臺灣位於亞熱帶氣候地區,每年防汛期(5月1日至11月30日)即遭逢颱風侵襲,其所帶來長延時、高強度之降雨,致使各地水患、坡地及土石流災情不斷。2009年莫拉克颱風帶來的降雨量更創下歷史新高,阿里山氣象觀測站最大24小時(1623.5mm)及48小時(2361mm)降雨更逼近世界紀錄(1825mm、2467mm),重創中、南部山區及屏東水患,災害損失極為可觀。 降雨是氣象觀測站的基本觀測項目,臺灣山區氣象測站相較於平地少,僅能於特定位置取得資料,故常有樣本不足或是空間分布不均等情況發生。當災害地點無雨量站時,過去常以鄰近測站為參考依據求得雨量資料,但雨量會隨著時間及空間而有所變化。然而,空間點狀分布的氣象測站資料推估成面化(Surface)空間的氣候分布狀態,是水資源管理、氣候變遷等研究議題的基本需求,也是許多水文數值模型的必要參數。 本研究以臺灣中部地區南投縣陳有蘭溪流域為研究區域,並以2004年敏督利颱風、2008年辛樂克颱風及2009年莫拉克颱風為研究案例。利用克利金(Kriging)、距離反比權重法(IDW)、最小曲率法(Spline)及徐昇多邊形法(Thiessen)進行逐時雨量(1、6、12、18、24、48小時及總延時)之數值內插分析,以測站抽樣方式比較出誤差較小之方法。其研究結果顯示,最小曲率法(Spline)為最適合描述陳有蘭溪流域降雨空間分布特性之方法,且以6小時以上之累積雨量來進行推估為最佳。另將推估面繪製成等值線圖,與環境因子相互套疊分析,其結果發現累積雨量及降雨延時與地形有密切關係,並擬定阿里山、玉山及望鄉為本流域主要降雨中心。
Taiwan is located in subtropical regions. Flood protection period (May 1 to November 30) that is hit by typhoons in every year, which will caused by flood, slope and landslide disaster induced by long delay and high intensity rainfall. The rainfall brought more record high by Typhoon Morakot in 2009. Largest 24-hour (1623.5mm) and 48 hours (2361mm) is closer to a world record rainfall (1825mm, 2467mm) by Alishan weather stations to maul heavily the center and southern mountains and Pingtung flood, the losses is very considerable. Rainfall is an observation item in weather observation station, Taiwan mountain weather station less than the flat, only to obtain data on some specific location, so often the sample is too small or uneven distribution occurs. When the disaster region no rainfall stations in the past often nearby reference station to obtain rainfall data, the rainfall over time and space to change. However, point distribution of the weather station data at space to estimate the space into the surface of the climatic distribution of the state, is the management of water resources, climate change, the basic needs of research topics is necessary in many hydrological numerical model parameters. In this study, watershed of Chenyulan river in Nantou county in central Taiwan, which in Mindulle of 2004 year, and Sinlaku of 2008 year, and Morakot of 2009 year as a Typhoon case study. Used to Kriging, Inverse Distance Weighting method (IDW), the minimum curvature methods (Spline) and Thiessen polygon method (Thiessen) for hourly rainfall (1,6,12,18,24,48 hours and total delay) interpolation of the numerical analysis to compare the sampling station less error methods. The results show that the minimum curvature method (Spline) as the most appropriate description of this case interpretation, methods of spatial rainfall distribution, and accumulation of 6 hours or more of rainfall to be the best estimate. The other side will be drawn into the estimated contour map, and overlay analysis between environmental factors. The results showed that the cumulative rainfall and rainfall duration is closely related with the terrain, development of Alishan, Yushan and Wang Xiang is watershed of Chenyulan river to rainfall center.