Simulating Errors in Statistical Tests of Leachate-lmpacted Ground-Water Quality

Data on 13 median leachate qualities at 56 landfills show a multivariate character, A simulation technique is used to include this multivariate character of leachate quality in assessing false negative error rates for statistical tests of ground-water quality. Although the technique makes gross simplifying assumptions regarding the conservative mixing of leachate and ground water, the technique is more appropriate than traditional statistical methods for simulating false negative error rates that neglect correlations among water quality parameters. The technique is applied to the simple case of a control chart, intrawell statistical test of ground-water contamination. Simulation results show that both the false positive and false negative errors of statistical tests can be reduced when tests include more parameters and a higher confidence level; thus, the technique can provide a “win-win” situation for regulators and the regulated.     

Ground-Water Quality at a Creosote Waste Site

An abandoned creosote facility in Conroe, Texas, has become a field site for the National Center for Ground Water Research (NCGWR) at Rice University. Ground-water contamination in the shallow aquifer beneath the site was characterized by sampling soils and water quality at 14 monitoring wells and 35 boreholes. Results from six sampling trips over two years for inorganic and organic chemical concentrations in the ground water show wells around the site were contaminated to levels above 800 μg/l for naphthalene, 400 μg/1 for methyl naphthalene, and 150 μg/1 for dibenzofuran. Conservative constituents, traced by chloride concentrations up to 75 mg/l, have migrated 300 ft (90 m) downgradient of the site. Organic contaminants have been adsorbed and microbially degraded in their migration from the waste source as evidenced by their attenuated concentrations. Detailed field pump tests have been performed to evaluate hydraulic conductivity at several of the shallow wells. The U.S. Geological Survey (USGS) Solute Transport Model (Konikow and Bredehoeft, 1978) has been used to predict chloride plume patterns and evaluate parameters which govern transport processes at the Conroe waste site.     

Evaluation of the Johnson and Ettinger Model for Prediction of Indoor Air Quality

Screening level models are now commonly used to estimate vapor intrusion for subsurface volatile organic compounds (VOCs). Significant uncertainty is associated with processes and models and, to date, there has been only limited field-based evaluation of models for this pathway. To address these limitations, a comprehensive evaluation of the Johnson and Ettinger (J&E) model is provided through sensitivity analysis, comparisons of model-predicted to measured vapor intrusion for 11 petroleum hydrocarbon and chlorinated solvent sites, and review of radon and flux chamber studies. Significant intrusion was measured at five of 12 sites with measured vapor attenuation ratios (α_m's) (indoor air/source vapor) ranging from ∼1 × 10⁻⁶ to 1 × 10⁻⁴. Higher attenuation ratios were measured for studies using radon, inert tracers, and flux chambers; however, these ratios are conservative owing to boundary conditions and tracer properties that are different than those at most VOC-contaminated sites. Reasonable predictions were obtained using the J&E model with comparisons indicating that model-predicted vapor attenuation ratios (α_p's) were on the same order, or less than the α_m's. For several sites, the (m were approximately two orders of magnitude less than the a 's indicating that the J&E model is conservative in these cases. The model comparisons highlight the importance in using appropriate input parameters for the J&E model. The regulatory implications associated with use of the J&E model to derive screening criteria are also discussed.     

Evaluation of the Johnson and Ettinger Model for Prediction of Indoor Air Quality

Screening level models are now commonly used to estimate vapor intrusion for subsurface volatile organic compounds (VQCs). Significant uncertainty is associated with processes and models and, to date, there has been only limited field-based evaluation of models for this pathway. To address these limitations, a comprehensive evaluation of the Johnson and Ettinger (J&E) model is provided through sensitivity analysis, comparisons of model-predicted to measured vapor intrusion for 11 petroleum hydrocarbon and chlorinated solvent sites, and review of radon and flux chamber studies. Significant intrusion was measured at five of 12 sites with measured vapor attenuation ratios (α_m's) (indoor air/source vapor) ranging from ∼1 × 10⁻⁶ to 1 × 10⁻⁴. Higher attenuation ratios were measured for studies using radon, inert tracers, and flux chambers; however, these ratios are conservative owing to boundary conditions and tracer properties that are different than those at most VOC-contaminated sites. Reasonable predictions were obtained using the J&E model with comparisons indicating that model-predicted vapor attenuation ratios (α_p's) were on the same order, or less than the α_m's. For several sites, the α_m were approximately two orders of magnitude less than the α_p's indicating that the J&E model is conservative in these cases. The model comparisons highlight the importance in using appropriate input parameters for the J&E model. The regulatory implications associated with use of the J&E model to derive screening criteria are also discussed.