Selective Remediation of Contaminated Sites Using a Two-Level Multiphase Strategy and Geostatistics

Abstract

Selective soil remediation aims to reduce costs by cleaning only the fraction of an exposure unit (EU) necessary to lower the average concentration below the regulatory threshold. This approach requires a prior stratification of each EU into smaller remediation units (RU) which are then selected according to various criteria. This paper presents a geostatistical framework to account for uncertainties attached to both RU and EU average concentrations in selective remediation. The selection of RUs is based on their impact on the postremediation probability for the EU average concentration to exceed the regulatory threshold, which is assessed using geostatistical stochastic simulation. Application of the technique to a set of 600 dioxin concentrations collected at Piazza Road EPA Superfund site in Missouri shows a substantial decrease in the number of RU remediated compared with single phase remediation. The lower remediation costs achieved by the new strategy are obtained to the detriment of a higher risk of false negatives, yet for this data set this risk remains below the 5% rate set by EPA region 7.

Application of Indicator Geostatistics for Probabilistic Uncertainty and Risk Analyses of Geochemical Data

“…외국에서는 토양 오염, 지하수 오염 등의 환경 관리를 목적으로 공간 분포 도 작성뿐만 아니라, 불확실성 모델링을 통한 오염 지역의 구분, 위험성 분포도 작성 등에 지구통계학을 많이 활용해 왔다 Juang and Lee, 1998;Barabas et al, 2001;Van Meirvenne and Goovaerts, 2001;Saito and Goovaerts, 2003;Saisana et al, 2004;Goovaerts et al, 2005;Goovaerts et al, 2008 …”

Geostatistical Simulation of Compositional Data Using Multiple Data Transformations

“…특히 지구통계학적 시뮬레이션은 단일 위치나 동시에 여러 위치에서 고려하고 있는 속성값의 불확실성 추정에 활용될 수 있다 (Goovaerts, 1997;Deutsch and Journel, 1998;Chilès and Delfiner, 2012). 예를 들어 특정 지역에서 측정 단위보다 큰 블록 단위에서 오염 임계치를 초과할 확률을 계산하 거나, 2차 분석에 사용되는 모델에 입력 자료의 불확 실성 전파 등을 정량적으로 모델링하는데 사용될 수 있다 (Kyriakidis and Dungan, 2001;Saito and Goovaerts, 2003;Wang et al, 2003;Park and Oh, 2006;Goovaerts et al, 2008;Oh and Han, 2010). 지구과학 자료는 값 자체의 연산이 가능한 연속형 자료와 서로 구별이 되는 몇 개의 범주로 구성된 범 주형 자료로 구분할 수 있다 (Davis, 2002).…”

Stochastic Simulations of Spatial Variability Based on Multifractal Characteristics

“…Moreover, multiple equiprobable realizations of the simulation process allow assessment of risk associated with uncertainty in the estimation of the studied area. Simulation algorithms have been successfully applied in numerous environmental remediation studies (see Saito and Goovaerts, 2003, as just one of many examples), in assessing soil compaction (Lapen et al, 2001), and in predicting crop yields (Pachepsky and Acock, 1998). The high values are of primary interest to environmental scientists and the optimal simulation strategies for high‐value characterization have been studied extensively; correct representation of the low values has achieved less attention.…”