| dc.description.abstract | The objective of soil sampling can be grouped into three types, including determination in average contaminant concentration, contaminated soil boundary, and confirmation of finished soil remediation. An appropriate sampling approach is a key point for effectively remedying the contaminated soil. Sampling design in the past was used to traditional types of data collection in order to reduce trouble but difference resulted from contaminant characteristics and remediation methods were neglected. According to EPA’s guideline of environmental data collection, the seven-step process for soil sampling comprises judgmental sampling, simple random sampling, stratified sampling, systematic and grid sampling, ranked set sampling, adaptive cluster sampling, and composite sampling. In this study, cases of the soil contamination are collected to make a comparison of the sampling methods. On the basis of sampling objective, the appropriate sampling methods were evaluated. The effects of the species of contaminants, such as heavy metal, high volatility compounds and low volatility compounds, and the types of soil remediation, such as in-stiu, on-site and off-site, on soil sampling were discussed. The view of statistic was applied to determine sampling size and data variances. The best sampling methods in Taiwan would be presented with various sampling objective.
The obtained result indicates that systematic and grid sampling in Taiwan is applied widely, because it could fit the various sampling objective. The simple random sampling and the composite sampling are effective for determining average concentrations of contaminants. For determining boundary of contamination, the judgmental sampling and the adaptive cluster sampling could be used under specific conditions. The judgmental sampling could obviously save cost, if the information of the contaminated soil has been known. Due to lacking analysis technique and soil information, the use of the stratified sampling and the ranked set sampling is confined. If the above-mentioned conditions can be improved, the efficiencies of the stratified sampling and the ranked set sampling are quite higher than that of the simple random sampling. | en_US |