區域數值天氣預報模式能進行高解析度之預報及模擬，而預報結果與初始條件和邊界條件有著直接關係。初始場能藉由資料同化方法加入觀測資料而獲得改善。衛星觀測能提供空間上廣大的覆蓋率，且能彌補傳統觀測較稀少的海洋區域。NASA/EOS/Aqua衛星上搭載著MODIS、AMSU與AIRS等儀器，結合AMSU/AIRS的反演產品能提供高品質的三維大氣溫溼度資訊，此產品之水平空間解析度可達45公里。另一方面，藉由物理反演法所反演出的AIRS 單一視場(single field-of-view; SFOV)反演產品，有著比AMSU/AIRS更佳的15公里空間解析度。不同空間解析度與反演產品準確度的二種衛星資料，對數值天氣預報的貢獻亦不相同，因此我們提出是否有最佳化的使用方法及時機，期能探討對於數種預報參數的改善能力。 本研究藉由資料同化系統分析AMSU/AIRS與AIRS SFOV兩種反演產品對於區域數值天氣預報模式的影響，將使用WRF與其3D-Var資料同化系統，針對2012年6月於臺灣所發生之梅雨期豪大雨事件進行模擬預報實驗。主要結果顯示由於微波與紅外線觀測上的特性，AMSU/AIRS的反演產品具有能提供雲區反演且空間涵蓋較廣的優勢，對於預報的影響較AIRS SFOV產品為大。但另一方面，若AMSU/AIRS與AIRS SFOV的反演資料空間涵蓋範圍一致，則因AIRS SFOV反演產品能提供較佳的空間解析度資訊，因而於溫溼度的預報改善較AMSU/AIRS明顯。因此，高光譜紅外線與微波若能適當的搭配使用，將能進一步提升預報能力。研究中亦發現， AMSU/AIRS反演產品有相對較小的溫度偏差及較高的空間覆蓋率，同化AMSU/AIRS的實驗組對於高度場與鋒面位置的掌握，均較使用AIRS SFOV之預報能力為佳。;The numerical weather prediction (NWP) and simulation model has been developed for decades. It has received substantial improvement in term of predictability. On the other hand, satellite observations and retrieved products may provide critical assistant over ocean than conventional observation. In particular, the combined sounding retrievals from Advanced Microwave Sounding Unit and Atmospheric Infrared Sounder (AMSU/AIRS) suggest a high quality estimation of atmospheric temperature and moisture profiles. Meanwhile, a higher spatial resolution from single field-of-view (SFOV) of AIRS sounding was developed followed by the previous combined AMSU/AIRS product. It is believe that finer spatial resolution retrievals could retain a better gradient structure in a weather system. In this study, we propose to use combined AMSU/AIRS and AIRS SFOV soundings to evaluate the performance for introducing these two different data sets. A heavy precipitation MCS case associated with a Mei-Yu frontal system during early June 2012 in the vicinity area of Taiwan is selected to demonstrate this concept. Weather Research Forecasting (WRF) and its three-dimensional variational module (3D-Var) is used to evaluate the forecast performance due to assimilating of NASA EOS AMSU and AIRS products. The preliminary result suggests that AIRS SFOV data set have better performance over combined AIRS/AMSU in temperature and mixing ratio forecast when both data are selected in the same spatial coverage. However, assimilating AMSU/AIRS data can improve the front location. Therefore, combine AMSU/AIRS and AIRS SFOV these two products may enhance the forecast capability.
