation is used for modeling the heavy rain. It successfully incorporates observational data of temperature, humidity and wind, and provides an optimal analysis being in favor of the formation of heavy rain. The 24-hour rainfall predicted by the new scheme is 109 mm rather than 2.5 mm forecasted by control experiment. Furthermore, the forecasted accuracy of location and time-variation of hourly rainfall reproduced by this model is also improved.Keywords：variational data assimilation, torrential rain, numerical modeling.

Three-dimensional variational data assimilation implemented in numerical modeling for Wuhan torrential rain in July 1998

On July 20～21, 1998, a torrential rain with 24-hour rainfall of 286 mm occurred in Wuhan, Hubei Province of China. In this study, a three-dimensional variational data assimilation system developed at the Fifth-Generation Pennsylvania State University/National Center for Atmospheric Research Mesoscale Model (MM5) along with conventional rawinsonde sounding and surface observation is used for modeling the heavy rain. It successfully incorporates observational data of temperature, humidity and wind, and provides an optimal analysis being in favor of the formation of heavy rain. The 24-hour rainfall predicted by the new scheme is 109 mm rather than 2.5 mm forecasted by control experiment. Furthermore, the forecasted accuracy of location and time-variation of hourly rainfall reproduced by this model is also improved.