頁籤選單縮合
題 名 | 水質指標模式之共線性分析比較研究=Comparison Study of the Water Quality Models by Collinearity Analysis |
---|---|
作 者 | 陳鴻烈; 蔡大偉; 呂銘淵; | 書刊名 | 水土保持學報 |
卷 期 | 42:1 2010.03[民99.03] |
頁 次 | 頁35-47 |
分類號 | 445.212 |
關鍵詞 | 共線性分析; 集水區管理; RPI; WQI; Collinearity analysis; Watershed management; |
語 文 | 中文(Chinese) |
中文摘要 | 本研究主要目的在於探討水質指標模式之共線性,以協助集水區水質管理指標之選擇與監 測。研究中選用淡水河流域為研究區域,在蒐集水質資料後,使用複迴歸分析法,分別建立 RPI (River Pollution Index)、WQI (Water Quality Index)兩種水質指標之迴歸模式,然後再針對模 式進行共線性分析,藉此了解水質指標模式之共線情形,並依據分析結果來比較兩個水質指標 模式。研究結果顯示在RPI 模式中,其組成參數具有顯著共線性,其中又以氨氮與凱氏氮之共 線情形最為顯著;而在WQI 分析結果則顯示此模式同樣具有顯著共線問題,同時檢定結果顯 示氨氮與凱氏氮共線性最強。經過比較RPI 與WQI 之共線性分析結果後,兩模式均為顯著共 線,且在共線參數的分佈上相當類似,均集中在氮營養鹽相關參數。在共線程度上,WQI 模式 明顯較RPI 嚴重,推測原因可能是WQI 為了更完整解釋水質狀態,使用範圍較廣的參數所導 致。針對模式共線性的解決方案,希望未來透過更深入的探討,以提出可減輕共線性並有效解 釋水質之替代方案,協助選擇最佳水質指標。 |
英文摘要 | The main purpose of this study was to compare the water quality models with collinearity analysis, which could be helpful for index selection and monitoring in watershed management. We selected the Dansuie River Watershed as a study site and collected all water quality data available. Multiple regression analysis was applied to build up the regression models in each water quality index: RPI and WQI. Each model was examined by collinearity analysis. The results could help us to compare these two water quality models. We found out several important results in this study: (1) In the RPI model, according to the analysis results there was significant collinearity in this model and the most collinear parameters were ammonia and Kjeldahl. (2) The result of the WQI model showed that high collinearity existed in the model and the strongest collinearity located at ammonia and Kjeldahl and (3) after comparing all the analysis results, both RPI and WQI models exhibited significant collinearity and the distribution of collinear parameters was very similar which concentrated on nitride parameters. WQI model had more serious collinearity than RPI because WQI applied larger range of parameters in order to offer better sensitivity in water quality. In the future, we will do further study on alternatives of reducing collinearity and keeping high explanation for choosing better index in water quality management. |
本系統中英文摘要資訊取自各篇刊載內容。