A Method for Assessing Leaching Potential for Petroleum Hydrocarbons Release Sites: Multiphase and Multisubstance Equilibrium Partitioning

Abstract

This article presents the rationale for the mathematical fate and transport model, which has been provided in the accompanying spreadsheet (GWProt). This spreadsheet model may be used as a simple and scientifically defensible regulatory tool for determining the risk-based soil clean up level of petroleum release sites to protect groundwater quality. The model incorporates either a three- or four-phase partitioning equilibrium mechanism, depending on the detection of Non-Aqueous Phase Liquid phase presence mathematically, as well as Raoult's Law convention and default dilution and attenuation factors. A database of contaminant-specific parameters, including solubility and organic-carbon partition-coefficient, molecular weight, and Henry's Law constant, is assembled for benzene, toluene, ethylbenzene, xylenes, and 12 other TPH equivalent carbon fractions. In addition to distributing organic chemicals among aqueous, sorbed solid, air, and NAPL phases, according to traditional partitioning equations, the algorithm incorporates equations for the conservation of mass and volume. A unique solution is obtained by solving a series of mass balance equations simultaneously using the iterative spreadsheet routine built in MICROSOFT EXCELTM Solver — with the restrictions that the volume is conserved and the sum of the mole fractions is equal to one. Sample calculations are presented for a range of parameter values to illustrate the use of the model and the relative leach-ability of a wide range of representative fuels. Sensitivity analysis was also performed to quantify the effects of uncertainty in the estimates of the key model parameters on model results. Model predictions were compared with the results from a water-fuel experiment. The noncar-cinogenic Hazard Index (HI) for groundwater through direct ingestion was calculated using predetermined oral reference dose (Rfd) values. Applications and limitations of the model are also discussed.