Year Sampled from 1.6 ppm to 41 ppm PCBs (U.S. Environmental Protection Agency, 2000c; New York State Department of Environmental Conservation, 2001a). For the same species in the Lower Hudson within eleven miles of the Federal Dam, concentrations range from 1.1 ppm to 11 ppm. PCB concentrations vary greatly within each species, and PCB levels in individual fish have exceeded these mean values by several fold. In recent years, maximum PCB concentrations in fillets from individual Hudson River fish have been found to be as high as 480 ppm in common carp, 290 ppm in white sucker, 160 ppm in American eel, 150 ppm in largemouth bass, 50 ppm in red-breast sunfish, 42 ppm in walleye, 39 ppm in smallmouth bass, 37 ppm in brown bullhead, 30 ppm in yellow perch, and 27 ppm in black crappie (NOAA letter, 2000). In the lower Hudson River, recent maximum concentrations of77 ppm in shortnose sturgeon liver, 42 ppm in Atlantic sturgeon gonad, and 31 ppm in striped bass fillet have been documented (NOAA letter, 2000) Fewer data are available for wildlife species other than fish, but several bird and mammal species sampled near the Hudson River also exhibit increased levels of PCBs in their tissues (McCarty and Secord, 1999; Secord and McCarty, 1997; Secord et al., 1999; New York State Department of Environmental Conservation, 2001b).

For Upper Hudson River fish, PCB concentrations declined substantially between the 1970s and 1980s and experienced an increase in the early 1990s due to the Allen Mill event (the collapse of a wooden gate structure adjacent to the riverbank at the GE Hudson Falls plant site that resulted in a release of PCBs). The most recent data (Fig. 24.2) show considerable year-to-year variability and less obvious declining trends (New York State Department of Environmental Conservation, 2001a,b). Comparing these tissue burdens of PCBs with published guidelines demonstrates that the current levels of exposure of fish and wildlife in the upper Hudson River drainage basin are high enough to cause concernfor environmental effects. To protect piscivorous wildlife in the Great Lakes, a guideline for total PCB loads in fish of approximately 0.1 ppm was recommended by the International Joint Commission (Newell, Johnson, and Allen, 1987; International Joint Commission, 1989;

Canadian Council of Ministers of the Environment, 2001).

Some of the strongest evidence of adverse PCB consequences to fish-eating animals has been documented for mink and otter, two mammals that are especially sensitive to PCBs (Golub, Donald, and Reyes, 1991; Heaton et al., 1995; Halbrook et al., 1999; Moore et al., 1999). When mink eat fish containing PCB levels comparable to those recently and historically reported in the Upper Hudson fish, they experience impaired reproduction, reduced offspring (kit) survival, and reduced kit body weight. Results of three long-term studies in which PCB-contaminated fish were fed to mink allowed development of a dose-response curve relating the rate of PCB ingestion (milligram of PCB ingested per kg body weight per day, mg/kg-day) to a decline in fecundity (Golub et al., 1991; Heaton et al., 1995; Moore et al., 1999; Halbrook et al., 1999). That analysis suggests that a daily dose of 0.69 mg PCB/kg-day (corresponding to approximately 5 ppm PCBs wet weight in their food) will result in a greater than 99 percent decline in mink reproductive fecundity, while approximately 0.1 and 0.025 mg/kg-day (equivalent to approximately 0.7 and 0.2 ppm) will result in 50 percent and 10 percent declines in mink reproduction and fecundity, respectively (Golub etal., 1991; Heaton etal., 1995; Moore etal., 1999;Halbrooketal., 1999).PCB levels in Hudson River fish exceed levels demonstrated to cause reproductive impairment in mink. Moreover, recent analyses of PCBs in the livers of mink and otter collected from the Upper Hudson River valley showed levels in some individuals that exceed values reported to cause negative impacts (New York State Department of Environmental Conservation, 2001 b,c).

Thus, our current knowledge strongly suggests that the health of some sensitive mammalian species, such as mink and otter, may be seriously impaired along the Upper Hudson River. EPA considers otter to be at slightly greater risk than mink, because otter diets have higher proportions of fish, and the agency has designated whole-body fish concentrations of 0.03-0.3 ppm (mg/kg) total PCBs (approximately corresponding to 0.012-0.12 mg/kg total PCBs in fish fillets) as the upper limit for protection of otter (U.S. Environmental Protection Agency, 2000c). Fish concentration goals designed to protect mink and otter should afford protection to the other less sensitive species that inhabit the Hudson River ecosystem. A corollary to this is that the less sensitive species should recover sooner in response to decreasing PCB levels in the Hudson River than the more sensitive species.

