本研究中擬結合可在地形上進行多雷達風場合成的技術WISSDOM (Wind Synthesis System using Doppler Measurements)及新發展的熱動力反演方法，並利用反演出來的三維氣象場產品，計算挑選過的預報因子，找出對流系統未來一小時強度與動向變化的關連性。 其中熱動力反演，是指將氣象雷達無法直接觀測得到的三維熱動力場資訊，包括溫度與氣壓場，從雷達合成的風場中萃取出來。傳統方法(Gal-Chen 1978)的優點在於求解一個與氣壓相關的Poisson方程式，沒有其它人為的假設，即可得到熱動力場。但其限制在於如果不搭配探空或飛機等額外獨立的觀測，則無法解析熱動力場的垂直結構。此外，還有無法在複雜地形上進行計算的缺點。相較於傳統方法，本研究發展的技術利用動量方程以及簡化的熱動力方程式，對於影響位溫但不確定性高的源匯項不給予參數化，來解決無法解析垂直結構的問題，方法類似Liou(2001)與Liou et al.(2003)，但經過進一步的改進使此得此方法能夠被運用於地形上。 本計畫可以提供在台灣複雜地形上，以雷達資料結合測站及模式背景場來反演的三維風場和氣壓、溫度擾動場，進而計算、導入及評估對TANC而言具有潛在預報能力的反演物理量，再依據此資訊計算若干挑選後的氣象預報因子(predictor)，並統計其與隨後風暴是否產生，以及產生後的強度與位置等的關聯性。 ;Thermodynamic retrieval is a process by which one can derive the unobservable thermodynamic parameters in a three-dimensional space such as temperature and pressure using multiple-Doppler radar synthesized wind fields. The traditional method suggested by Gal-Chen (1978) can solve a Possion equation to obtain the pressure, but needs extra information from other instruments such as soundings or aircrafts to determine the vertical structure of the thermodynamic fields. In addition, it will be difficult to apply the method of Gal-Chen (1978) over terrain. In this study we will obtain the three-dimensional wind synthesized over terrain by a system called WISSDOM (Wind Synthesis System using Doppler Measurements). Based on the wind information we develop a new thermodynamic retrieval scheme, with the goal to identify the relationship between the convection initiation and the meteorological parameters. Compared with the traditional approach, our method employs the three-dimensional momentum equations. A simplified thermodynamic equation is also added, but without conducting any parameterization for the unknown source and sink terms of the potential temperature. By adding this equation, it is possible to alleviate the ambiguity occurs in the vertical structure of the thermodynamic fields. This method is similar to that proposed by Liou (2001) and Liou et al. (2003), but with further improvement so that the retrieval can be performed directly over complex terrain. This research can provide three-dimensional meteorological fields over complex terrain with high temporal and spatial resolutions. Based on the above information, we will select several predictors, whose relationship with the development of the thunderstorms as well as their magnitude and locations will be studied extensively.
