PRELIMINARY MODEL CALIBRATION REPORT EXECUTIVE SUMMARY

October 1996

The U.S. Environmental Protection Agency (USEPA) is currently conducting a study of the Hudson River PCB Superfund Site, reassessing the interim No Action decision that the Agency made in 1984. The purpose of the Reassessment is to determine an appropriate course of action for the PCB contaminated sediments in the Upper Hudson River in order to protect human health and the environment. PCBs (polychlorinated biphenyls) were discharged into the Upper Hudson River from two capacitor plants in Hudson Falls and Fort Edward, New York. The Superfund Site extends from Hudson Falls to the Battery in New York City, a distance of approximately 200 river miles. Unacceptable levels of PCBs in fish tissue have resulted in fishing bans and fishing restrictions throughout the river.

This report provides an update on the mathematical modeling efforts being conducted as part of the Reassessment. It is meant as a preliminary or interim report, in that the purpose of the report is to provide interested parties with information about the data and assumptions that are being used in the models, prior to completion of the actual modeling work. Therefore, many of the conclusions are preliminary and may change as the models are further refined and calibrated. When the models are completed, the modeling results will be presented in the Baseline Modeling Report and model predictions for various remedial alternatives will be included in the Phase 3 Report (Feasibility Study).

Study Objectives

The models described in this Preliminary Model Calibration Report were designed to answer the following questions:

1. When will PCB levels in the fish population recover to levels meeting human health and ecological risk criteria under continued No Action?

2. Can remedies other than No Action significantly shorten the time required to achieve acceptable risk levels?

3. Are there contaminated sediments now buried and effectively sequestered from the food chain which are likely to become "reactivated" following a major flood, resulting in an increase in contamination of the fish population?

The overall goal of the modeling analysis is to develop and field validate useful and scientifically credible mass balance models in order to answer these questions. The modeling approach is based on the principle of conservation of mass, that is, the quantity of material that enters a section of the river must be equal to the quantity of material that leaves the section, plus any internal sources or minus any environmental losses. A large body of information from site specific field measurements, laboratory experiments and a search of the scientific literature was synthesized within models for the Upper Hudson River and the tidal freshwater portion of the Lower Hudson River. Models were developed for the transport and fate of PCBs in the water column and sediments, and for PCB body burdens in fish. The integration of these different models allows for the simulation of transport and fate of PCBs that enter the river from the upstream boundary at Fort Edward, from various tributaries, across the airwater interface and across the sediment water interface.

Transport and Fate Model Development

The overall concept involved the development and application of a set of individual models to describe hydrology, solids dynamics and PCB dynamics in the river water and sediments. The principal time frame of interest is from 1983 through 1994. Diverse and extensive data from numerous sources were used in developing and calibrating the models.

The Reassessment database contains information from: USEPA, New York State Department of Environmental Conservation (NYSDEC), U.S. Geological Survey (USGS), General Electric (GE) and private and academic research investigators. The most intensive datasets available are from the USEPA Phase 2 investigations conducted in 1993 and 1994. The USEPA database for the Reassessment is described more fully in the Database Report which was issued in October, 1995. The database itself was issued in March, 1996, and is available on CDROM.

Upper Hudson River PCB Model

The Upper Hudson River PCB Model (HUDTOX) is a mass balance model that includes hydrology, solids and PCBs in river water and sediments. HUDTOX was applied to the Upper Hudson River from the northern tip of Rogers Island (upper end of Thompson Island Pool) to Federal Dam at Troy. HUDTOX provides the ability to simulate total PCBs, as well as specific PCB congeners (BZ#4, BZ#28, BZ#52, BZ#[90 + 101] and BZ#138), based on their particular physical and chemical properties. To date, HUDTOX has been calibrated to field data for the period January 1 through September 30, 1993, coinciding with the USEPA Phase 2 monitoring program.

Thompson Island Pool Hydrodynamic Model

This model computes localized water velocities corresponding to different river flows for areas within Thompson Island Pool (where the contaminated sediments are most concentrated). Results from this model are used as input to the Thompson Island Pool Depth of Scour Model.

