Discrete, composite and multi-incremental samples (MIS) samples of surface and subsurface soil were collected around 24 fence posts (eight posts per sector), under the fence posts (by digging out the posts) and analyzed for metals and SVOCs. The samples were collected around the corners of each post because the post footing concrete typically had large cracks that would likely result in a preferred contaminant migration pathway from the post to underlying soil.

Soil sampling was conducted following the ITRC (2012) guidance for incremental sampling methodology. Linear decision units (DUs) were defined as approximately 100 foot lengths of fence, separated into interior and exterior portions on either side of the fence and extended laterally about 3.5 to 6 feet from the fence and vertically to depths of 1 ft. ISM was completed using a systematic random sampling approach. Each end of the DU boundaries were measured and recorded in the field using a handheld global positioning system (GPS) unit with submeter accuracy. The boundaries of each DU were marked before sampling. The linear DUs were typically staked with string to maintain the sampling line while drilling. All data (discrete, composite and MIS) were used for the risk assessment.

For six individual sites (focusing only on the three fence sites as sources of CCA), the total concentrations ranged from 0.96 to 957 mg/kg while the 95% UCL of the mean calculated for each location ranged from 27 mg/kg up to 478 mg/kg. Table 11-6 includes the arsenic concentrations in soil.

Table 11‑6. Arsenic concentrations

11.9.2.1 Soil Type Read More

11.9.2.2 Source of Arsenic Read More

11.9.2.3 Future Land Use Scenarios Read More

IVBA testing of 10 surface soil samples was performed by Dr. Nick Basta at Ohio State University in Columbus, Ohio, based on a simulated gastric fluid extraction procedure which mimics a gastric pH of 1.5 (the CAB method prior to its publication). Costs were about $200 per sample.

Electron microprobe analysis (EMPA) of two surface soil samples was performed by Dr. John Drexler of University of Colorado at Boulder, Colorado, to analyze the chemical speciation of arsenic present in the site soils to identify the dominant form of arsenic present. Costs were approximately $500 per sample.

The bioaccessibility estimate (%IVBA) ranged from 51% to 64% (mean of 58%). The IVBA analyses were used to develop a site-specific RBA value indicative of the potential for arsenic absorption in the gastro-intestinal tract for humans (its bioavailability) based on the form of arsenic present in site soils. Higher RBA estimates (closer to 100%) indicate a higher potential for arsenic absorption and exposure. The site-specific RBA estimate of 47.4% was used to develop site-specific arsenic risk-based concentrations (RBCs) for human receptors. This value was calculated using the mean IVBA estimate of 58% in the following literature-based equation based on in vitro to in vivo correlation for a combination of agricultural and mining soils: %RBA = 0.72 (%IVBA) + 5.64 (Bradham et al. 2011).

The qualitative findings from the EMPA results indicated that the dominant form of arsenic was associated with manganese (III) oxyhydroxide (MnOOH), representing 98% of the relative arsenic mass. Similar iron oxide particles represented the remaining relative mass of arsenic. The MnOOH contained approximately 0.5% arsenic by weight and 2% copper by weight. The likely source for the arsenic was identified as a copper arsenate compound (consistent with site history), that, when weathered, released its arsenic and copper to be sorbed onto the MnOOH in the soil. Some of the particles of MnOOH are directly related (cemented) to organic (plant) material. Almost all of the arsenic in iron oxides and manganese oxides may be available for solubilization and may be considered bioaccessible. These qualitative findings were refined using IVBA.

A value of 47.4% was used to develop the target concentrations for site soils by applying the %RBA to the ingestion portion of the back-calculated equation for direct contact with soil. The site-specific RBA is slightly lower than USEPA’s current default value of 60%.

After applying the site-specific RBA to the oral ingestion exposure route, residential human health cleanup goals for arsenic that corresponded to the target cancer risk levels of 1 x 10^-6, 1 x 10^-5, and 1 x 10^-4 were estimated at 1, 10, and 100 mg/kg of total inorganic arsenic, respectively. In comparison, site-specific background values for arsenic ranged from 1.88 to 3.22 mg/kg among the different soil types.

The arsenic cleanup goals for human health, reflective of the site-specific RBA, demonstrated that arsenic exposure under a hypothetical residential scenario is not the driving risk pathway for the site, and cleanup ended up being based on protection of sensitive plant species (51 mg/kg).

The estimated remediation volume and cost using HHRA cleanup goal for arsenic of 1 mg/kg based on the default RBA (60%) was 20,000 cubic yards and $15 million. The estimated cost of the RBA study was $27,000 total ($12,000 for sampling and analytical and $15,000 for risk assessor labor). The estimated remediation volume and cost using ecological cleanup goal of 51 mg/kg was 8,000 cubic yards and $8 million. Therefore, the Air Force spent $27,000 to reduce the expected remediation volume from 20,000 cubic yards to 8000 cubic yards, which equates to a total reduction in costs of approximately $7 million.

A key regulatory challenge on the project originated from the different federal and state policies governing risk assessment. Specifically, new Air Force policy regarding risk assessment as of June 2015 dictates that risk assessments at Air Force Facilities in California no longer follow California State guidance in the cleanup of non-NPL facilities, particularly guidance relating to the choice of toxicity values and the cancer risk that may trigger a response action. The California agency’s request for the Air Force to calculate human health cleanup goals for the residential scenarios using both federal and state exposure and toxicity values to estimate risk was declined by the Air Force as it is contrary to the new policy.

However, as part of the responses to state comments on the HHRA, arsenic cleanup goals for the residential scenario were estimated for target risk levels of 1 x 10^-6, 1 x 10^-5, and 1 x 10^-4 using state exposure factors and toxicity values. These state-based cleanup goals were equivalent to or below the federal cleanup goal for arsenic based on a 1 x 10^-4 target risk level and were also below the site-specific background concentration for arsenic. The state does not require cleanup to concentrations less than site-specific background, and use of the federal-based cleanup goals satisfied the state’s cleanup requirements consistent with unrestricted land use. Therefore, the state maintained a policy disagreement with the approach used to develop the human health risk assessment, but acknowledged that the proposed soil cleanup volume for the project based on the ecological cleanup goal (lower than the federal human health cleanup goal) meets the state’s acceptable risk management range for unrestricted land use and allowed the bioavailability-based decisions to stand. This decision allowed all agencies to agree that the selected remedy was protective of human health for this site and the reduced volume of excavation was used for execution of the feasibility study.