Besides being a source of PCB contamination to consumers, fish themselves are vulnerable to these chemicals. Recommended levels for protecting fish from exposure to PCBs range from a median threshold value of 1.1 ppm total PCBs in whole body (Meador, Collier, and Stein, 2002) to 25-70 ppm in adult fish liver (Monosson, 2000), and 5-125 ppm in the body of fish larvae. Current levels of PCB contamination in Upper Hudson River fish often exceed those associated with health effects on fish and wildlife. Because PCBs have such wide-ranging effects on the health of biota, and are so persistent once exposure occurs, it is very likely that current levels of contamination are causing injury to species that depend on the Upper Hudson River ecosystem.

KEY QUESTION 2. Are the PCBs in the upper Hudson River sediments an important continuing source of contamination to the lower Hudson under average flow conditions?


1. PCBs leaking from the GE plant sites and re-mobilized from the sediments continue to add PCBs to the Hudson River and to the food chain.

2. In recent years, contaminated sediments have become the dominant source of PCBs to the river. As a result of source controls being implemented at the plant site, contaminated sediments are expected to serve as the dominant source of PCBs to the river for years to come.

3. Analyses conducted to date by both GE and EPA using relatively coarse-scale numerical models (QEA, 1999; U.S. Environmental Protection Agency, 2000d) lack the required fine-scale spatial resolution in the sediment transport model and use of an overly simplistic PCB distribution and bioaccumulation model. These deficiencies limit the ability of either model to accurately project future

Figure 24.3. PCB concentrations in weeklywater sample collections at the upstream (A) and downstream (■, ♦) ends of the Thompson Island Pool in 1997 through 1999. Increased levels at the downstream end indicate that the contaminated sediments in the Thompson Island Pool are the major current source of PCBs to the Upper Hudson River water column, contributing on the order of 180 kg/y. (Plot prepared by Jennifer Tatten as part of RPI (2001) based on data from General Electric Company as reported in the database supplied by GE to NYSDEC).

Figure 24.3. PCB concentrations in weeklywater sample collections at the upstream (A) and downstream (■, ♦) ends of the Thompson Island Pool in 1997 through 1999. Increased levels at the downstream end indicate that the contaminated sediments in the Thompson Island Pool are the major current source of PCBs to the Upper Hudson River water column, contributing on the order of 180 kg/y. (Plot prepared by Jennifer Tatten as part of RPI (2001) based on data from General Electric Company as reported in the database supplied by GE to NYSDEC).

PCB levels in the Upper Hudson River, with or without active remediation. More sophisticated field evaluations and models would greatly improve efforts to define and monitor the remediation of the Hudson River.

The current releases of PCBs from the GE facilities are substantially less than those during active operation of the plants. GE has spent and will continue to spend considerable amounts of money to stem the flow of PCBs from their properties. Nonetheless, given the large amount of PCB contamination on these sites, and their immediate proximity to the Hudson, a small but significant amount of PCBs is expected to continue to enter the river from the GE sites for many years. Based on the amount of PCB in the river near the GE facilities, this small amount of leakage is presently estimated to be no more than 3 ounces per day (or 30 kilograms per year, see next paragraph), whereas the average PCB releases from the facilities were 2,700 to 16,000 kilograms per year between the 1940s and 1977 (United States Environmental Protection Agency, 2000a). In addition to this recognized leakage of PCBs from the plant sites, PCBs that were previously discharged from the plants and now reside in river sediments downstream from the plants are being released into the river's waters. PCBs may be released from the sediments during resuspension by currents and by diffusion and mixing of PCBs.

