全球暖化對臺灣極端降雨量影響之評估

Abstract

近年來，氣候變遷議題已為世界重視並關注。極端降雨事件之頻率增加、強度增強可視為氣候變遷之警訊。本研究探討全球暖化對極端降雨量影響之評估。研究於空間尺度為臺灣全域平均及雨量單站之日雨量，以累積雨量方式決定日雨量門檻值，極端降雨由日雨量門檻值訂定，由此決定出年極端降雨百分率為大於門檻值的累積雨量占年總雨量的比率。本研究蒐集近50年(1961-2010年)臺灣雨量站之雨量，評估全球暖化對臺灣極端降雨量影響。 研究指出溫度與極端降雨百分率變化，具雙變數常態分布關係。因此，條件常態分布下，序率方式評估全球暖化對極端降雨百分率影響。並以機率方式評估全球暖化，極端降雨百分率超過Clausius-Clapeyron增加之1.07倍速率。 結果指出當全球平均溫度升高1K，臺灣全域之極端降雨百分率，其超越機率為0.82。使用雨量站之日雨量亦指出類似之結果，除了臺灣中部與中央山脈西側之雨量站，大部分之雨量站亦有可能超越Clausius-Clapeyron速率。研究評估之結果可提供決策者因應氣候變遷制定調適策略與衝擊評估之依據。

The effects of global warming on frequency and intensities of extreme storm events have been demonstrated by many catastrophic events in recent years. This study aims to assess the effect of global warming on rainfall extremes in Taiwan using fifty years (1961 – 2010) of daily rainfall data collected at 15 rainfall stations around Taiwan. Rainfall extremes, or the top 10% bin rainfalls, of islandwide-average and station-specific daily rainfalls were defined by threshold daily rainfalls which were determined by a cumulative rainfall approach. Yearly top 10% rainfall percentages were then calculated based on yearly rainfall extremes. In this study, changes in the top 10% rainfall percentages and changes in the global mean temperatures were found to be characterized by a bivariate normal distribution. Thus, the effect of global warming on changes in the top 10% rainfall percentages can be stochastically assessed based on the conditional normal distribution. It was found that, under one degree increase of global mean temperature, the probability of having a super Clausius-Clapeyron rate of change in the top 10% rainfall percentage for the islandwide-average daily rainfalls equals 0.82. Similar analyses using station-specific daily rainfalls also showed that most stations are likely to experience super Clausius-Clapeyron rate of change in the top 10% rainfall percentages, except those stations located in central Taiwan and west of the Central Mountain Range.