- 學位論文
應用地理統計及空間尺度轉換於污染地區特徵分析與台灣地區重金屬污染場址復育驗收準則探討
Application of Geostatistics and Upscaling Methods on Pollution Area Characteristics Analysis and Guideline for Verifying the Attainment of Soil Remediation for Heavy Metal Contaminated Sites in Taiwan
摘要
利用有限混合分佈模式(Finite Mixture Model, FMM)區別彰化地區東西二圳、三圳系統灌區裡污染來源的特徵，結果顯示本方法對於特性的掌握較以往方法或經驗更具有意義，而以Cr, Cu, Ni, Zn四種重金屬為例，區分的污染來源特徵分別為點狀污染源、帶狀污染源及正常區域。 結合移動視窗迴歸法(Moving Window Regression, MWR)與FMM之優點，可將台灣地區(大樣區)土壤之As含量區分為三群以做為區域背景值評估的依據。另外解析彰化地區(小區域)二種不同採樣尺度資料之差異成因，結果顯示不同採樣尺度之As含量經過尺度換轉之後差異不大，僅有彰化南興里、南美里、平和里南端農地之差異較大，原因為農地引灌了受污染的水源所造成。 本文建議之污染場址改善驗收準則包含四個部分，第一：任何一個驗證樣本之濃度均需低於管制標準值，第二：採樣結果之累積機率97.5%對應之濃度值需低於管制標準值，第三：採樣結果之累積機率50%對應之濃度值需低於監測標準值，第四：改善整治工程之均勻度指標值需小於2。
並列摘要
Finite Mixture Model (FMM) was used to analyze the characteristics of soil heavy metal pollution source around the 2nd and 3rd East-West irrigation systems in Chang-Hua County. The result shown that this method is more effective than other methods used before when dealing with the pollution source identification. And three statuses can be classified as point pollution, diffused pollution and normal region. Combining Moving Windows Regression (MWR) with FMM, the soil As content in Taiwan can be divided into three categories area, this may reflects the soil formed from different mother rocks. Besides, we use upscaling method to describe two different sampling scales in Chang-Hua. The result shows that As concentration in two scale are about the same in the whole study field, except in Nan-Shin village, Nan-May village and ping-He village. The different is caused by the polluted irrigation water. The guideline for Verifying the attainment of Soil Remediation for heavy metal contaminated site includes four parameters: first, the concentration of every sample should be less than the control standard; second, the concentration of 97.5% probability of cumulative curve should not exceed the control standard; third, the median concentration of sampling should be below the monitoring standard; and fourth, the indicator of evenness for the remediation works should be less than two. In comparison with traditional methods, our proposed guideline is shown to be effective in verifying remediation results.