人為影響對2009年莫拉克颱風極端降雨變化的歸因分析

Abstract

天氣與氣候極端事件到底是不是真的已經與過去發生的事件截然不同，一直是極具爭議性的問題。本研究針對最近發生的極端天氣事件（如熱帶氣旋）進行機率事件歸因研究，期望能克服模式模擬極端天氣與氣候事件的能力限制，同時也能夠透過獨特的數值實驗設計釐清並量化過去氣候變遷中人為的貢獻。本研究中主要分為三部分，（1）過去百年人為所造成的大氣與海洋變化的估算，運用第五階段耦合模式比較計劃（CMIP5）的氣候模式數值實驗資料庫，可以將人為排放溫室氣體與氣溶膠的作用與已知的氣候系統自然變動加以區分，同時也以不同模式的估算涵蓋其不確定性。（2）評估雲解析風暴模擬模式對於侵臺颱風與其伴隨降雨的模擬能力，了解其掌握大尺度背景環流、熱力結構以及地形對颱風影響的能力。（3）利用歷史情境和只有自然驅力情境的系集模擬，進一步展開人為因子對於颱風影響的量化評估。 運用機率事件歸因的統計分析，並以莫拉克颱風（2009）為例，研究發現過去百年人為所造成的氣候暖化，對於颱風路徑並沒有顯著性的影響，整體颱風強度指數的增加雖然只有 5%左右，但是統計上非常顯著。而伴隨颱風的極端強降雨分析顯示，對於區域強降雨超過 500 mm以上的極端事件發生機率，人為的影響非常可能（大於 90%）會使極端降雨機率增加至少 10%。若以相對於颱風中心的角度分析，極端降雨超過 2000 mm 以上的事件發生機率，人為的影響可能（大於 66%）會使極端降雨機率增加至少 10%，甚至也不排除有 10%的機會，人為的影響使極端降雨發生的風險增加了一倍。進一步透過颱風環流與熱力結構分析發現，極端降雨增加的原因不只來自於水氣的變化，上升氣流加強以及其與極端降雨增加的空間對應關係，顯示動力效應提供了額外的助力。Whether or not we can found a distinct change in the recent weather and climate extremes that is detectable from natural variability is a rather controversial research topic. Our project aimed at using probabilistic event attribution framework to study anthropogenic impact on recent weather extremes (e.g. tropical cyclone). There are three important golds of this study. First is on the estimation of anthropogenic warming of the past century. It is evaluated by the difference from 20th century historical run with all forcing and natural forcing only experiments in the CMIP5 climate model data archive. The uncertainty of estimation would be quantified by the spread of multi-model ensemble. The second milestone is the detailed validation on whether a high-resolution regional model can reasonably simulate the evolution of a typhoon affecting Taiwan and the associated rainfall. Finally, the ensemble simulation of historical scenarios and natural only scenarios is used to quantify the impact of human factors on typhoons. By applying the probabilistic event attribution statistics to the ensemble simulations of 2009 typhoon Morokat. We found that while the anthropogenic impact did not change the typhoon track systematically, and there is a significant increase of 5% on the intensity index of tropical cyclone. More importantly, our result suggests that there is a 10-15% increase in the risk for tropical cyclone rainfall extremes when anthropogenic forcing included in the Typhoon Morakot simulation. The increase is not only supported by the more abundance of water vapor. The additional dynamical impact from the enhancement of ascending motion corresponded well to the location of rainfall extreme increases.