本研究分析淡水河流域長期觀測水文資料在時間與空間上之變異，資料年限為1961年至2007年不等，首先主要利用線性迴歸與無母數趨勢檢定來評估時序資料變化趨勢，並利用主成份分析與群集分析進行空間變異分析。 分析淡水河流域長期溫度結果為趨勢增加且具正向顯著性，在年降雨量與日平均降雨強度變化趨勢亦有顯著性的增加，但連續不降雨日則較不顯著。利用主成份分析降雨之結果顯示:第一主成份可代表絕大部分之變異量達62.2%，第二主成份為13%，第三主成份為10%，進一步探討各個主成份可能代表之變異來源，第一主成份為綜合性指標，第二主成份為地形因子影響，第三主成份為區位因子影響。將主成份scores與ENSO和SOI大氣指標進行相關性分析，結果發現聖嬰現象的確對於淡水河流域降雨造成影響，再利用群集分析將流域降雨空間分布分類討論日平均降雨強度變化，發現近10年日平均降雨強度區位增強主要為石門水庫集水區。最後嘗試利用主成份分析結果，推估未設站位置降雨，假設現有測站降雨未知下，進行降雨推估比較，其結果與觀測值相當吻合，在逐月Nash檢定值有不錯的結果。 Long-term observed hydrological data including precipitation, discharge, and temperature of the Tamsui River basin were analyzed to investigate temporal and spatial variations. First both linear regression and non-parametric trend examination were applied to examine trends and variations of hydrological data in time domain, then the Principle Component Analysis (PCA) and the Cluster Analysis (CA) were further used to investigate spatial variations. Temperature trend is positive significant in the basin. Annual precipitation and daily rainfall intensity are increase and significant, while variations of dry spell days are insignificant. In whole analysis trend is increase more in this section. Monthly rainfalls were analyzed by the PCA. The PCA1 (the first principle component) stands for comprehensive climate characteristics, the PCA2 illustrates topography effects, and the PCA3 is location factor. Correlation analyses between ENSO/SOI indices and PCA scores show basin rainfall was slightly affected by the EI Nino. Results of the CA reveal that daily rainfall intensity in the Shishmen Reservoir has the most significant increasing trend watershed in the past decade. Results of PCA were further applied to estimate rainfall at ungauged locations. Rainfall at assumed unknown gauge can be reasonable estimated.
