本研究之目的為利用多部都卜勒雷達觀測資料，改善模式的初始場，以提升模式定量降水預報之能力。此方法主要包含三大部分：(1)多都卜勒雷達風場合成、(2)熱動力場反演、(3)水氣場調整。
吾人選取2008 西南氣流實驗計畫(SoWMEX)中所觀測到的IOP8個案，作為本研究的實驗對象。使用中央氣象局七股雷達(RCCG)、墾丁雷達(RCKT)及美國國家大氣研究中心(National Center for Atmospheric Research; NCAR)所屬的SPOL 雷達，於2008 年6月14日1200 UTC時所觀測的回波及雷達徑向風資料，反演出大氣三維風場及熱動力場，並對水氣場進行調整，最後同化至模式中。本研究使用NCAR Weather Research and Forecasting (WRF) Model作為預報模式。
經過本方法調整模式初始場，實驗顯示模式對降雨預報的改善可達3 小時，雖然仍有高估之趨勢，但相較於未同化前的降水分佈會更趨近於觀測。此外，由本研究的實驗還發現水氣的調整確有其必要性，而以多部雷達網連的方式增加資料覆蓋量，對反演與同化結果會有重要的影響。
相較於四維變分(4DVAR)及系集卡爾曼濾波器(EnKF)等同化技術，本方法可使用較少的計算機資源與觀測資料完成計算步驟與同化，開始進行短期的定量降水預報。未來預計將本方法用於測試午後對流或颱風降雨系統的預報上。

The purpose of this study is to improve the model initial field and hence the quantitative precipitation forecast (QPF) by using multiple-Doppler radar observational data. The assimilation algorithm includes three major components: multiple-Doppler radar wind synthesis, thermodynamic retrieval, and moisture adjustment. A case during IOP8 of 2008 Southwest Monsoon Experiment (SoWMEX) is selected. The radar data in use are the reflectivity and radial wind of the CWB RCCG and RCKT radars, and the SPOL radar from NCAR. The three-dimensional winds, retrieved from the radars and sounding data, are utilized to calculate thermodynamic fields. The moisture field is also adjusted when certain criteria are satisfied. The computational platform employed for this study is the Weather Research and Forecasting (WRF) model.
Compared with the experiments without data assimilation, the assimilation technique developed in this study significantly improves the accuracy of the model rainfall forecast for about three hours. The results from a series of experiments indicate that the moisture adjustment is necessary. The use of multiple-Doppler radar data is also crucial, because a larger radar data coverage is expected to lead to better results. The proposed
method can use relatively fewer computing resource and observational data within a shorter observational period to conduct short-term QPF. In the future, it is planned to apply this method for the forecast ofprecipitation in afternoon thunderstorms or tropical cyclones.