應用多變量統計技術研析地下水水質特徵－以烏溪流域為例

Evaluating Water Quality Characteristics of Groundwater Applying Multivariate Statistical Techniques-Case Study of Wu River Basin

江漢全(Hann-Chyuan Chiang) ； 廖仲威(Jhong-Wei Liao) ； 李蕙君(Hui-Jun Li) ； 林冠宇(Guan-Yu Lin)

《農業工程學報》 61卷2期 (2015/06) Pp. 75-85

https://doi.org/10.29974/JTAE.201506_61(2).0005

摘要

烏溪屬台灣重要河川之一，分布在南投縣、台中市及彰化縣境內，整體流域面積為2,026平方公里，主流長119公里。台中及埔里兩沖積盆地為烏溪流域地下水資源蓄存區，造就烏溪流域豐沛的地下水資源，但地下水水質受到氣候、人為活動及地質等諸多因素之影響，因此評估地下水水質特徵，成為利用地下水資源十分重要的一環。多變量統計方法中的因子分析係藉由降低資料維度，保留最相關的資訊以顯示各變數間之關係，掌握影響水質特性之主要原因或污染源，並藉由Piper水質菱形圖相互印證，然而群集分析可藉現有資料之特性進行相似性分類，並結合土地利用之情形，助於瞭解其特性差異之分布。故本研究應用多變量統計方法中的因子分析及群集分析，探討烏溪流域上游補注區至中下游流出區的地下水水質特性，因子分析之結果顯示，本區受到鹽類交換因子、碳酸鹽礦物因子以及人為污染因子之影響為主，而群集分析之結果顯示，井群分布主要依人為污染及地質因素為主要分布依據。

關鍵字

地下水水質 ； 因子分析法 ； 群集分析法 ； Piper水質圖 ； 烏溪

並列摘要

Wu River is one of the important rivers in Taiwan, located in Nantou County, Taichung City and Changhua County. Its overall watershed area reaches 2,026 square kilometers, and mainstream 119 kilometers. Taichung basin and Puli basin are the reservoirs of groundwater which creates abundant groundwater resources in Wu River Basin. Since the groundwater quality is affected by the factors including climate, geology, human activities, etc., the assessment of groundwater quality characteristics becomes a key part of utilizing the groundwater resources. Factor analysis retains the most relevant information to show the relationships of various variables and understand the principal cause of pollution which influences the water quality characteristics by reducing the data dimension, as well as using the Piper diagram to confirm the results. Cluster analysis, on the other hand, uses the characteristics of existing data to carry out similarity analysis, and combine the situation of the land used for assisting to understand the distribution of the data characteristic differences. In this study, factor and cluster analyses, which are both multivariate statistical methods, were used to investigate the characteristics of groundwater quality in Wu River basin from upstream recharge area to downstream discharge area. The result of factor analysis indicates that the area was significantly affected by salt exchange factor, carbonate minerals factor and the man-made pollution factor. In addition, cluster analysis shows that man-made pollution and major geological factors affected the distribution of wells.

並列關鍵字

groundwater quality ； factor analysis ； cluster analysis ； Piper water quality diagram ； Wu River