To estimate the relative importance of these two sources of PCBs to the Upper Hudson River, we examined monitoring data collected by GE (O'Brien and Gere Engineers, 1998; QEA, 2000, 2001; RPI, 2001) at two locations downstream of their facilities (Fig. 24.3). The first site is at Rogers Island downstream of GE's Fort Edward Plant (Fig. 24.1). Here PCB concentrations are (relatively) low and quite constant. By multiplying the PCB concentration in the river by the river's flow rate at Fort Edward, we estimate that about 30 kilograms of PCBs per year were moving down the river at this point in the late 1990s.1 In contrast, PCB concentrations in the waters passing over the Thompson Island Dam six miles downstream were much higher and more variable than at the upstream site (Fig. 24.3).

1 Weekly PCB mass loadings were calculated as the product of measured PCB concentrations (nanograms/ liter x 1012 nanograms/ kilogram = kg PCB/liter) and the corresponding total flow for the period (usually weekly; cubic feet/second converted to liters/week). The annual mass loading of PCBs was calculated as the sum of weekly loadings.

Figure 24.4. Approximate mass balance for PCB fluxes in the Thompson Island Pool for 1998, based on data shown in Figure 24.3.

From Sediment ~ 150 kg/year

Figure 24.4. Approximate mass balance for PCB fluxes in the Thompson Island Pool for 1998, based on data shown in Figure 24.3.

These higher PCB concentrations, multiplied by the river flow, yield an estimate of 180 kilograms of PCBs per year passing over the Thompson Island Dam. We conclude, therefore, that about 150 kilograms of PCBs per year enter the river as it moves through the Thompson Island Pool (Fig. 24.4). The only plausible source of these PCBs is release from the Thompson Island Pool sediments. These sediments are highly contaminated with PCBs that can be released into the water column under a variety of flow conditions and there are no other likely significant PCB sources to this stretch of the river. It is important to note that these releases have occurred during relatively typical flow conditions.

These observations are consistent with our understanding of PCB behavior in rivers. Measurements of PCBs in the river indicate that the release of PCBs from sediments in the Upper Hudson River, includingthose belowthe Thompson Island Pool, is currently occurring and that this release is the dominant source of PCBs to the Hudson River downstream from the GE facilities at Hudson Falls and Fort Edward. GE has asserted that this current ongoing PCB supply is transient, resulting from the contamination of near surface sediments in the Thompson Island Pool (and, presumably, a number of other spots downstream) by the Allen Mill gate failure (1991) and more recent releases from the plant sites. Both the GE and EPA models indicate that the Thompson Island Pool is a region of net deposition and GE maintains that its ongoing program to control releases from the plant sites will lead to burial by relatively clean materials in short order, isolating these sediments and associated PCBs from the overlying water column (see QEA, 1999; also GE interpretation of model projections - As a result, the company says, dredging of contaminated sediments would be counterproductive, invasive, and expensive because it could expose deeply buried, highly PCB-contaminated sediment layers and increase downstream transport of contaminated sediments. If this were the case, monitored natural attenuation of PCB impacts by allowing new sediments to bury the contaminated sediments within the Thompson Island Pool and elsewhere would clearly be the preferred course.

If the Thompson Island Pool were a quiescent area of net deposition, one would expect that the sediments would accumulate in a rather orderly fashion, layer by layer, forming a stable, stratified deposit in which the deeper, older sediments and their associated contaminant burden would be efficiently isolated from the surface layers and the overlying waters. Transport and material exchange would be confined to the immediate surface layers even during the extreme flow events. This idealized description of sediment accumulation, however, is not consistent with the bulk of the available data. While there are a few sediment cores that show orderly and progressive deposition as evidenced by radionuclide dating, there are many more showing a disturbed and irregular sediment column in which the record of sediment accumulation cannot be readily deciphered (Bopp et al., 1985;United States Environmental Protection Agency, 1997; O'Brien & Gere Engineers, 1999). In contrast to the well-ordered cores, these irregular distributions of properties provide clear indication that significant areas of the sediment deposit resident within the Thompson Island Pool are subject to time-variant disturbance involving vertical distances similar in magnitude to the observed depths of contaminant burial. When viewed collectively (rather than selectively), these disparate field data indicate that the Thompson Island Pool sediment deposit is not an ordered, stratified mass with near horizontal uniformity in sediment properties, but rather is more accurately described as a spatially heterogeneous "patchwork quilt." In this deposit, sediment characteristics and the associated PCB concentrations display significant spatial variability. These variations affect the ability of the sediment to be moved by bottom currents under average ambient flows as well as the aperiodic high energy storm event. As a result, a given flow condition might find some areas of the pool experiencing net deposition while other areas erode. A change in flow state could significantly alter the locations of deposition and erosion and might change the pool from net depositional to net erosional or vice versa.

