Assimilation Of FY4A/GIIRS Infrared Water Vapor Channels Radiances And Its Application In Regional Numerical Weather Prediction

Satellite radiance data are widely used non-conventional observation data which can compensate the shortage of conventional observations data in the oceans and plateaus.The usage of satellite radiance data plays a key role in enhancing the accuracy of major catastrophic weather forecasts such as meso-and micro-scale weather forecast,typhoon and heavy rainfall.Based on the high spectral resolution of hyperspectral infrared water vapor channels radiance,this study carried out its assimilation and application in the regional numerical weather prediction for the quality control process in assimilation research.In the first place,according to the characteristics of GIIRS water vapor channels combined with "entropy subtraction method",water vapor Jacobian matrix,noise equivalent temperature difference and brightness temperature deviation statistical characteristics,50 water vapor channels which are sensitive to water vapor concentration and containing the most information are selected.The “box” scheme is utilized in thinning the data,which solves the problem of the low efficiency of observation data utilization and the waste of computational resources caused by the large scale of the state vector,and removes the correlation between different observation data.Besides,based on the method of field of view,the cloud detection of GIIRS radiance is implemented,and the rejection of data affected by cloud is realized.Afterwards the threshold control is adopted to further constrain the brightness temperature bias.Eventually,this study adopts the bias corrections of 128 view points.According to the correction result,most of the channel brightness temperature bias are more Gaussian.In this paper,the control and assimilation experiments are carried out in the forecast area for analysis and testing in order to ensure that the hyperspectral infrared water vapor channels radiance can be accurately assimilated in the numerical weather prediction system.The results illustrate that the assimilation of FY4A/GIIRS water vapor channels radiance data in GRAPES_MESO has positively improved the prediction of extraordinary rainfall.