本研究計畫以WRF 模式為計算平台，針對都卜勒雷達觀測建立LETKF 資料同化系統，並評估其改善中尺度短期數值天氣(含定量降水)預報之效率及準確度的能力。初期以同化虛擬之都卜勒徑向風資料為主，亦即以 OSSE 方式設計數值實驗，後續之計畫則擬逐步增加回波及雙偏極化雷達參數之同化，最後則進一步針對實際觀測資料進行測試。本研究規劃預計執行三年，預期要達到之研究目標如下，本次申請的計畫則為第二年: (1) 第一年：針對多重巢狀網格之 WRF 氣象模式，在LETKF 系集同化方法之架構下，建立雷達資料同化系統。初期目標為同化單或多部都卜勒雷達之徑向風資料，進行OSSE 實驗，分析其對中尺度天氣系統之影響。 (2) 第二年：開發此系統同化其他雷達產品之能力，如回波、差異反射率 (Zdr)、反射率差(Zdp)、比差異相位差(Kdp)等，建立相關觀測算符；進行 OSSE 實驗，分析同化不同產品對預報中尺度天氣系統之影響。 (3) 第三年：針對強降水事件(如颱風、梅雨鋒面)進行OSSE 實驗，評估 WRF 模式在不同系集條件（初始化、物理參數化等）下，此系集資料同化系統對各氣象場（含定量降水）預報之影響程度。進而利用實際觀測資料評估此系集資料同化系統應用於實際個案之可行性。 This research plans to use WRF (Weather Research and Forecasting) Model as the computational platform, and develop a Doppler radar data assimilation system based on the LETKF (local ensemble transform Kalman filter) method. The ability of this system in terms of improving the efficiency and accuracy of a mesoscale short-range numerical weather prediction (including QPF) will be evaluated. In the early stage of this project, only simulated radial wind data will be assimilated, while reflectivity and other dual-polarimetric parameters are expected to be included in the later stage. Applications to use real observations will also be investigated. This proposal is designed for three years, and the research goals for each year are listed in the following. This application is for the second year. (1) Based on a nested WRF model and LETKF method, this research will build a radar data assimilation system. The goal of this year is to establish the assimilation algorithm for single or multiple Doppler weather radars. All experiments will be tested under the OSSE framework. The impact of using radar data on meso-scale weather forecast will be evaluated. (2) Develop the observation operators for other dual-polarmetric radar variables, such as reflectivity, differential reflectivity (Zdr), reflectivity difference (Zdp) and specific differential phase (Kdp), and compare the performances of the model forecasts after assimilating these additional radar parameters. (3) Evaluate the WRF-LETKF forecasts of various meteorological variables for selected heavy precipitation events (e.g. Mei-Yu front, typhoons) under different model setups (initialization, physical parameterization, etc). Finally, we will assess the feasibility of applying this data assimilation system to real case studies. 研究期間：10008 ~ 10107
