Research On Data Assimilation Of COSMIC RO Data For Typhoon Prediction And Its Quality Evaluation

Nearly twenty years, the ability to forecast the path and the intensity of the typhoon is constantly improving with the development of numerical models. The structural design and physical processing of models are increasingly comprehensive, thus the significant of the initial conditions for the numerical prediction results are increasing. Currently, meteorological satellite remote sensing data can provide large-scale, high time resolution observation of atmospheric information, significantly improved the quality of initial condition of the numerical prediction model. Especially marine, highland, desert and polar and other areas with scarce conventional data distribution. However, due to a variety of complex issues and limitations of assimilation system, only5percent of satellite observation can be used for the assimilation system. Subject to lots of information on the impact of clouds and precipitation are not fully utilized and these data usually has important influence on the forecast for the mesoscale weather system such as tropical cyclone and rainstorm etc. Compared to remote sensing satellites, GPS radio occultation using other means of detection. It’s excess phase, bending angle. And Doppler shift are nontraditional meteorological variables. GPS radio occultation is not sensitive to atmospheric particulates of the area of cloud and rain. The GPS radio occultation has high data accuracy and vertical resolution, its’observation data are unbiased and have a good complementary with other satellite. So the research on GPS occultation has a great significance.The COSMIC (Constellation Observation System of Meteorology, Ionosphere and Climate) second level data wetPrf was used in this paper. Firstly COSMIC data was quality evaluated and analysis respectively with radiosonde data and NCEP FNL Global Analysis Data, to reveal the characteristics and provide scientific basis for the quality control of the assimilation system. Further for the vertical resolution of the COSMIC is to high compared to the GRAPES model layers, it could cause the redundancy effect analysis. The author developed a thinning scheme for the COSMIC data that coupled with a numerical model vertical stratification, and the prediction results were better in the typhoon forecast assimilation experiment. Finally the author used the Global/Regional Assimilation and Prediction System to run some experiments of typhoon forecasting then evaluated and analysis the effect of COSMIC data assimilation in numerical prediction. Following is the paper’s main research and conclusions Summary:1. In the region of China, the temperature of COSMIC data is slightly lower than the radiosonde data and FNL data. The root mean square error and bias of T and Vp is acceptable. In the world. According to different heights, different levels and different regions, COSMIC data is of high quality, but the error distribution characteristics with different height or area is slightly different. The overall, COSMIC data has a well reliability and accuracy, not only provides the new data for the typhoon forecast in China, at the same time, the error distribution characteristics provide a scientific basis for the quality control of COSMIC data assimilation.2. According to the vertical structure characteristics of GRAPES model, the author design a COSMIC thinning scheme that more fit for the model, it can effectively reduce the date redundancy problem in data assimilation. Those experiments’ results showed that the thinning scheme can significantly improve the quality of data assimilation.3. GRAPES assimilation prediction system using the new data thinning scheme and COSMIC data to run some assimilation experiments of the typhoon forecast. The results show the scheme and data has obvious effect on the offshore typhoon forecast in China, is a effective way to further enhance the typhoon path and intensity prediction.