GPS Precipitable Water Data Are Appled To Adjoint Assimilation Systerm

One of the primary means that numerical weather prediction model toenhance the capacity of weather prediction is modification of data error.Regular stations distribute sparsely, observation time is limited, anddifficult to reflect changes in space-time atmospheric water vapor, soradar and satellites data and other unconventional information as asupplement to the conventional information applied to numericalprediction models. Global Positioning System (GPS) of atmospheric watervapor can conduct continuous moisture measurement, sounding quiteconventional precipitation data accuracy. GPS precipitation data hasbeen precedented as an effective precipitation data for the model,currently, GPS precipitation data variational assimilation technologyhas become the main Drive and trends for applying the GPS data to numericalforecasting model.The numerical simulation and adjoint assimilation of GPS precipitable water experiments are conducted in the process of a heavy rainoccurred in Beijing, Tianjin and Hebei area between 23 and 24 July2005. Of these the GPS data are simulated and assimilation program,Cressman radius 15km, 30km, 45km were selected to conduct sensitivitytests and analysis, three methods are used to test the degree of improvements GPS precipitation data do to model humidity field and effects of adjoint assimilation. They are 24-hour precipitation TS score, RMS between simulated humidity field and of the real situation,the real precipitation map. MM5 simulation adding GPS precipitationdata can not only improve the accuracy of forecasting precipitation,After adjoint method, temperature, height, wind field and other model variables are modified to a certain extent, the RMS experimentsresults of temperature field, Wind field show: GPS precipitation data can improve initial humidity field at the time of initial temperature height and wind fields in different degrees. Moreover, Cressmanradius should be choose properly in order to realize GPS precipitation data information contained in precipitation.