河溪流量特徵是目前評估全球氣候系統暖化、土地利用變遷或其他環境變遷對集水區水文反應的重要指標,且與洪水或乾旱、水利結構物建設與操作、溪流或近岸海洋環境系統、集水區物質輸送有密切關聯。近期研究從雨量觀測資料分析指出,過去半世紀以來無論全球或臺灣之強降雨頻率顯著增加且微降雨頻率約減半,這顯示未來可能受到更嚴重的乾旱與洪澇災害。因此有必要調查河溪流量特徵的變化以及對降雨變遷的反應,推估未來可能改變的趨勢。本研究分析臺灣1970-2010年48個山地集水區日流量特徵 (包括頻率、規模、延時、時間點、變異度),並針對其中4個流量品質良好集水區的暴雨事件進行降雨變遷投影,輸入至2個分散型水文模式 (TOPMODEL與HBV) 以預測未來洪水事件的變化。 研究結果顯示,流量特徵分析方面,高海拔地區平均流量與基流量較高,且高流量發生的頻率高且延時長,但變異性較低;低海拔地區則變異性較高。在水資源區差異,南部暴雨流量與變異性較其他地區大。高流量時間點從西北部以逆時針繞中央山脈遞增。經過地勢特徵分析,流量特徵主要和河川密度與起伏特徵有關聯。在趨勢分析,以低流量相關的特徵改變最顯著。中南部低海拔區與宜蘭地區基流增加,與降雨乾旱研究不一致。基流在淡水河、大甲溪、曾文溪與高屏溪上游則減少,增加這些下游人口密集地區水資源減少的危機。在極端事件的流量反應模擬結果顯示:總降雨量約增加2倍,且最大降雨增加至2.23倍。其造成總逕流增加約2-3倍之間,但在洪峰流量可達2-4倍。地表逕流甚至增加4-5倍。本研究分析臺灣流量特徵連結至暴雨事件,提供防災與資源規劃之參考,並且增進環境變遷下的水文循環變化的基礎知識。
Streamflow characteristics are the intricate response of climatic system, landuse change, and other environmental changes. Meanwhile, they are also the basis for flood and drought warning, design and operation of hydraulic constructures. Previous studies analyzed historical rainfall records showed that the intensity of heavy rain increased doubled and light rain became about half worldwide. The signification rainfall intensification implies the droughts and floods may threat to natural and anthropogenic system frequently. This study analyzed the streamflow characteristics, including frequency, magnitude, duration, timing and varibility, of 48 mountainous catchments in Taiwan since 1970 to 2010. Besides, the possible rainfall projections are generated and inputted to 2 distributed hydrological models (HBV and TOPMODEL) for assessing the streamflow responses to rainfall intensification. Results revealed the mean of daily flow and baseflow in mountainous catchments is higher, but lower in variability. In contrast, the stormflow and varibility in lower and southern catchments are higher. The timing of wet season circulates from north-western with counter-clock. Besides, flow characteristics are found to highly correlated to drainge density and relief-relevant geomorphologic chacteristics. In trend analysis, the baseflow in central and southern low-land and Ilan region is increasing. On the other hand, the baseflow decreases in important water resources region, such as Danshui, Dajia, Tzengwen, Gaoping. In simulation between historical and projected rainstorm, accmulated rainfall increase 1.79-folds, and maximum 6-hr rainfall increase 2.32-folds. Under this circumstance, the accumulated flow may increase 2.27-folds, peak flow increase 3.52-folds, and the surface flow increase 4 to 6-folds. The amplification in peak flow is significant which should be the foundation of disaster prevention and water resources plan, and improve the understanding of hydrological cycle under environmental change.