本計畫擬採用一高解析度(網格間距小於4.0公里)之數值模式為一計算平台，發展一適用於台灣地區，針對風暴/對流尺度(storm/convective scale)劇烈降雨系統之氣象雷達資料同化方案，並著重在探討此方案對於極短期(1~ 3 小時)定量降水預報的改進程度。吾人計畫使用的數值模式為Weather Research and Forecasting (WRF) model，準備使用的資料同化方法則以三維變分法為基礎。初期主要的工作項目在改進現有之多雷達風場合成技術，以提供較佳的三維風場資訊。此外，並評估於不同的條件下，當同化雷達資料後，對模式預報的影響。這些條件包括：雷達與分析區域的相對位置、雷達資料的觀測誤差、不完整的資料覆蓋、背景場的選取及其扮演的角色、雷達掃描策略的時空分佈、最佳的同化窗口等。後續之規畫則為真實個案之測試，真實個案的資料來源可為全台作業與研究用雷達，或是2008年西南氣流實驗當中，由NCAR S-POL、移動式雷達TEAM-R、氣象局七股或墾丁雷達所觀測的劇烈降水系統。In this project we plan to use a high-resolution (with grid size less than 4.0 km) numerical model as a platform to establish a storm/convective scale radar data assimilation algorithm suitable for Taiwan’s geographic conditions. The major focus is to study the impact of the radar data assimilation on the short-term quantitative prediction forecast (QPF). The Weather Research and Forecasting (WRF) model and the 3DVAR method will be the candidates for this assimilation scheme. In the first stage of this work, we will develop an improved multiple-radar synthesis technique by which a better three-dimensional wind field can be obtained. In addition, several factors that affect the assimilation results will be investigated. They include the relative positions of the radar sites with respect to the model/assimilation domain, the radar data errors, incomplete data coverage, the role of background fields, the spatial/temporal resolutions of the radar scanning strategy, the optimal assimilation window, the assimilation frequency, and so on. The design for the second stage of this work will be the testing of this assimilation algorithm in real cases. The data of the real cases can be from all the operational and research radars in Taiwan, or the observations collected by NCAR S-POL , TEAM-R , CWB RCCG and RCKT during TiMREX field experiment to be held in 2008. 研究期間：9802 ~ 9812