Thompson Island Pool Depth of Scour Model

The quantity of sediments likely to become "reactivated" (scoured) following a major flood depends on the velocity of the river flow. River velocities are computed using the Thompson Island Pool Hydrodynamic Model. The Thompson Island Pool Depth of Scour Model is used to determine the range of scour depths and quantities of resuspended (scoured) solids and PCBs during high flow events. The maximum flow simulated using this model corresponds to a 100-year flood.

Lower Hudson River PCB Model

An existing mass balance model developed by Thomann et al., (1989) was used for hydrology, solids and PCBs in Lower Hudson River water and sediments. The Thomann model was applied to the portion of the Hudson River below Federal Dam at Troy. The model represents total PCBs in terms of the sum of individual PCB homologues. The model was validated using revised PCB loads over Federal Dam without the need for recalibration of the original model parameters.

[Note: EPA understands that as of September 1996, the Thomann model is being updated under a grant from the Hudson River Foundation, and that certain modifications have been made to the published model. EPA is evaluating whether the updated model will be available or appropriate for use in the Reassessment.]

Development of Fish Body Burden Models

The overall concept involved development and application of a set of models for relating body burdens of PCBs (expressed as Aroclor equivalents, individual congeners or total PCBs) in fish to exposure concentrations in Hudson River water and sediments.

Bivariate Statistical Model

The Bivariate Statistical Model relates measured PCB levels in water and sediments (two variables, or "bivariate") to measured PCB levels in fish. This model was applied to the Upper Hudson River and to a segment of the Lower Hudson River near Albany. The Bivariate Statistical Model was developed using the historical PCB Aroclor database.

Probabilistic Bioaccumulation Food Chain Model

The Probabilistic Bioaccumulation Food Chain Model relies upon feeding relationships to link fish body burdens to PCB exposure concentrations in water and sediments. The model combines information from available PCB exposure measurements with knowledge about the ecology of different fish species and the relationships among larger fish, smaller fish, and smaller animals in the water column and sediments. The Probabilistic Model was developed using both historical and current field data, and was applied to the Upper Hudson River and to a segment of the Lower Hudson River near Albany. In contrast to the Bivariate Statistical Model, which provides average body burden estimates, the Probabilistic Model provides information on uncertainty and variability around these average estimates.

As part of the development of the Probabilistic Model, speciesspecific profiles (i.e., descriptions of feeding behavior, range and movement) were developed for Yellow Perch, Largemouth Bass, Pumpkinseed Sunfish, Brown Bullhead, White Perch, Spottail Shiner, Shortnose Sturgeon and Striped Bass. These profiles include characteristics that could potentially affect bioaccumulation of PCBs.

Thomann Food Chain Model

The Thomann Food Chain Model is part of the PCB transport and fate model for the Lower Hudson River. The model links PCB exposure concentrations to PCB body burdens in White Perch and Striped Bass. The Thomann model does not explicitly consider the effect of contaminated sediments on accumulation of PCBs in the food chain.

Principal Report Findings

Since this is a preliminary calibration report, it does not present definitive answers to the principal Reassessment questions. However, a number of preliminary conclusions have been drawn based on the work presented here, including:

• The PCB mass balance model for the Upper Hudson River (HUDTOX) provides a reasonable representation of hydraulics, solids dynamics and PCB dynamics for the period of simulation corresponding to the USEPA Phase 2 monitoring program.

• The principal external loadings of total PCBs to the Upper Hudson River during the period of the HUDTOX simulation were across the upstream boundary at Fort Edward (74 percent).

• During the period of the HUDTOX simulation, the model computed a large gain in total PCB mass across Thompson Island Pool between Fort Edward and Thompson Island Dam.

• Large increases in water column concentrations of dissolved phase PCBs, especially for lower chlorinated congeners, are observed to occur across Thompson Island Pool. These increases appear to be caused by an internal source within Thompson Island Pool.

