A Study Of Mesoscale Numerical Simulation Of Cloud Drift Wind Data Assimilation

These years, the accuracy of numerical weather prediction has been greatly promoted with the development of various mesoscale numerical models which could replay the development of various weather systems well. And the quality of NWP mainly depends on the exact initial field which could be improved to enhance the simulation and prediction ability of model. But it is hard to forecast the mesoscale system for the low space-time resolution of conventional observations. With the development of meteorological satellite and radar technology, non-convection data has been used abroad in operation and research, and become the important data source for meteorological data assimilation. Now as one of the productions of stationary meteorological satellite cloud drift wind (CDW) data apply lots of upper-wind data and could improve the initial field of mesoscale numerical model effectively.On the basis of using the mesoscale numerical model MM5(V3.6) and the FY2C cloud drift wind data inversed by National Satellite Meteorological Center, The U, V and T, which belong to infrared window channel and water vapour absorption channel, are assimilated into the initial field by using nine experiments. And the numerical simulation is conducted in the process of a heavy rain occurred in the mid- and low-reaches of Yangtse River during the period from 00:00 UTC July 11 to 00:00 UTC July 12, 2005 in this paper. The results show that cloud drift wind data assimilation could improve the simulation ability of MM5. And the main rainstorm focus is simulated well in the experiment of combining with the U, V, T of infrared window channel and water vapour absorption channel. The output of model is been analyzed and diagnosed and it comes to the conclusion that MM5 could replay the dynamical field of this heavy rainfall, and there is MPV < 0 in the low-level of rain area, where the MPV negative center locate in the rainstorm focus.