本研究探討以地理統計之克利金法,考量土壤空間變異性,利用有限之先期調查測值,推估調查全區域的濃度分佈,用以輔助篩選後續調查地號之方法,進行超標率及檢出率之分析評估,並比較不同採樣輔助方法之優劣,可實際應用在農地污染調查之上。 研究結果顯示,在案例一中若採樣點數精簡為原本的一半(835筆)時,多變量指標克利金法、一般克利金法、逐步隨機抽樣法,檢出率分別為67.4%、66.9%、49.9%;案例二中若採樣點數精簡為原本的一半(490筆)時檢出率則分別為88.3%、53.4%、49.9%。不論在案例一或案例二,克利金法在檢出率的表現上明顯優於其他輔助方法,尤其在案例二中當採樣點數精簡為原本的一半時,可以得到88.3%之檢出率,明顯高於其他方法。而空間自相關方法僅能有3種組合可以進行佈點,效能僅較逐步隨機抽樣方法佳。區域排水渠道推估法証明污染與灌溉排水相關,亦較逐步隨機採樣法優良,但無法抓住污染熱區的範圍,所以呈現的結果僅較逐步隨機採樣法優良,但仍落後於克利金法。 本研究依循農地重金屬採樣分兩階段進行之常規,建議第一階段使用如瓶架式網格系統式均佈方法作事前評估,第二階段依先前評估結果輔以克利金法進行採樣規劃,必定可節省成本。
A new sampling strategy for soil surveillance was developed in this study, which is different from systematic and grid sampling. Due to spatial variability in natural, geostatistics are commonly applied to estimate the spatial distribution of heavy metals pollution in soil. To enhance efficiency, different strategies are used, assisting with surveys to reduce the sampling numbers. Both over limit rate and screening rate are calculated as evaluation index. The result showed that kriging method is superior to the other methods in all cases. Especially in reducing the soil samples in half, i.e., case II, 88.3% of detection limit is achieved. For Spatial autocorrelation, it only has 3 combinations. Its performance is between kriging method and random sampling. However, for Regional drainage method, although it can prove that the pollution which is related to irrigation and drainage, but it can’t identify the hot spots. This study suggests that in soil surveillance of heavy metals pollution the sampling process can be divided into two stages: the system-type method (e.g., Bottle Rack Grid method) is used in first stage. And then, based on the previous results, kriging method is used to do finer sampling design. It can effectively enhance the implement of agricultural land investigation and restoration.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。