Yesterday, EPA released the final draft of the Remedial Investigation (“RI”) for the Lower Duwamish Waterway Superfund Site. The complete text is available here. The Duwamish Superfund Site consists of the lower five miles of the Duwamish River, which empties near the Seattle waterfront. EPA listed this site in 2000 in response to contamination concerns, and four parties (Boeing, the City of Seattle, King County and the Port of Seattle) agreed shortly thereafter to conduct a Remedial Investigation and Feasibility Study for the site. I worked on matters related to this site while at my former law firm, and find it to be a fascinating intersection of science, policy decisions, and law.

This RI is the result of investigations that have been conducted since 2000, and data available from as far back as the early 1990s. It is intended to describe what is known about the Duwamish site, including the history of the waterway; dynamics of sediment transport and deposition within the waterway; the distribution of contaminants in sediments, water and tissues within the waterway; information about past and ongoing sources of chemicals to the waterway; and the results of both human health and ecological risk assessments performed using the information gathered about the site.

The RI is massive. The main text alone is over 800 pages in length, and it would be impossible to summarize it completely in a blog post. What I want to highlight is how decisions made with regard to the human health risk assessment (“HHRA”) dictate further decisions regarding the cleanup. The results from the HHRA will drive the cleanup of the Duwamish because risk-based cleanup levels derived from this risk assessment will be much lower than those from the ecological risk assessment, and will also be much lower than those derived from other possible sources of cleanup levels, i.e., human health concerns will become the primary driver in remediating the site. This HHRA highlights the intersection of issues such as treaty rights of tribes, conservatism in assessing risk, and how EPA handles uncertainty in making policy decisions regarding cleanup implementation. There is a good summary of the human health risk assessment in the Executive Summary of the RI, starting at Page ES-16, and the complete text of the HHRA is available as a lengthy appendix to the RI. The HHRA evaluates numerous different exposure scenarios, including dermal contact and different levels of fish consumption. The “driver” scenario is the one defined as “reasonable maximum exposure” (“RME”) and is based on seafood consumption rates from a study of the Tulalip Tribe.

Right now, available data for the Duwamish indicates that not much seafood is being harvested from the Duwamish. EPA believes this is a result of closures for shellfish due to bacteria, possible pollution concerns, and more desirable fishing locations adjacent to the Duwamish. The RME does not represent a present-day scenario, and is instead intended to represent upper-end seafood consumption in the future, after the cleanup has occurred. Some of the specifics of the RME:

  • Salmon consumption was not included in the RME because adult salmon accumulate contaminants in the open ocean or Puget Sound, and only a small amount of contaminants accumulate in juvenile salmon born in the Duwamish.
  • Both adult and child exposure were based on consumption rates developed from surveys of the Tulalip Tribe. For adults, it was assumed that 97.5 grams of seafood from the Duwamish were ingested per day, with 13.1 meals per month, over 70 years. Children were assumed to ingest 39.0 grams per day, with 5.2 meals per month over a 6-year period.
  • The “diet fraction” or fraction of seafood that comes from the Duwamish, was set at 1, i.e., the scenario assumes all of the seafood being consumed is coming from the Duwamish.
  • Contaminants of concern in fish and shellfish assumed to be consumed under this scenario included PCBs, arsenic, dioxins and cPAHs.
  • EPA calculated that the excess cancer risk (i.e., number of cancers attributable to this exposure scenario at current concentrations of contaminants) from the site is 3 in 1,000 for PCBs. EPA considers a 1 in 10,000 to 1 in 1,000,000 excess cancer risk acceptable for CERCLA cleanups. Risks from arsenic, dioxins and cPAHs were likewise found to be elevated based on EPA’s calculations.

Implications for the Cleanup

Although the RI is not intended to set cleanup levels (which EPA will select when it chooses the ultimate remedy for the site), the RI contains some calculations on what sort of risk would be associated with different sediment concentrations post-cleanup. It does this by connecting sediment concentrations of contaminants to tissue concentrations in fish through the use of a food web model.

These “Risk-Based Threshold Concentrations” (RBTCs) provide some guidance on what possible cleanup levels may be. For PCBs, a 1 in 10,000 excess cancer risk post-cleanup equates to a 1-25 microgram per kilogram dry weight PCB concentration in sediment. The 1 in 100,000 and 1 in 1,000,000 risk level sediment concentrations are all below 1 microgram per kilogram PCBs in sediments. In fact, sediment concentrations at these risk levels could not be calculated because even if you set sediment concentrations at zero in the model, tissue concentrations in fish would still exceed acceptable levels due to the background level of PCBs in water alone.

To put these RBTCs in perspective, regular urban sediments often have concentrations in the range of 50-to-100 micrograms per kilogram, mostly due to atmospheric deposition of PCBs and other diffuse, urban sources. Stated differently, the results of the HHRA for the Duwamish indicate that it would be necessary to lower sediment concentrations up to 100 times below these background levels to achieve the lower end of what EPA considers acceptable risk at a remediated Superfund site. This is something that is most likely impossible to achieve.

How this plays out in terms of the ultimate cleanup is still to be determined. Under state regulations, where a risk assessment yields contaminant concentrations below background, cleanup levels are set at background, which the draft Feasibility Study (FS) states is in the 50-to-100 microgram per kilogram level. The draft FS—which is expected to be finalized later this year by EPA—estimates that achieving the 100-microgram per kilogram level would involve dredging 315 acres of the river over a 43-year time period, at an estimated cost of $1.28 billion. Cleanup alternatives of lesser scope are anticipated to cost $250 to $650 million and take 15 to 30 years to complete (see page 20 of the draft Executive Summary of the FS). Where that money will come from is a very good question—EPA has identified approximately 250 “potentially responsible parties” to date, and even on the low end this equates to an average of $1 million per party, assuming all 250 parties have links to the site, and are found to be liable under CERCLA.

Given the stakes, the direction of the Duwamish cleanup will certainly shape environmental law and policy in the Pacific Northwest for years to come. The human health risk assessment approach used here will likely be replicated at other sediment sites, with correspondingly interesting challenges created by the low cleanup levels that result from these high seafood consumption rates. In fact, the same approach has already been used at the Lockheed Site located downstream of the Duwamish next to Harbor Island. Similar issues are being raised regarding fish consumption rates for the Portland Harbor Superfund Site in Oregon. For instance, EPA’s comments on the draft of the HHRA for that site (see page 3) suggest that consumption rates even higher than those used in the Duwamish may be applied for Portland Harbor. Businesses that end up involved in these types of cleanups will be facing potentially enormous cleanup costs, and, as evidenced by the length of times estimated to remediate the Duwamish, many years of implementing these cleanups.