本子計畫屬於一整合型計畫，該整合型計畫的依據是108 年度科技部自然司防災科技學門專題研究計畫課題重點說明(防災氣象領域，學門代碼：M1710)之研究課題1-3(臺灣與周邊地區雷達遙測應用於氣象防災的技術開發、 驗證與應用)，發展雷達資料品質控管處理與應用等相關前沿技術。本子計劃具體的工作項目為精進雷達資料反演技術，以獲得在複雜地形上高時空解析度包括熱力場的三維氣象變數場，再根據此高品質的資料，研究人員可嘗試了解影響對流演化的關鍵因子，開發可進行短期預報劇烈天氣的高解析度雷達資料同化技術。申請人近年來發展的多都卜勒雷達風場合成系統WISSDOM (WInd Synthesis System using DOppler Measurements)，可以同時使用來自任何數目氣象雷達的觀測，以及結合中尺度預報模式、地面測站、探空、剖風儀、衛星反演的海上低層風場等資料，合成出涵蓋雷達baseline以及沿著複雜地形上的三維風場，特別適合用在台灣島上多山環境及周邊海域的風場分析。 本子計畫的研究將搭配WISSDOM所合成的風場，結合地面測站的溫度壓力觀測，研發可以在地形上直接反演熱力場的運算方法，再將由直接觀測與經數學反演的在地形上的三維氣象變數場，初始化一個高解析度的數值模式，本計畫將選取多個具代表性的真實個案進行測試，評估對模式在山區預報降雨能力的改進程度。 ;This proposal is a sub-project of an integrated project, which is based on the 2019 disaster prevention policy released by MOST with the purpose of developing radar data QC procedures and advanced data application technologies. The goal of this sub-project is to improve the radar data retrieval methodology so that high spatiotemporal resolution of three dimensional meteorological fields over complex terrain can be retrieved. By using these high quality data sets, it is possible for researchers to identify the key factors controlling the evolution of the convection, and design proper high resolution data assimilation technique to conduct short term forecast of severe weather events. In recent years the applicant has developed a multiple-Doppler radar wind synthesis method WISSDOM (WInd Synthesis System using DOppler Measurements). WISSDOM is capable of using data collected by any number of weather radar, and merging wind information from mesoscale model forecast, surface station, radiosonde, wind profiler, and low level wind over ocean derived by satellite, to synthesis the flow field along radar base line as well as over complex terrain. WISSDOM is particularly useful in analyzing wind fields over Taiwan’s complex topography and vicinity area. This sub-project will use the WISSDOM synthesized wind field, and combine the surface station observed temperature and pressure, to develop an algorithm to directly retrieve the thermodynamic fields over terrain. Then, the observed and retrieved three dimensional meteorological parameters will be applied to initialize a high resolution numerical model. In this research multiple representative real cases will be selected and tested so that one can evaluate the improvement of the model ability in forecasting severe rainfall events in mountainous regions.