We believe that the heterogeneous nature of the Thompson Island Pool sediment deposit in space and time makes it impossible to specify the "age" of the PCBs being added to the water passing by. Whether the PCBs being added to the water at present are simply remnants of those introduced by the Allen Mill gate failure or contaminants introduced much earlier and subsequently remobilized by physical andbiologicalprocesses, or some combination of these two sources, cannotbe accurately determined from the field data alone. Nonetheless, we conclude that under the prevailing average flow conditions the sediments of the Thompson Island Pool are a continuing source of PCBs to the overlying waters.

Having concluded that PCB release from the sediments of the Upper Hudson River is the dominant current source of these contaminants to the water column and food web, the next question is how long this condition willpersist in the river. Will PCBs continue to bleed from the sediments indefinitely, or will natural processes gradually sequester the PCBs within the river's sediments? If all of the PCBs in the Thompson Island Pool sediments (approximately 15,000 kg - U.S. Environmental Protection Agency, 2000a) were available to be reintroduced back into the river and the rate of release continued at the present level (150 kg/year), there would be sufficient PCB in the sediment to support release for 100 years. This approximation is not realistic, however, as some of the 15,000 kg are undoubtedly trapped within the sediments, and one would not expect the release rate to remain constant in the face of declining PCB inventories in the sediments. To refine this estimate requires a coherent understanding ofthe movements ofwater, sediments and PCBs in the river, as well as addressing the difficult problem of quantifying remobilization of sediment. Predicting the future consequences of environmental actions is quite difficult, especially in a dynamic river system that has already been altered through the construction of locks and dams, reservoirs, canals, and dredged channels. Numerical models are tools used to estimate how PCB levels in the Hudson River sediments, water, and biota will change in the future, with and without active remediation. If the main motivation for active remediation is to reduce PCB levels in the future, our ability to design and evaluate the effectiveness of proposed remediation depends almost entirely on the accuracy of such models.

Both GE and EPA have developed numerical models that describe PCB transport in the Upper Hudson River (QEA, 1999; U.S. Environmental Protection Agency, 2000d). While these two models share many similarities, there are also some key differences related to the extent of PCB release from the sediments. The two models predict similar levels of PCBs in the Upper Hudson River during the next several decades.2 Both models predict slowly declining PCB levels in the Upper Hudson River over the next several decades as the system continues to respond to the gradual depletion of PCBs in the 'active' layer of sediment. In other words, the results of the models are driven by the underlying

2 Fora side-by-side comparison of the USEPAand GE models, see pages 143-5 in National Research Council (1983).

assumption that the sediments are a source of PCBs to the river water, but that the magnitude of this source will gradually decrease over the next several decades. This decrease results from the continued burial of PCBs by ongoing deposition of clean "new" sediment and from the release into the overlying water. Neither model predicts that the PCB levels will approach zero within the next 65+ years, reflecting both the likely continual chronic release of PCBs from upstream and the inherently slow response time of the system. As discussed above, it is not clear to us that the Thompson Island Pool is net depositional. Therefore, we question whether ongoing burial will significantly deplete PCBs in the surface sediment as fast as predicted by these models.

Because the long-term recovery of the river from PCBs depends explicitly on the amount of PCBs in the river sediments and the rate at which these PCBs are removed from the active surface sediment, our ability to assess the future course of PCB levels in the Hudson River, with or without active remediation, depends upon our ability to model sediment transport processes. This is a challenging exercise because the sediment transport regime with the upper river is highly dynamic and is significantly variable in space and time. River sediments are constantly being reworked and those which settle in one location are often later resus-pended and displaced. A fraction of these materials may accumulate within the Thompson Island Pool, while others move downstream. The extent of this "trapping" of sediments within any stretch of river is difficult to estimate. The retention efficiency of the Thompson Island Pool (that is, the fraction of the solids entering the pool that remain in the pool for long times) is believed to be low, and the associated sedimentation rates are low (on the average of a few tenths of a centimeter per year, averaged over the entire pool) (QEA, 1999). Temporal variations in sediment transport and accumulation result in a heterogeneous sediment deposit whose characteristics vary significantly over small vertical and horizontal distances. As a result, the bottom throughout the upper river is a complex mosaic of fine sands, silts, clays, wood chips, and other organics formed by the combination of constantly changing currents and sediment supplies. Predicting sediment andPCB transport withinsuch a system requires the use of a numerical model with sufficient spatial resolution to accurately represent this heterogeneity. Unfortunately, the models used by both the EPA and GE employ relatively coarse spatial segmentation that effectively masks the heterogeneity of the river bottom. Only the GE model attempts to address the complexities associated with the transport of sediments of mixed composition. This approach, although commendable, is essentially untested, leaving its accuracy open to question.

