基於 WRF‑LETKF 雷達資料同化系統(WRF-LETKF Radar Assimilation System， WLRAS)架構的對流尺度集合資料同化（ensemble data assimilation; EDA）系統已廣泛應用於台灣短期降水預報研究。本研究探討該對流尺度 EDA 系統的兩項近期改進。首先，在標準操作中，包括徑向風速與反射率在內的雷達觀測資料，會在扇形區域內進行重取樣（resample），並以距離相關的加權函數處理。然而，此方法會犧牲資料密度，並影響掌握對流強度的能力。本研究將極座標下生成的超級觀測（superobservations）改為在笛卡爾座標下的超級觀測。其次，本研究將微脈衝差分吸收雷達（MPD， Micropulse Differential Absorption Lidar）水汽剖面觀測資料。MPD能以高時間解析度提供大氣水汽垂直結構資訊，並具備直接量測水汽場之能力。本研究將其同化至對流尺度EDA系統，以彌補雷達觀測缺乏直接水氣相關測量的限制。 本研究以 2022 年 5 月 31 日台北盆地的午後雷暴個案進行實證分析，該事件發生於「台灣地區豪雨觀測與預報實驗」（TAHOPE）與「太平洋極端降水預報計畫」（PRECIP）第二個密集觀測期間，屬於強降水事件。結果顯示，相較於極座標方式，笛卡爾座標下的雷達資料能有效提升海上資料覆蓋率，同時在高資料密度區維持對流強度。這種資料分布有助於呈現降水強度。即便對流胞在台北盆地被觸發並迅速發展成大規模對流系統，額外同化位於上游 60 公里的 MPD 資料可進一步增加水氣量，進而改善模式雷達回波同化的表現，並大幅改善極短期降雨預報。;A convective-scale ensemble data assimilation (EDA) system based on the framework of the WRF-LETKF radar assimilation system has been widely applied to study short-range precipitation prediction in Taiwan. This study investigates two recent advancements in the convective-scale EDA system. First， in the standard operation， the radar observations， including radial velocity and reflectivity， are resampled with the data within a fan-shaped area using a distant-dependent weighting function. However， this will sacrifice the data density and affect the ability to capture the intensity of the convection. We replace the superobservations generated in the polar coordinate with the ones arranged in the Cartesian coordinate. Second， observation from a Micropulse Differential Absorption Lidar (MPD) is assimilated in the convective-scale EDA system to compensate for the limitation of radar observation lacking the direct moisture-related measurement. The MPD is an innovative instrument that can provide water vapor profiles at high frequency. The impact of these recent advancements is investigated with an afternoon thunderstorm case in the Taipei Basin on 31 May 2022， a heavy rainfall event during the intensive observation period #2 of the joint Taiwan-Area Heavy Rain Observation and Prediction Experiment (TAHOPE)/Prediction of Rainfall Extremes Campaign in the Pacific (PRECIP) field campaign. The results indicate that compared to polar coordinates， the Cartesian radar data effectively enhance data coverage over the ocean while maintaining convection intensity in high-data-density regions. This data distribution is beneficial for representing the precipitation intensity. Even though the convective cells were initiated and burst into an extensive convective system in the Taipei Basin， the additional assimilation of the MPD data 60 km upstream increases moisture availability， leading to improved representation of model radar reflectivity. More results from DA and prediction experiments will be presented.
