The Study On The Local Application Of The Meso-scale Model MM5

High resolution meso-scale numerical mode is the foundation of the refinement weather forcast. For the moment, MM5 meso-scale numerical model obtains extensive use in our country. This mode has many alternative physical process and schemes. It can be applied to many different initial value fields. But the default settings of MM5 is primarily for America and may not reasonable for other place. Both here and abroad substantive simulations show, there may be many difference in the selection of the physical processes and processing schemes in different areas or differ season. Thus, in order to get the best result of the mode, many simulations need to be done. The localization of the meso-scale model MM5 include the selection of the initial data and convection parameterized schemes, the applications of the output of MM5 mode, the application of MM5 mode's output under MICAPS system.The simulations of MM5 to 14 cases with different cumulus convection parameterized schemes show that, none is always better than the others. The large-scale circulation features simulated with different cumulus convection parameterized schemes tend to have the same result, but it often cause different dynamic and thermal structure of the meso-scale and micro-scale systems. The simulations to Huanghe-Huaihe Cyslones show that NCEP and T213 data almost have equal ability to simulate the path of Huanghe-Huaihe Cyslones. Generally, the simulative precipitation field by using NCEP data is better than that by using T213 data. However, when the western North Pacific subtropical high occurs obviously in the modeling area, the simulative precipitation field and the path of Huanghe-Huaihe Cyslones with T213 data is better than NCEP data. With the cyclone's positions and intensities simulated by MM5 as inputs, a liner neural network had been constructed. It can be found that the liner neural network could improve MM5's stimulant results. By study the data format of MM5 and MICAPS, the output of MM5 can be integrated into MICAPS system. Based on the output data of MM5, the stepwise-regress and the Kalman filtering method are used to forecast diurnal air quality. It can be found that the Kalman filtering method can give better results than stepwise-regress.This theses contains the selection of mm5 mode's initial value field , the selection of physical process and schemes, the interpret application of MM5 mode's output , the application of MM5' output under MICAPS system and other aspects. It could offer overall reference to other places when they unfold refinement weather forcase with the meso-scale mode.