In addition, we feel that the numerical models used by both EPA and GE to describe PCB transport and accumulationin biota are too simplistic in their chemical descriptions. Although a large amount of high quality measurements of PCB components were made in the Hudson River, the models treat the complex and variable mixture of PCB components as a single 'chemical' (called Tri+ PCB, equal in concentration to the sum of all PCB components in the Hudson with three or more chlorines). The behavior of the PCB mixture varies markedly depending on the properties of the individual PCB components, especially as a function of the number of chlorines. The PCB composition changes with space and time in the Hudson. We are concerned that extrapolating a PCB model into the future that has been calibrated primarily on data collected over a relatively short period in which the PCB composition has not varied markedly introduces important uncertainties into the projections of long-term recovery. Based on our knowledge of PCB behavior, we believe that the recovery time of the more highly chlorinated PCB congeners (those that accumulate most in the food web) could be longer than that projected by the models.

Both the EPA and GE models appear to reasonably match previous field measurements. One should not conclude from this general agreement, however, that the underlying processes are correctly modeled. As noted above, we are concerned that the lack of fine-scale spatial resolution in the sediment transport model and the use of an overly simplistic PCB distribution and bioaccumulation model limits the ability of either model to accurately project future PCB levels in the Upper Hudson River, with or without active remediation. More sophisticated field evaluations and models would greatly improve the efforts to define and monitor the remediation of the Hudson River.

KEY QUESTION 3. What are the chances that the PCBs currently buried in the upper Hudson River will be released sometime in the future under extreme weather conditions? Findings:

1. The extent of remobilization of "buried" PCB-contaminated sediments during episodic high flow events (for example, 100-year or 200-year floods) may have been underestimated and remains a concern.

2. Based on current releases of PCBs from sediments and potential remobilization of "buried" PCBs during episodic events, we do not see monitored natural attenuation as a sufficient remedy.

As if modeling sediment and PCB movements in the dynamic Upper Hudson River was not difficult enough, the modeling of extreme weather events, such as a 100- or 200-year flood, is particularly challenging. Models are calibrated with available data, which typically do not include extreme events and often do not include flood periods. The spatial patterns of sediment erosion and deposition vary as functions of river flow. It is quite likely that an extreme event such as a 100-year storm will occur in the Upper Hudson River during the recovery period. Whereas a 100-year storm is an event that occurs, on average, every 100 years, there is a 10 percent probability of a 100-year storm occurring in the next 10 years, a 25 percent probability in the next 30 years, and a 40 percent probability within the next 50 years. A question central to the PCB issue in the Upper Hudson is the depth of remobilization of sediments under different flow conditions. More energy in the river in the form of water currents can cause a deeper disturbance of the sediments and a greater release of the associated PCBs to the water column. To assess the potential impact of high flow events, both GE and EPA modeled the bottom current velocities under a high flow of 47,000 cubic feet per second (QEA, 1999; U.S. Environmental ProtectionAgency, 2000d). The two models predict substantially different amounts of non-cohesive sediment remo-bilization in the Thompson Island Pool, with the

EPA model predicting as much as 13 cm eroded (averaged over the pool - see U.S. Environmental ProtectionAgency 2000d) versus 0.14 cm from the GE model (QEA, 1999). This important discrepancy underscores the difficulty in using hydrodynamic and sediment transport models to estimate sediment remobilization during extreme events in the Upper Hudson River.

