本文主要在探討氣候變遷下濁水溪流域所受到的影響，本研究首先將濁水溪流域依地形分為山區與平地兩個範圍，利用頻率分析及趨勢檢定，分別研究各月降雨參數的差異與趨勢改變。接著藉由馬可夫鏈理論，建立一階馬可夫轉移機率矩陣，來預測出未來三個月的降雨情況。最後利用趨勢檢定結果修正降雨參數，進而合成雨量來模擬未來十年的氣候型態。 分析結果顯示，山區與平地在降雨量上有明顯差異，山區的總降雨量約為平地的1.6倍，造成此差異的原因為降雨強度與降雨機率的同時影響。月雨量頻率分析上，發現三參數對數常態分佈對於各月份的模擬性相當高。趨勢檢定結果明顯發現，7月份的降雨量與降雨強度有增大的現象，而平地乾季則有降雨天數減少的現象。最後氣候預測的兩個案例中，總雨量皆有些微增加的現象，但在乾濕季的降雨表現上則有明顯的不同，若氣候根據過去的趨勢持續變動(case1)則表現出未來十年內乾旱的程度會加劇；若氣候維持在一定程度中震盪(case2)則顯示未來十年內乾季會擁有比以往更多的雨量。 This paper aims to discuss the influence of climate change in Cho-Shui river basin. First, Cho-Shui river basin was divided into mountainous regions and level ground according to the landform of the area. The difference and tendency changed monthly rainfall was studied by frequency analysis and tendency test. Second, first-order transition probability matrix by Markov chain theory was built, and monthly rainfall in the next three months was predicted. Lastly, rainfall parameter by tendency test result was estimated and simulating the climate pattern in the future ten years. Based on the analysis result, there are significant differences between mountainous regions and level ground. The rainfall in the mountainous regions is 1.6 times the level ground. The reasons of the differences are the influencing factors, rainfall intensity and rainfall probability, acting simultaneously. On monthly rainfall by frequency analysis, it is found that Three-Parameter Log-normal Distribution was adequate in every month. It can be discovered obviously by tendency test result that both total rainfall and rainfall intensity increase in July, the dry season raining days decreased in the level ground. On the two climate forecast cases, the total rainfall all increased slightly, but there is difference between wet season and dry season in rainfall display. In case 1, it will increase the degree of drought in the future ten years. In case 2, there will be higher rainfall in the dry season in the future ten years.
