梅雨季超大豪雨個案之模擬與診斷分析

Abstract

本研究統計1987至2008年梅雨季台灣地區發生「超大豪雨」事件發現，以台灣西南部測站發生機率最高，多集中在5月下旬至6月中旬；且與梅雨鋒之MCS關係密切。本文選取降雨分佈範圍最廣、降雨量最豐，且降雨延時最久的2005年6月12至15日及2006年6月8至11日兩個案進行研究。 從綜觀環境條件與降水特徵顯示，兩個案共通特徵包括（1）低對流層屬暖濕條件性不穩定大氣，（2）最大相當位溫梯度以及水氣梯度分布於700 hPa以下，（3）低層輻合均為個案中MCS之重要激發機制，（4）有明顯低層西南噴流，大氣呈現對流性不穩定狀態，（5）中層短波槽線的移入，以及（6）高層輻散分流場的存在，提供MCS發展的有利環境條件。最大的差異則為2006年個案的垂直風切、低層噴流、中層槽線、高層噴流以及垂直偶合均較2005年個案來的顯著。降雨特徵方面，2005年個案主要降雨區域在於西南部平原，其中在嘉南地區平地降水大於山區，高屏地區則降水最大處集中在山區，最大降雨區出現於屏東山區，最大日降雨量為585 mm，日夜變化顯著；2006年個案主要降雨區域在於中部及南部山區，其中在中部及高屏地區降水最大處多集中在山區，在嘉南地區山區降水則大於平地，最大日降雨量為718 mm，日夜變化較不顯著。 兩個案之MCS皆伴隨顯著集中之垂直渦度，針對MCS垂直渦度之診斷分析顯示，兩個案渦度發展與維持的機制主要來自於渦度輻散項及垂直平流項，系統移動的主要機制則是受到水平平流效應主導。由水氣收支計算可得到水平水氣通量輻合對於對流系統之水氣來源扮演正回饋作用，其中2006年個案中水氣通量（水平輻合）較2005年個案來得大。2006年個案之鋒生過程顯著，中低層鋒生過程扮演導致鋒面對流系統之發展與維持重要角色；而低層噴流生成的條件主要來自於橫向非地轉風垂直分布之噴生作用。 選擇對區域預報能力、對中尺度天氣系統與天氣現象掌握相當不錯的WRF模式進行模擬，結果顯示兩個案在綜觀環境的模擬與實際觀測結果大致相似，惟在風場模擬結果偏弱，溫度場則明顯偏高。另外，在降水部份，模式對台灣地區降水分佈具有相當良好之掌握，惟強度上在2006年降水強度明顯偏弱，且東部地區降水掌握較不理想。 綜合而言，2005年個案中梅雨鋒面結構為垂直近乎不傾斜的淺系統，槽線系統不明顯，鋒面水平溫度梯度微弱，低層具有強水平氣旋式風切，較似正壓系統，個案中MCS發展及維持機制，主要是透過CISK 過程來維持；其過程是由於低層鋒面帶上之輻合激發對流發展，對流潛熱釋放造成局部氣壓梯度增強，因而增強地轉風；而潛熱釋放亦有利低層輻合之增強，產生非地轉風，再透過科氏加速增強西南風，使LLJ增強，形成一正向的回饋作用。而2006年個案屬發展中斜壓系統特徵較為明顯，個案中鋒面及其伴隨之MCS發展及維持機制，主要是透過斜壓過程來維持；其過程是由滯留鋒面帶上存在的低層輻合，中層鋒生效應顯著，再配合高層之噴流輻散移入，且地面低壓系統位於中高層槽前之不穩定區，地面至高層系統垂直偶合顯著，進而誘發地面低壓系統增強並發展，造成局部氣壓梯度增強，亦使得LLJ增強，形成MCS發展的有利環境條件。

From the statistic results showed that the “extremely torrential rain” events produced by the Mei-Yu fronts in Taiwan area from 1987 to 2008 were mostly occurred over the southwestern parts of Taiwan between late May and mid-June. In this study, the two cases about the wide-spread and long-lasting rainfall due to MCSs in Taiwan area on 12-15 June 2005 and 8-11 June 2006. In terms of the synoptic environment and precipitation characteristics, there were some common characteristics in these two cases. For instance, warm and moist air with the conditional instability atmosphere in the lower troposphere, it was associated with the highest equivalent potential temperature (θe) gradient and strong moisture contrast to below 700 hPa. The lower-level convergence was the major produced mechanism of MCS and the low-level southwesterly flow generated convective instability. The mid-level trough and upper-level divergence were the favorable conditions for the developments of mesoscale convection systems(MCSs). The difference between case 2005 and case 2006 is that the vertical wind shear, low-level jet (LLJ), mid-level trough, upper-level jet and vertical coupling in case 2006 are obvious than case 2005. The case of 2005, heavy rainfall occurred over the southwestern plain of Taiwan, especially in the Chianan Plain, it was heavier than in the mountain area. The maximum accumulation of daily rainfall (585mm) was occurred in the mountain area of Ping-dong County and the diurnal change was obvious in case 2005. However, the maximum daily rainfall in case 2006 (718mm) was heavier in mountain area than in the Chianan Plain, and its diurnal change was not obvious. The analytic results show that the mechanisms of MCSs’ development and maintenance in both cases were derived from vorticity divergence and vertical advection, the movement of MCSs was dominated by the horizontal advection effect. From the moisture budget, the horizontal moisture flux convergence showed positive feedback in the convective systems. The moisture flux (horizontal convergence) in case 2006 was more than case 2005. The frontogenesis process in case 2006 was obvious; it showed that the front system was accompanied with a strong temperature gradient. The mid and low-level frontogenesis process played an important role in the development and maintenance of causing the front convective systems. The jets were formed with the difference of horizontal ageostrophic wind in vertical direction. The WRF modeling results showed that the simulation patterns were similar to the observation. Only the wind field was weaker and the temperature field was higher than observation. In addition, the model precipitation simulation pattern did well in Taiwan area, but the intensity was weaker in case 2006. It was not ideal in eastern Taiwan area was not ideal in case 2006. In the conclusion, the Mei-Yu front in case 2005 was shallow in structure with weak temperature gradient, strong horizontal wind shear and with no significant trough systems. The mechanism was similar to the equivalent barotropic warm core structure. The Conditional Instability of the Second Kind (CISK) was the primary mechanism of development and maintenance for MCSs in case 2005. As the MCSs intensified and persisted, the latent heating was highly efficient in producing significant enhancement of Mei-Yu front systems. The LLJ intensified by Coriolis acceleration of ageostrophic wind was induced by the effect of latent heat release. The LLJ and the effect of latent heat released reinforce each other in convection system through a positive feedback process. However, the baroclinic structure of Mei-Yu front in case 2006 was very clear. The primary mechanism of MCSs for development and maintenance was depended on the baroclinic process. The low-level convergence, mid-level frontogenesis, upper-level divergence and the depression system were vertically superposed in the front system. The deepening and strengthening of the surface low was derived from the vertical coupling process, which supplied feedback enabling LLJ to be intensified and promoted a favorable environment for MCSs to development.