| 摘要: | 台灣5 到 6月份的梅雨鋒面以及7到9月間的颱風，除了系統本身主要環流和降雨之外，經常會引進強烈的西南氣流並夾帶著劇烈豪雨，對台灣地區造成重大災害。因此，定量降水估計及預報的準確性就顯得相當重要。 而雷達觀測能提供高時空解析度的觀測資料，透過雙偏極化雷達觀測可得到偏極化參數Z（reflectivity）、ZDR（differential reflectivity）、ΦDP（differential phase）、KDP（specific differential phase）及ρhv（correlation coefficient）等，利用這些偏極化參數得到的降雨估計演算法，如：R(Z)，R(Z,ZDR)、R(KDP)等不同的關係式估計降雨，近十幾年來在不同波段的雙偏極化雷達被廣泛的應用，降雨估計的精確度也有不錯的結果。 本研究主要是利用在西南氣流實驗期間，中央大學車載X-波段雙偏極化雷達（TEAM-R）的觀測資料（Z、ZDR、KDP）估計降雨，與雨滴譜儀（二維雨滴譜儀2-DVD和JWD撞擊式雨滴譜儀）、局屬自動雨量站的實際觀測作比較。來評估X-波段雙偏極化雷達在台灣南部地區估計降雨的適用性。 使用的資料時間為2008年6月14日的個案（IOP8）。結果顯示在降雨估計方面，利用R(KDP)其精確度來的比R(Z)還要好。利用R(KDP)估計降雨產生誤差的分布情形，隨著觀測距離越遠與地勢高度增加。使用的資料高度也跟著增加，與地面雨量站或雨滴譜儀站的變異性也就越大。根據前人研究，降雨估計所能接受的誤差範圍約在15到20％以內。因此利用TEAM-R雷達資料估計降雨時，20公里以內是不錯的估計範圍。另外，因受ZDR衰減修正不足的影響，在利用R(Z,ZDR) 估計降雨時，會有高估的情況產生。 In Taiwan, the Mei-yu fronts in May and June and the typhoons in summer and early fall often induce strong southwesterly flows and consequent heavy rainfall, resulting in severe disasters. Therefore, the accuracy of quantitative precipitation estimation and forecast is of great importance. Weather surveillance radars provide observational data at fine temporal and spatial resolutions; moreover, dual polarimetric radars offer parameters such as reflectivity (Z), differential reflectivity (ZDR), differential phase (ΦDP), specific differential phase (KDP) and copolar correlation coefficient (ρhv). In recent decades, a number of rainfall estimation algorithms utilizing these polarimetric parameters, e.g. R(Z), R(Z,ZDR) and R(KDP), are applied to dual polarimetric radars with different wavelengths and perform well in accuracy. This study mainly exploit the polarimetric data (Z, ZDR and KDP) of the NCU’s mobile X-band polarimetric radar (TEAM-R) during SoWMEX/TiMREX to estimate rainfall. These polarmetric rainfull estimates then were compared with measurement from disdrometers (2DVD and JWD) and the CWB’s automatic rain gauges. The suitability of the quantitative precipitation estimates retrieved from the X-band polarimetric radar in southern Taiwan was evaluated. The data on June 14, 2008 (during IOP-8) were analyzed. In rainfall estimation, R(KDP) is shown to perform better than R(Z), but the errors utilizing R(KDP) increase as the target distance gets larger and the terrain gets higher. The higher the used data is, the larger the variability between the radar and the rain gauge (or disdrometer) can be. According to former studies, the acceptable range of errors in rainfall estimation is approximately 15~20%. For this reason, TEAM-R’s data in a range within 20 kilometers are acceptable. In addition, R(Z,ZDR) usually overestimates the rainfall due to insufficient correction for the attenuation of ZDR. |