We are also concerned that the flows used to model the impact of extreme events do not adequately account for high flows from the Sacandaga Reservoir, which drains to the Hudson upstream from the GE plants. Since the Sacandaga River was dammed in 1930, one storm (May 1983) was large enough to cause water to spill over the dam and raised flows in the Sacandaga River above 12,000 cubic feet per second3 which is 50 percent higher than the worst-case Sacandaga River flows used in the sediment transport modeling. In addition, the operation of the Sacandaga dam has recently changed. Relicensing agreements between Orion Power and surrounding communities on the Sacandaga Reservoir and along the downstream HudsonRiver dictate that Orion Power willkeep the reservoir at higher levels both during summer and winter months (Bucciferro, 2000). The new agreement signals a shift in management practices away from one favoring flood control, toward one favoring recreational uses of the reservoir and river. This loss inreservoir capacity decreases the dam's ability to hold back precipitation during extreme events, and increases the likelihood of flows through the Upper Hudson River that have not been experienced since the dam was constructed seventy years ago.

Neither the GE nor the EPA model adequately explains the observed current PCB releases from the Thompson Island Pool. We believe this is partly due to the coarse spatial and temporal resolution of those models and their corresponding inability to properly represent small-scale and ongoing redistribution of sediments within the pool. As we mentioned previously, both GE and EPA maintain that the Thompson Island Pool is net depositional, without any supporting geophysical evidence. The

3 River flow data have been recorded daily since 1907 at the

Sacandaga River at Stewarts Bridge near Hadley, NY (USGS

Station 01325000).

overall result of their modeling is that less than 20 percent of the total reservoir of PCBs in the Thompson Island Pool willbe released over the next thirty years without dredging, with the remainder buried indefinitely (QEA, 1999; U.S. Environmental Protection Agency, 2000b). Due to shortcomings of the modeling with respect to the ongoing redistribution of sediments under low to moderate flows and large-scale changes under extreme flood events, we believe the eventual release of PCBs from the Thompson Island Pool could be much greater than the 20 percent of the current PCB reservoir predicted by the models. We believe that both GE and EPA have likely underestimated the magnitude and probability of PCB release from the sediments and subsequent transport downstream.

KEY QUESTION 4. Can active remediation be implemented in such a way that it provides a net long-term benefit to the Hudson River? Findings:

1. In other locations, active remediation of contaminated sediments resulted in lower contaminant levels and risk in wildlife. While the most sensitive species will continue to be impacted for decades, other less sensitive species will benefit sooner from declining PCB levels resulting from active remediation.

2. With the best dredging techniques, onlyavery small fraction of PCBs are released to the water, likely less than 2 percent of the total PCBs dredged. In the Thompson Island Pool, this short-term releaseis comparable to the rate at which PCBs are currently being released from the sediments. Thus, with properly designed and executed techniques, dredging may result in no more than a doubling of the present day PCB flux during the project period.

3. Effectively managing the dredged materials stream is critical to the success of the active remediation.

4. Dredging with appropriate techniques is technically feasible, but requires rigorous oversight to minimize contaminant dispersion and community disruption.

5. There will be short-term impact of the dredging operations on local communities and habitats but, properly managed, these impacts need be no greater than those of other large construction activities (road/bridge construction, navigation dredging, lock and dam repair and maintenance).

The estimated average concentration of PCBs in surface sediment in Thompson Island Pool is approximately 40 ppm, with maximum concentrations reaching 2,000 ppm (U.S. Environmental Protection Agency, 2000a). Elsewhere, concentrations of this magnitude and less required or led to remedial actions under state and Federal laws. For example, sediment remediation in Commencement Bay near Tacoma, Washington is proposed to reduce the PCB level to 0.45 ppm, although the National Oceanic and Atmospheric Administration and the Department of the Interior, the Federal stewards of natural resources, have requested a lower target of 0.2 ppm in the interests of chinook salmon andfish-eating birds (Weiner, 1991; United States Fish and Wildlife Service and National Oceanic and Atmospheric Administration, 1999). Target PCB concentrations have been set at 1 ppm for cleanup of the Housatonic River (Connecticut - Massachusetts), the St. Lawrence and Raquette rivers (New York), the Kalamazoo River (Michigan), and the Delaware River (New Jersey - Pennsylvania); at 0.5 ppm for the Sheboygan River and harbor (Wisconsin); and at 0.25 ppm for the Fox River (Wisconsin).4

