多變量分析應用於台灣淡水河流域之水質分析

Investigation of Water Quality of Dansuie River Basin in Taiwan Using Multivariate Analysis

鄭任伯(Jen-Po Cheng)

指導教授 ： 范書愷

元智大學/工程學院/工業工程與管理學系/碩士(2008年)

https://doi.org/10.6838/YZU.2008.00211

摘要

本研究針對台灣淡水流域(基隆河、新店溪、景美溪、大漢溪及主支流)進行水質評估分析。研究中將水質及魚種調查資料以量化的方式，利用多變量分析來解析水質污染特性，並探討水質變化對魚種的影響。本研究利用因素分析由多項水質指標萃取出綜合型水質指標，再以集群分析將水質樣本分成兩群、三群及四群，並歸類各調查水質樣站，來歸納出不同的分群數目，會有不同的的污染特性。並以判別分析來驗證及討論集群分析的分群效果。接著透過路徑分析來建立溶氧水質模式，探討溶氧在分群方式為兩群時，溶氧與其他水質指標彼此的因果關係，進而針對其分析結果反應水質污染整治重點。最後藉由研究溶氧在各類魚種的變化，配合各群的水質污染特性及整治重點，探討不同的魚種在不同水質狀況下其分佈的情形及未來生態整治上的方向。

關鍵字

多變量分析 ； 集群分析 ； 集群分析 ； 判別分析 ； 路徑分析

並列摘要

This thesis presented a water quality analysis of Dansuie River Basin in Taiwan (Keelung River, Shindian River, Jingmei River, Dahan River, and its main tributary) using a sequence of multivariate methods. In this research, water quality and fish species were studied to characterize the water pollution with multivariate analysis, and the relationship between water quality and fish species were investigated. First, factor analysis was adopted to extract several comprehensive water quality indices from general water quality indices, and further, the water quality samples were classified to 2-4 clusters using cluster analysis to represent different severity of water pollution. Each water quality monitoring station was categorized according to the extracted water pollutants. Discriminant analysis was used to verify the outcome of cluster analysis. Furthermore, DO water quality model was established with path analysis to examine the relationship between DO and the other water quality indices in different clusters. In water pollution remediation, the key water quality index for improvement was represented by path analysis. Finally, the interval estimation of DO of each fish species was calculated, and according to this result, the distribution and the water environment of each fish species were discussed.

並列關鍵字

Multivariate analysis ； Factor analysis ； Cluster analysis ； Discriminant analysis ； path analysis

參考文獻

1. Ahumada, R., Vargas, J., and Pagliero, L. (2006). Simple model of dissolved oxygen consumption in a bay within high organic loading: an applied remediation tool. Environmental Monitoring and Assessment, 118, 179-193.

2. Alberto, W. D., Pilar, D. M. D., Valeria, A. M., Fabiana, P. S., Cecilia, H. A., and Angeles, B. M. D. L. (2001). Pattern recognition techniques for the evaluation of spatial and temporal variations in water quality. A case study: Suquía River Basin (Córdoba-Argentina). Water Research, 35(12), 2881-2894.

3. Andrews, D. F., Gnanadesikan, R., and Warner, J.L. (1971). Transformations of multivariate data. Biometrics, 27, 825-840.

4. Arhonditsis, G. B., Stow, C. A., Steinberg, L. J., Kenney, M. A., Lathrop, R. C., McBride, S. J., and Reckhow, K. H. (2006). Exploring ecological patterns with structural equation modeling and Bayesian analysis. Ecological Modeling, 192, 385-409.

5. Arhonditsis, G. B., Paerl, H. W., Valdes-Weaver, L. M., Stow, C. A., Steinberg, L. J., and Reckhow, K. H. (2007). Application of Bayesian structural equation modeling for examining phytoplankton dynamics in the Neuse River Estuary (North Carolina, USA). Estuarine, Coastal and Shelf Science, 72, 63-80.