同化全日衛星輻射量為當代資料同化議題中極具挑戰性的工作，但也因為衛星資料解析度，準確度大幅提升衛星資料同化對數值天氣預報將預期可有重要的影響。而由微波或紅外線所觀測之向上輻射量為與水氣，雲水，冰晶散射特性等大氣變數之複雜函數。此外，所觀測的輻射量對於散射特性非常敏感，而本身我們確定這些散射特性了解不夠，亦無法正確模擬。在本計畫中，我們使用統計方法來掌握多重散射特性以最大化輻射傳遞及大氣變數之關係。以此建立衛星觀測算符並使用於資料同化架構中，再進一步測試其對颱風同化及降雨預報之影響。 ;The assimilation of all-sky (clear and cloudy) satellite radiances is one of the most challenging yet potentially rewarding issues currently facing numerical weather prediction. The upwelling radiances observed by a passive microwave or infrared satellite are a complex function of geophysical variables such as water vapor and temperature as well as scattering from cloud liquid or frozen hydrometeors. The observed radiances are highly sensitive to the undermodeled and under-quantified nature of these scattering signatures. In this proposal we adopt a statistical approach to utilize multiple scattering properties in order to extract the maximally certain relationships within corresponding geophysical and radiative transfer databases. We propose to use this observation operator to test the impact on tropical cyclone and precipitation forecasting.
