Formatted Title

Considerations for Empirically-Derived Default Attenuation Factors for Vapor Intrusion Screening

Background/Objectives

Most regulatory agencies develop risk-based screening levels for vapor intrusion (VI) based on U.S. EPA’s default attenuation factor (AF) of 0.03 derived from subsurface (SS) and indoor air (IA) data.  Concerns exist over the applicability of the AF for commercial/industrial buildings, soil-gas vapor data, buildings with slab-on-grade construction, and VI sites located in temperate geographic climates.  Recent studies intended to address these limitations have shown that AFs can vary by several orders of magnitude depending on the assumptions used for data filtering and the statistical method for analysis. Understanding how these factors affect the AF is critical to the development of a technically defensible AF for VI screening. The goal of this presentation is to demonstrate how the AF varies as a function of a) assumptions used for data filtering and data pairing within individual buildings and b) the method used for statistical analysis.  

Approach/Activities

An empirical AF database was created, containing over 14,000 IA and SS vapor data pairs from more than 100, predominantly trichloroethylene (TCE) and tetrachloroethylene (PCE), sites in California.  The vapor data were extracted from site investigation reports contained in California Environmental Protection Agency’s Geotracker and Envirostor electronic databases.  Various methods were used to identify low-concentration SS and IA vapor data potentially affected by background (non-VI) sources. Populations of data pairs, sites, and buildings varied depending on the assumptions used for background filtering and the statistical method used for AF derivation. Key variables and considerations for AF derivation are highlighted. 

Results/Lessons Learned

Analyses of the filtered database resulted in AFs that were:

• one to over two orders of magnitude less than the AF 0.03 recommended by US EPA;
• highly dependent on the SS vapor concentration and statistical method for analysis;
• generally consistent with values reported in peer-reviewed literature from other VI sites;
• weakly correlated with many key variables thought to affect VI; and
• better aligned with theoretical relations when accounting for spatiotemporal variability within individual buildings.

Results of this study are being used to guide a new national empirical AF study that will involve analyses of additional vapor data collected at numerous VI sites across the US.