Evaluation of Vapor Intrusion from a Subsurface Diesel Plume Using M ultiple Lines of Evidence

Vapor intrusion from affected soils and groundwater into overlying structures has been identified as a potential mechanism for human exposure to volatile organic chemicals (VOCs), including petroleum hydrocarbons. However, in many cases, background concentrations of organic vapors in indoor air, associated with household products, vehicles, or ambient air, can lead to incorrect conclusions regarding the occurrence of subsurface vapor intrusion impacts. To investigate alleged vapor intrusion impacts from a subsurface diesel plume, a series of investigations has been conducted near a railway facility in Mandan, North Dakota, where organic vapors had been detected in both the subsurface and in indoor air. The results of this investigation demonstrate how multiple lines of evidence, including statistical cluster analysis, can be employed to distinguish between background indoor air quality and organic vapors associated with actual subsurface vapor intrusion. To distinguish vapor intrusion impacts from background indoor air sources, this paper describes data management procedures and statistical methodologies for: i) comparing the organic vapor composition of indoor air, ambient air, and subsurface vapors; ii) comparing indoor air concentrations to background indoor air data compiled nationally and locally; iii) measuring vertical vapor concentration profiles within vadose zone soils and between subslab and indoor air, and iv) evaluating the temporal effect of soil vapor extraction (SVE) system operations on organic vapors in the subsurface and in indoor air. For the Mandan site, this evaluation showed that the indoor air composition in target buildings overlying the diesel plume closely matched the composition in background buildings located away from the affected area and exhibited a moderate match to ambient outdoor air. In contrast, all subsurface vapor sources (sub-slab gas, SVE discharge, well headspace gas) formed a very distinct cluster from that of the indoor air and ambient air data sets, such that the subsurface vapors could not create the vapor composition observed in the indoor air. The multiple lines of evidence indicated that indoor vapor sources were the likely cause of the VOCs measured in indoor air, with no evidence of contribution from subsurface vapor intrusion.