|
Firstly, using cloud detection to Abigail Van \The AMSU data physically inversion, skies, clouds and precipitation weather, temperature, humidity, water, substances inversion results, and AMSU observation error re-estimated. For inversion results, using FNL information, cloud satellite CloudSat data and Herd Plan under cast radiosonde data for test validation data. The results showed that: 1, the AMSU-B ch2 channel the initial observation incremental and precipitation probability combination for cloud detection scheme to reasonable the AMSU data into the three types of skies, clouds and precipitation. The cloud detection scheme can be as different weather the AMSU data inversion and assimilation quality control standards. Observation error re-estimates can be more objective observation errors, improved temperature and humidity inversion effect. 3 for the temperature variable, under clear sky conditions variational method can produce accuracy higher than the a priori information on the atmospheric layers inversion results; cloud area and rain, the variational method is low-rise and high-rise in atmospheric accuracy higher than the prior information inversion results, and in the middle of the atmosphere (500hPa), the accuracy of the variational method of inversion results due to the emergence of cloud and precipitation lower than the a priori information, this occurs because is the observation operator CRTM simulation of cloud and precipitation particles microwave scattering effect is not precise enough. In addition, in all weather conditions, the inversion temperature and DOTSTAR dropsonde data deviation less than 3k: temperature inversion in the horizontal direction and the FNL data close. Accurate for variable humidity, the variational method in the upper atmosphere (above 400hPa) produce more accurate inversion results than the prior information; inversion results but in the middle and lower atmosphere (400hPa), the variational method The degree is lower than the prior information. The IWP for solid water path, liquid water path LWP, precipitation rate RR and other water substances variables, the variational method is capable of generating CloudSat satellite data inversion results. Inversion test a priori information AMSU data by the statistical regression method inversion, so this test can compare the differences variational method and statistical regression method inversion effect: temperature variable, clear sky conditions variational method the The inversion is better than the statistical regression method, Cloud area and rain, the variational method is only in the atmosphere low-rise and high-rise superior to the statistical regression method; humidity variables, only changes in the upper atmosphere (above 400hPa) The points method inversion effect was better than the statistical regression method.
|