• The processes controlling PCB dynamics in Thompson Island Pool are not fully understood at the present time. One hypothesis that could explain the large increases in PCB concentrations across the pool is sediment water advective flux of pore water PCBs due to groundwater inflow. Such a pore water advective flux would be relatively more important for lower chlorinated PCB congeners due to their greater water solubilities.

• The Thompson Island Pool Hydrodynamic Model produced results that were in good agreement with available information for river flow velocities and water elevations.

• For a 100-year flood event, the Thompson Island Pool Depth of Scour Model predicts that 1,838,600 pounds of solids and 55 pounds of total PCBs will be scoured from the cohesive sediment areas. This mass of PCBs represents less than 1 percent of the total reservoir of PCBs in the cohesive sediment areas of Thompson Island Pool, based on measurements of the inplace reservoir of PCBs from the 1984 NYSDEC survey.

• For a 100-year flood event, the Thompson Island Pool Depth of Scour Model predicts a median depth of scour of 0.41 inches in the cohesive sediment areas. Considering the uncertainty in model predictions, the average depth of scour for this event could range from 0.08 to 2.46 inches.

• The Bivariate Statistical Model for fish body burdens indicates the relative importance of both water column and local sediments as pathways for bioaccumulation of PCBs in Upper Hudson River fish. Reported Aroclor 1016 burdens are mainly attributed to water column concentrations for all species. Reported Aroclor 1254 burdens, which include more highly chlorinated PCB congeners that tend to accumulate in fat, show a wide range in the relative importance of water column and sediment pathways among different species. Results for Aroclor 1254 are consistent with species feeding behavior: for species that feed in the water column, the water column pathway tends to dominate, while for bottom feeders, the sediment pathway tends to be dominant. Fisheating species at higher levels in the food chain appear to accumulate Aroclor 1254 from both water column and sediment pathways.

• The Probabilistic Bioaccumulation Food Chain Model indicates that water pathways contribute significantly to PCB body burdens in forage fish (including Pumpkinseed Sunfish) and Yellow Perch. Water and sediment are both important for Largemouth Bass, and sediment is the main exposure pathway for the bottom feeding Brown Bullhead. These results compare favorably with results from the Bivariate Statistical Model. Results from the original Lower Hudson River modeling effort by Thomann et al., (1989) were successfully reproduced.

Future Baseline Modeling Efforts

The conclusions presented in this Preliminary Model Calibration Report indicate that significant new understanding has been gained about PCB transport, fate and bioaccumulation in the Hudson River. The modeling work for this Reassessment is continuing and more definitive conclusions will be presented in the Baseline Modeling Report. The purpose of this future modeling work is to reduce uncertainties contained in the preliminary models. Future plans include continued development of both the transport and fate mass balance models, and the fish body burden models. They also include applications of these models to additional field data that became available after completion of this preliminary model calibration work.

Future work with the HUDTOX model will include development of a more finely resolved spatial segmentation grid for Thompson Island Pool, application to daily suspended solids data collected during the Spring 1994 highflow period and re calibration using the complete, validated, Phase 2 field data for 1993. Finally, a longterm (19841993) hindcasting calibration will be conducted to confirm the predictive capability of the model over a decadal time scale.

Future work with the Thompson Island Pool Depth of Scour Model will include extension of the modeling framework to include both cohesive and noncohesive sediment areas, and application to the complete, validated, Phase 2 field data for 1993.

Future work with the fish body burden models will include application to NYSDEC 1995 fish data, further analysis of exposure pathways involving water column invertebrates and exploration of patterns of congener uptake between and among different fish species. In addition, use of a model based on fugacity (Gobas, 1993), or chemical potential, will be explored.

The final version of HUDTOX will be used to simulate PCB concentrations in the water column and sediments due to No Action and various flood events. For these simulations, the output of the HUDTOX model for the Upper Hudson River will be linked to the Thomann model for the Lower Hudson River. In turn, the PCB outputs from these Upper and Lower Hudson River models will be linked to the fish body burden models. Finally, predictive modeling simulations to evaluate the effects of various remedial scenarios will be presented as part of the Phase 3 Report (Feasibility Study).

For information about this page, contact: kluesner.dave@epa.gov