Assimilation of Precipitation Data to the JMA Mesoscale Model with a Four-dimensional Variational Method and its Impact on Precipitation Forecasts

Two sets of forecast-analysis cycle experiments were performed with and without direct assimilation of precipitation amounts using JMA mesoscale four-dimensional variational data assimilation system (Meso4D-Var). With a devised cost function of precipitation observation, which is derived from the exponential distribution, Meso 4D-Var successfully assimilated precipitation data in various weather situations throughout a one-month experiment period. The result of experiments shows that the precipitation assimilation by Meso 4D-Var improves the model forecasts of both weak and moderate precipitation and it ameliorates a spin-up problem of the model precipitation.