API PUBL 4674-1998

When soils are impacted by leaks or spills, or wastes are placed in impoundments, the potential exists for contaminant vapor migration to enclosed spaces (buildings, conduits, etc.) and leachate migration to groundwater. Regulations have historically considered the leachate impacts on groundwater; however, the issue of vapor migration has only recently been formally and quantitatively considered. This focus has been brought about in large part by the move toward more structured risk-based corrective action (RBCA) approaches (e.g., ASTM 1995), coupled with an increased awareness of this pathway.

The significance of the vapor intrusion pathway and natural attenuation of vapors in the vadose zone is currently the subject of intense debate.When common screening-level algorithms (e.g., Johnson and Ettinger 1991, Little et al. 1992) are combined with conservative soil properties, geometries, and exposure assumptions, the resulting risk based screening levels (RBSLS) are very low. In fact they are often one-tenth to one-thousandth the existing cleanup guidelines in many states. For example, the sample calculation in the ASTM RBCA Standard (ASTM, 1995) suggests that benzene concentrations in excess of 5 µg/kg-soil could be of concern if one wishes to be protective to a 10^-6excess cancer risk level.

Many intuitively feel that the current generation of screening-level predictive models is too conservative and leads to unnecessarily low cleanup levels. Some point to the fact that the algorithms generally do not account for biodegradation and other possible vadose zone attenuation mechanisms. It is reasonable to expect that some chemicals of interest degrade as they migrate, especially those originating from petroleum spills (e.g., benzene). If this is true, then these chemicals should be found at concentrations much less than those predicted by the current generation of screening level algorithms. This hypothesis is supported to some degree by the Fitzpatrick and Fitzgerald (1997) Massachusetts indoor air survey, the data of Fischer et al. (1996), and others who have observed and reported on petroleum hydrocarbon biodegradation in the vadose zone under natural conditions (e.g., Ostendorf and Karnpbell 1991).

Unfortunately, little data exist to refute or support existing algorithms, or to quantify the degree of over-conservatism. This lack of data is a result of many factors, including the fact that interest in this pathway is relatively new. From a comparison of model predictions with published radon intrusion data, Johnson et al. (1991) and Little et al. (1992) conclude that the screening algorithms should predict reasonable results when contaminants are present in soil gas immediately adjacent to a basement (e.g., Nazaroff et al. 1987). Yet to be reported are rigorous comparisons of model predictions and measurements for well-characterized sites where the contaminant sources are located at a distance from the buildings. Recently, Fitzpatrick and Fitzgerald (1997) presented their conclusions from a study of Massachusetts sites where indoor air samples were collected. Their goal was to review site characteristics and then identify specific trends and field conditions that most influence vapor migration and vapor intrusion into buildings. They also were interested in assessing the validity of generic state guidance derived from use of the Johnson and Ettinger (1991 ) algorithm. In summary, they noted that the generic Massachusetts guidelines overestimated vapor intrusion impacts for petroleum fuel hydrocarbon sites; however, they also found that the generic screening guidelines sometimes under-predicted indoor concentrations at sites where chlorinated organic vapors were present. Contrary to the popular belief that the models are overly conservative, the authors concluded that the generic Massachusetts guidelines were not conservative enough for site screening purposes, at least for chlorinated compound sites.

Given limited data and limited understanding, the potential for high sensitivity to site specific conditions, and the tendency to lean toward conservativeness when developing regulations, it seems unlikely that technically defensible alternatives for developing generic screening levels will surface in the near-term. The inevitable consequence is that many sites containing volatile carcinogens are unlikely to satisfy generic RBSLS for this pathway. Thus, this pathway will need to be addressed on a more site-specific basis, and options are needed to ensure that this is done in a technically defensible manner. Some state-level regulatory agencies are already struggling with developing site-specific guidance for assessing this pathway.

In answer to this need, options for addressing the vapor migration pathway on a more site-specific basis are proposed here. These include more refined use of existing screening algorithms for layered geologic settings, as well as use of updated algorithms that consider biodegradation. These options stem from consideration of available data, existing algorithms, theoretical considerations, and empirical experience (Jury et al. 1983, Kampbell et al. 1987, Nazaroff et al. 1987, Garbesi and Sextro 1989, Jury et al. 1990, Johnson et al. 1990, Loueiro et al. 1990, Ostendorf and Kampbell 1990, Johnson and Ettinger 1991, Johnson and Perrot 1991, Hodgson et al. 1992, Little et al. 1992, Unlu et al. 1992, Ostendorf 1993, Jin 1994, Acomb et al. 1996, Auer et al. 1996, Fischer et al.1996, Jeng et al .1996, Lahvis and Baehr 1996, Smith et al. 1996, Uchrin 1996, BP 1997, DeVaull et al. 1997, Li 1997, Sextro 1997, Stout 1997). The data collection and data reduction activities can easily be arranged in a sequence of increasing complexity, increasing data requirements, and likely increasing cost. Whether or not this approach is reasonable and defensible can only be determined by application to actual field sites followed by review of the results and experiences. It is recognized that with application, knowledge will continue to grow, and opinions and recommended practices are likely to evolve and become refined over the next few years.

In order to provide insight to the technical challenges, the reader is first provided an introduction to current approaches for developing generic risk-based screening levels, and a discussion of other technical considerations important to the development of the options described here. This is followed by the proposed array of options for assessing the significance of the vapor intrusion pathway on a more site-specific basis. At the end of this report, a vision for a much simpler site-specific assessment is presented and accompanied by a discussion of the developments necessary to progress toward that goal.