Abstract
Under the United States Environmental Protection Agency's (U.S. EPA) CERCLA program, soils suspected of lead (Pb) contamination are evaluated to assess the impact of soil Pb exposure on blood Pb levels. The decision to remediate relies, in part, on whether the measured soil exposure point concentration (EPC) excee ds an action level. The U.S. EPA established data quality objectives (DQOs) to support data collection used to estimate the EPC and assess confidence in remediation decisions. To support DQO processes at sites where site-specific soil Pb relative bioavailability (RBA) is assessed, a statistical simulation model was developed that estimates false compliance/exceedance decision error probabilities based upon uncertainty in the RBA-adjusted EPC, employing model inputs defining (1) the sampling protocol being evaluated, (2) variability in total and bioavailable soil Pb across the assessed area, and (3) analytical measurement uncertainty. A framework for utilizing the simulation model is presented using a hypothetical site informed by concentration and soil Pb bioavailability distributions from an actual Pb-contaminated site. Pre-sampling, false compliance/exceedance decision error probabilities were predicted for various sampling protocols. A DQO-compliant sampling protocol was then selected, and accuracy and precision in the measured EPC were assessed relative to a specified risk-based action level.