島嶼地形影響颱風路徑偏轉及打轉之機制研究

Abstract

過去的觀測和數值研究皆顯示，颱風在接近台灣地形時，可能會受地形影響而有偏轉的路徑出現，有時此偏轉運動十分地顯著，甚至偶爾會出現路徑打轉的現象。這些顯著的偏轉或是打轉路徑的發生，很可能會延長及擴大風雨影響台灣的時間和區域，因此這是一個非常重要的颱風預報問題。而過去對此現象的研究仍非常有限，亦缺乏以包含完整物理過程模式所進行之高解析度模擬與深入探討，因此有關地形引起的偏轉運動之動力機制仍有待釐清，而這不僅是颱風預報問題，也是一個相當有趣的基礎學術議題。 本研究將利用MM5中尺度模式進行高解析度 (最高解析度為3公里)的數值模擬，探討柯羅莎颱風 (2007) 打轉運動和台灣地形之關係，及其背後的動力機制；研究結果顯示柯羅莎的偏轉與地形的高度密切相關，當台灣地形高度降低時，柯羅莎的偏轉幅度也減少，反之亦然。若以解析度9公里進行模擬，則無法模擬出柯羅莎顯著的南偏路徑，分析顯示9公里和3公里解析度對地形描述的差異，可能並非造成其模擬路徑有所不同之主因，而此兩種解析度對渦旋結構及其中物理過程之描述對模擬結果的影響則尚待進一步研究。 此外，若以MM5進行包含完整物理過程的高解析度理想氣流和理想渦旋實驗，則發現颱風在接近(離開)地形中、北部(南部)時，有顯著的南(北)偏運動發生；初始位置越北(南)的颱風，其接近(遠離)地形時的南(北)偏的情形就越顯著。 個案模擬與理想實驗均顯示，颱風受地形影響而出現偏轉路徑的同時，颱風和地形之間的低層風速有顯著增強之情形發生，颱風內核區的非對稱流場(asymmetric flow) 也因此產生變化，分析顯示此非對稱流場與偏轉運動有密切關聯。軌跡線分析顯示低層空氣質點經過地形與颱風間的狹小通道時有匯合的情形發生，顯示此風速增強與所謂的通道效應(channeling effect)之概念相當一致。 為了全面性了解此議題，我們計畫進行更多相關的敏感度實驗，探討背景流場速度、渦旋強度與結構、地形和渦旋運動方向之間的角度(impinging angle)所扮演的角色，釐清渦旋遇到地形時發生顯著的偏轉或打轉運動的流制(flow and parameters regimes)。

Both observation and numerical studies showed that typhoons are prone to be deflected when approaching high topography. Sometimes, the track could be sharply deflected, and looping motion might occur. The warning areas, which are lashed by severe rainfall and gust, could shift greatly due to the sharp turn. This is a very important typhoon forecast problem. However, research issues with such unusual typhoon motion have not been well addressed in literatures. Few idealized numerical experiments have been conducted based on a full-physic model with high resolution. The mechanisms leading to the sharp turn, which is an interesting scientific problem, have not been well investigated. High resolution (3-km) simulations are conducted using MM5 model to investigate the mechanisms leading to the loop motion of the typhoon, and the role of Taiwan topography. The results show that high topography plays an important role in Typhoon Krosa’s looping. The southward track deflection is reduced when the terrain elevations are lowered, and vice versa. An experiment with coarser resolution (9 km) cannot well capture the southward turn. Differences in the topography from two model resolutions appear not critical to the simulated track. Impacts of the model resolution on the vortex structure and physics processes still need to be addressed in the future. Idealized experiments are also conducted using MM5 model with a 3-km mesh. These simulations show that the more the vortex approaches (leaves) the northern (southern) topography, the more southward (northward) deflection occurs. Both the real-case and the idealized studies indicate that strong winds at the channel between the storm and topography are enhanced when great deflection occurs. The asymmetric flow within the inner-core appears to contribute to the track deflection. Meanwhile, the backward-trajectory calculation of the low-level air parcels shows the confluent trajectories when the air column passes over the channel between the high terrain and the vortex. The results of the backward trajectory analysis are consistent with the concept of the channeling effect. To fully investigate this issue, more sensitive experiments will be designed in the future. The impact of the magnitude of the ambient flow, vortex intensity and structure, and the impinging angle will be further studied to identify the flow and parameters regimes leading to the sharp track deflection and the looping motion.