A Study On Assimilation Of Satellite Cloud Image And Its Application On Rainfall Numerical Modeling

It is one of the difficulties of mesoscale-numerical forecasting in rainfall. Data assimilation is an effective approach to improve the numerical forecasts skill. Based on earlier research results, this paper aims at improving the forecast skill for rainfall by experiments, which add satellite infrared cloud image into a forecast model MM5 quantificationally. Statistical regressive method are employed to retrieve temperature and humidity fields in low-level atmosphere, and then assimilated into MM5 by variational technique. Secondly a method is applied to add cloud TBB data to numerical model, which corrects the temperature and amplify the humidity simultaneously. Experiments for the case of the rainfall occurred in the Changjiang River during 9-10,July,2003, are carried out. Numerical experiments results on modeling the rainfall, show that: l)The rainfall field modeled is improved and much similar to real rain field when TBB data is assimiliated into initial fields. The correction methed mentioned above,employed in this paper can assimilate TBB data to numerical model directly, and it has phisical significance clearly other than regressive method also. 2) The experiments verify that correction method gives better results than conventional regressive method in intensively convective zone.