Ambient PCB reduction through use of activated carbon in the Lower Duwamish Waterway

Abstract

Decades of industrial activity along the Lower Duwamish Waterway have resulted in significant pollution of the sediments, water, and aquatic life in the area. Through the U. S. Environmental Protection Agency’s (EPA) Superfund Program, the Lower Duwamish Waterway is being remediated for numerous Contaminants of Concern, including Polychlorinated Biphenyls (PCBs). The Record of Decision, published in November 2014, identified Enhanced Natural Recovery as one remedial technology slated for implementation to help reduce through sequestration dissolved (and therefore readily bioavailable) concentrations of contaminants in the waterway. As part of the implementation, activated carbon is proposed as an amendment to Enhanced Natural Recovery pending the results of design phase pilot testing. A hydrophobic organic contaminant mass transfer model, developed by Luthy, Choi, and Cho (2013), was utilized to predict PCB concentration reductions in sediment pore water for three PCB congeners. The model relied upon a series of parameters specific to sediment characteristics, properties of the activated carbon, and kinetic characteristics specific to the PCB congener modeled. A combination of Lower Duwamish Waterway-specific parameters and default parameters set in the model was used. Model results predicted a 69% reduction in PCB 180 concentrations over a one year period under the most probable scenario for activated carbon. Similar porewater reductions of 68% and 81% were predicted for the PCB 153 and PCB 101 congeners, respectively. The mass transfer model was then applied to field data from Puget Sound Naval Shipyard, another sediment site with PCB contamination where activated carbon was placed by the U.S. Navy in 2012 as part of remedial efforts at the site. Results from the model were compared to actual concentration reductions seen at Puget Sound Naval Shipyard at the ten month post construction period. The model closely predicted a PCB 180 concentration reduction of measured reduction seen at the site. However, model results for PCB 153 and 101 under predicted actual reductions but fell within the 95% confidence intervals of the observed mean results. The results of this predictive modeling have implications for remedy implementation in the Lower Duwamish Waterway.