環境品質調查資料空間變異分析之探討

Abstract

本研究針對地下水、土壤等環境品質調查資料，利用統計方法分析，增加資料的利用價值。地下水水質方面，以民國82年至94年台灣地區區域性地下水監測資料為依據，利用多變量分析並配合地理資訊系統(GIS)將資料重新呈現，結果顯示台灣地區地下水受到五個因子主要影響，分別為因子一「海水鹽化因子」、因子二「重金屬污染因子」、因子三「酸鹼值因子」、因子四「有機污染因子」以及因子五「錳金屬影響因子」，其累積貢獻度為77.34%，將各因子造成原因加以討論。土壤重金屬Ni小樣區及點尺度方面，利用移動視窗法以及半變異元建立彰化地區Ni金屬之空間結構，其污染影響距離約為700公尺，另外可由迴歸分析得出和美鎮附近有持續污染發生的潛勢，應持續進行監測管制。台金濂洞煉銅廠附近由於冶煉銅礦伴隨產生之As煙塵外洩，進而污染到附近環境，利用各種統計分群方法，顯示以有限混合分佈模式進行單變數研究可得到有效之分群意義。

The statistics methods were used to analyze the environmental quality data, such as groundwater data and soil heavy metal content data, and hope to increase the value of those data. Using multivariate analysis and Geographic Information Systems (GIS) to analyze the groundwater quality data in Taiwan from 1993 to 2005 can find that there are five main factors for groundwater in Taiwan. Those factors are factor 1 (Saline Factor), factor 2 (Heavy Metal Pollution Factor), factor 3 (pH Factor), factor 4 (Organic Factor), and factor 5 (Manganese Factor). The cumulative percent of variance is 77.34%, the results and its cause are discussed. Use moving window method and semi-variogram to analyze the structure of Ni in Chang-Hwa. Find that the effect range of Ni in Chang-Hwa and find it is about 700m. Then, using regression analysis to find that there is contamination continuous occurring in Homei, and should be monitored. The arsenic contamination in Chinkuashih is caused by the smoke from mining industry. After using five statistics methods to analyze the univariate data, it shows that finite mixture model can get the effective classified group.