Chromium leaching experiments were conducted using soils rich in chromite ore processing residue (COPR). Experiments were performed with acrylic columns under a constant hydraulic head and a free board. Air was purged continuously through the influent solution (simulated rainwater) in order to create an oxidative condition in soils. It was found that the mass of Cr(VI) leached was strongly dependent on temperature. The leaching process accelerated above 8°C. Under the conditions of an influent solution pH of 4.3, air atmosphere, and temperatures ranging from 3 to 38°C, the total mass of Cr(VI) leached was in the range of 0.206 to 0.260 mg/g, which accounted for about 1% of the total Cr (25.5 mg/g) in the soil sample. A one-dimensional transport model was used to simulate the leaching of Cr(VI) from soils. Model simulation results revealed that the leaching process was governed by adsorption/desorption and was attenuated by redox reactions.

Soil vapor extraction (SVE) is a proven effective in situ technology for the removal of volatile organic compounds (VOCs) from the subsurface. However, most systems designed to date are based on empirical methods, rules-of-thumb, or results obtained from time-consuming and expensive pilot-scale field tests. Thus, it is very difficult to extrapolate laboratory results to field applications, to test hypotheses, to evaluate different subsurface scenarios, to accumulate experience, to optimize operation parameters, to predict field performance, and to modify operation parameters in response to unanticipated subsurface conditions. In this paper, the development of a three-dimensional semianalytical model for the simulation of transport and fate of VOCs in the subsurface during a multiple-well soil vapor extraction operation is presented. The model provides the necessary but simple-to-use engineering design tools for the process. Illustrative examples on the use of the model developed are also presented.

Sensitive parameter distributions were identified with the use of probabilistic analysis in the RESRAD computer code. RESRAD is a multimedia pathway analysis code designed to evaluate radiological exposures resulting from radiological contamination in soil. The dose distribution was obtained by using a set of default parameter distribution/values. Most of the variations in the output dose distribution could be attributed to uncertainty in a small set of input parameters that could be considered as sensitive parameter distributions. The identification of the sensitive parameters is a first step in the prioritization of future research and information gathering. When site-specific parameter distribution/values are available for an actual site, the same process should be used with these site-specific data. Regression analysis used to identify sensitive parameters indicated that the dominant pathways depended on the radionuclide and source configurations. However, two parameter distributions were sensitive for many radionuclides: the external shielding factor when external exposure was the dominant pathway and the plant transfer factor when plant ingestion was the dominant pathway. No single correlation or regression coefficient can be used alone to identify sensitive parameters in all the cases. The coefficients are useful guides, but they have to be used in conjunction with other aids, such as scatter plots, and should undergo further analysis.

The deterministic multimedia dose/risk assessment codes RESRAD and RESRAD-BUILD have been widely used for many years for evaluation of sites contaminated with residual radioactive materials. The RESRAD code applies to the cleanup of sites (soils) and the RESRAD-BUILD code applies to the cleanup of buildings and structures. This work describes the procedure used to enhance the deterministic RESRAD and RESRAD-BUILD codes for probabilistic dose analysis. A six-step procedure was used in developing default parameter distributions and the probabilistic analysis modules. These six steps include (1) listing and categorizing parameters; (2) ranking parameters; (3) developing parameter distributions; (4) testing parameter distributions for probabilistic analysis; (5) developing probabilistic software modules; and (6) testing probabilistic modules and integrated codes. The procedures used can be applied to the development of other multimedia probabilistic codes. The probabilistic versions of RESRAD and RESRAD-BUILD codes provide tools for studying the uncertainty in dose assessment caused by uncertain input parameters. The parameter distribution data collected in this work can also be applied to other multimedia assessment tasks and multimedia computer codes.

An increasing weight of evidence has indicated that under certain conditions chemicals above background levels in soils may not be easily released and thus may not have an adverse environmental effect. This has led to a broadening body of knowledge about the extent and rate of release (ROR) of hydrocarbons from contaminated soil. A groundwater fate and transport screening model was developed that incorporates chemical ROR kinetics for the express purpose of identifying where and when the incorporation of ROR information may affect model predictions, as well as site remediation and management decisions. The results from this evaluation indicate that at many sites the incorporation of chemical release kinetics in such models leads to important differences in predicted chemical concentrations that reach a receptor. This paper also discusses how ROR information may be used to improve model predictions at a site, how to evaluate the importance of ROR at a site, and how to identify the most influential input parameters.

This work evaluates the merits and the magnitude of potential errors expected from Hung’s one-dimensional groundwater transport model as integrated into EPA’s Prediction of Radiological Effects due to Shallow Trench Operation (PRESTO) model. PRESTO is a multimedia model designed to assess both the maximum individual doses for a critical population group and the cumulative genetic and fatal health effects for a general population that result from soil contamination and near-surface waste disposal. Hung’s groundwater transport model was incorporated into PRESTO to balance efficiency with accuracy for the purpose of regulatory analysis. This 1D advective model accounts for dispersion through a spatially-dependent correction factor while maintaining the proper overall time-integrated concentration. Generally, 1D models suffer from two important limitations. They are incapable of calculating the transverse concentration distribution of a contaminant plume, and they may introduce excessive errors caused by their inherent simplifications and assumptions. This paper reevaluates the consequence of these limitations for generic regulatory analyses currently practiced in the United States; for this purpose, the calculation of contaminant distributions in a plume is not necessary because only nearby well-water concentrations are required for dose calculations. For typical releases, errors in Hung’s model relative to calculations from a 3D finite-difference groundwater code, MODFLOW/SURFACT, can be maintained below 3%. In light of the limited impact of its simplifying assumptions and its merits in terms of computational efficiency, this work concludes that the Hung model is a useful 1D groundwater transport algorithm for near-field regulatory and compliance-driven human health risk assessments. In these applications, the receptor well is assumed to be located no more than 100 m away from the edge of the contaminated site.