US EPA

8748841 

Meetings & Symposia 

A Comprehensive evaluation of the biovapor model for prediction of petroleum vapor intrusion 

Hers, I; Jourabchi, P 

2014 

Air and Waste Management Association 

218-243 

English 

The use of models for prediction of petroleum hydrocarbon vapor intrusion can be of significant benefit for site evaluation purposes where screening approaches based on generic criteria or exclusion distances indicate the potential for a complete exposure pathway or unacceptable health risk. Models can be used for multiple purposes: Estimate indoor air concentrations and/or derive soil vapor criteria as part of a risk assessment; test the conceptual site model; evaluate the effect of changes in site conditions including predictions for future conditions; and prioritize site actions. For petroleum hydrocarbon compounds, it is essential that a model that incorporates aerobic biodegradation is used for prediction of vapor intrusion. Several available analytical and numerical models incorporating biodegradation are reviewed, followed by a comprehensive evaluation of the BioVapor model, an analytical model based on the Johnson and Ettinger model framework that includes oxygen-limited aerobic biodegradation. The BioVapor model was chosen for evaluation because of ease of use, availability and technical rigor. The key assumptions and inputs for the BioVapor model are explained, and sources of data and where appropriate reasonable default values are described including boundary conditions for oxygen flux to below buildings, data sources and a method for estimation of a first-order decay constant, and hydrocarbon source concentrations for different petroleum source types. Model sensitivity analysis as quantified by the benzene attenuation factor indicate biodegradation input parameters with a moderate to high sensitivity for the range of values assumed are the first-order decay constant, natural soil respiration, soil type and foundation air flow rate. When methane generation from ethanol degradation of an E10 gasoline blend is modeled, there is at least an order of magnitude increase in the benzene attenuation factor due to increased oxygen demand from methane oxidation. The results demonstrate the importance of conducting a sensitivity analysis. Scenarios where it may not be appropriate to use the BioVapor model and where instead a multi-component numerical model may be more appropriate to use are described (e.g., where there is significant methane generation or heterogeneous soils). Case studies for four sites representing a range of conditions including sites with light nonaqueous phase liquid (LNAPL) and dissolved contamination sources are presented, in which model results are compared to measurement data. The model predictions are generally consistent with observed behavior for different compounds and in most cases provide for a conservative prediction of soil vapor and/or indoor air concentrations when compared to measurement data.