臺灣博碩士論文加值系統

本研究是以此環保署建議之監測及管制臨界值, 做為鑑識之指標,以全省
各縣巿大樣區土壤重金屬污濃度量測資料進行比對, 並配合地理統計(
Geostatistics) 的克利金法, 進行理論變異元模式之套配, 探討各區域
最佳之半變異圖及等濃度線圖, 藉此鑑識本省各地區監測與管制區域分佈
。對全省15個縣分析結果, 共有13個縣具有超過監測值以上之區域欲鑑
識, 本研究以不同的半變異圖模式來吻合實驗半變異圖, 其中台北縣與南
投縣, 套用球狀模式做計算, 其它11個縣巿都套用高斯模式做計算。而只
有台北縣經自然對數轉換有明顯的改善半變異圖; 並對彰化縣大樣區土壤
重金屬Zn量測資料分別比較不同模式是否對估計值造成影響進行分析, 結
果發現球面模式與指數模式所得之估計值較相近, 而與高斯模式所得之估
計值差異較大, 最後將鑑識結果並做成圖以供參考。

The purpose of this research is trying to find out the soil
contaminated sites in Taiwan. First, the surveyed data of 1600
ha-grid and the tentative threshold of the monitoring and
control values which EPA suggests are compared. Second, the
Kriging me- thod of geostaistics is used to find out the best
semivariogram and concentration contour in each county.
Finally, the contami- nated sites are identified. The results
shows that up to 13 counties among the 15 coun- ties have the
value exceeding the monitoring threshold and need to be
identified. Various semivariogram models are used in this
research to fitted Experimental semivariogram. Spherical model
analysis is used on the evaluation of Taipei and Nanto
counties. The Gaussian model is adapted to analysis the other
11 counties. Upon using the nature-log transformation, only the
value of Taipei county show the obvious progress in the
semivariogram analysis. Furthermore, the Zn contaminated in the
soil of Chang- hwa county is measured to find out if different
methods will have influence on the estimated value. The
conclusion is the estimated value gained from Spherical model
and Gaussian model is closer. However, the estimated value
gained from Spherical model is less similar to the value from
Gaussian model. The identification result is made into the
graph and enclosed in Appendix as reference.