Scenario‐Specific Attenuation Factors for Vapor Intrusion Screening

A comprehensive study was undertaken to derive attenuation factors (AFs) for vapor intrusion (VI) screening that are more applicable to variable site conditions than the generic default AF of 0.03 recommended by the United States Environmental Protection Agency (USEPA). The study involved the development of an empirical AF database consisting of over 26,000 paired measurements of indoor air ( C IA ) and subsurface ( C SSG ) vapor concentration data for 37 chemicals at 1541 buildings, 330 sites, and 32 states across the United States. After extensive filtering to exclude low‐quality data and limit potential effects from background (non‐VI related) sources, the database was reduced to 1474 trichloroethene and tetrachloroethene C IA and C SSG data pairs from 271 buildings, 86 sites, and 15 states. Building‐specific AFs were used to address the bias and ambiguity caused by buildings with large numbers of C IA and C SSG data pairs. Descriptive statistics and analysis of variance (ANOVA) performed on the filtered database indicated that the most significant variables affecting building‐specific AFs were subsurface sample type (i.e., subslab versus near‐slab soil gas), building type, US Climate Zone, building construction date, and building foundation type. Scenario‐specific 95th percentile AFs defined for particular combinations of these variables varied by over an order of magnitude and were up to 10× less than 0.03. The 95th percentile AFs were broadly supported by reliability analyses, AF versus C SSG trend evaluations, and an AF analysis of radon data and consistent with values derived in prior empirical AF studies. Some of the less significant variables were time and lateral distance between C IA and C SSG sampling ( P‐ values of 0.0008 and 0.002, respectively). The AFs determined from this study can be used to improve VI screening for a range of building, sampling, and geographic conditions and help focus resources toward sites posing the greatest potential risk.