Although active PCB remediations have not always been successful due to design and operational problems (General Electric, 2000), there are other examples where biological benefits have followed active remediation. We note that whether a specific remediation is deemed 'successful' depends upon the criteria established for that project. Short-term degradation resulting from the dredging activity can mask eventual benefits, and one must recognize that judging the 'success' or 'failure' of a remediation will likely require a long-term view. Examples where biological benefits followed active remediation include Sweden's Lake

4 See the Major Contaminated Sediments Sites (MCSS) Database, a joint effort of the General Electric Co., Applied Environmental Management, Inc., and Blasland, Bouck & Lee, Inc. at index.html for details on specific contaminated sediment sites in North America.

Jarnsjon, where after two years PCB concentrations in one-year-old Eurasian perch in the lake and fifty miles downstream were half those before dredging (Bremle and Larsson, 1998a,b; Bremle, Okla, andLarsson, 1998). Removal ofPCB-contaminated soils and near shore sediments along the Upper Hudson River at Queensbury, New York (upstream from the two GE plants) led to significant declines of PCBs in yellow perch except near a remnant hot spot.5

Active remediation has relieved stress from other contaminants as well. Marsh and open-water sediments along the Lower Hudson River at Cold Spring were badly polluted with heavy metals (mainly cadmium but also cobalt and nickel). These sediments were excavated or dredged. The marsh area was covered with an absorptive clay fabric liner and clean fill and then replanted. Subsequently, five years of monitoring showed notable decreases of cadmium in the bodies of local plants, birds, invertebrates, and test fish.6

Similar findings have been reported for sediments contaminated with polycyclic aromatic hydrocarbons (PAHs), a class of synthetic organic compounds that are toxic to animals. Brown bullheads living over PAH-contaminated sediments in the Black River, Ohio, had high prevalences of liver tumors. Dredging of the sediments brought a temporary increase in tumors among resident bullheads, but bullheads spawned after the dredging had no tumors four years later (Baumann and Harshbarger, 1995,1998). The age class structure of the bullhead population improved, and the benefit of dredging was greater than that observed after onshore source control of the PAHs (Baumann, Blazer, and Harshbarger, 2001). Similarly, liver tumors in English sole at a PAH Superfund site in Eagle Harbor of Puget Sound, Washington, decreased fifteen-fold over the six years after the site was capped with cleaner sediments (Myers et al., 2001). Eleven years of monitoring before this remediation had shown

5 Ronald Sloan, New York Department of Environmental Conservation, personal communication. New York State Department of Environmental Conservation (2001b) Hudson River PCB Biota Database. Bureau of Habitat, Division of Fish, Wildlife and Marine Resources, Albany, NY, Updated 13 March 2001.

6 See html for a history of the Foundry Cove site at Cold Spring, NY.

no evidence of natural PAH attenuation in either the sediments or the fish (Baumann and Harshbarger, 1995,1998).

These case studies indicate that active remediation of contaminated sediments can more effectively reduce toxic pollution in most aquatic systems than natural dissipation of the pollutants. In addition to reducing surface contaminant concentrations, dredging will greatly reduce the reservoir of buried contaminants that could be remobi-lized during an extreme event. Lessening the risk of event-driven release of PCBs is one of the most valuable long-term benefits of dredging. Our professional opinion is that removal of contaminated sediments from the Upper Hudson River will accelerate recovery of the river.

Dredging will bring problems, of course. Some contaminants inevitably will be released when dredging disturbs the sediments. Previouslyburied muds with high PCB concentrations might be encountered and disturbed. Aquatic habitats will be disrupted in and downstream of dredging areas. Management of waste sediments will be a large and challenging operation. Nearby human communities will be bothered by noise, lights, odors, and temporary closures of roads and navigation channels. We believe these problems are less serious than commonly perceived and can be minimized.

Dredging technology has greatly improved in the past decade (National Research Council, 2001a). An ability to "surgically" dredge has developed in response to demand for such technology around the world, and firms specializing in remediation dredging (as distinct from navigation dredging) now exist. As with any engineering project, success or failure of Hudson River dredging will depend equally on the quality of the project design and the rigor and responsiveness of the project's oversight. Both factors can be encouraged and facilitated by performance-based contracting, but it will be very important to carefully specify the expected outcomes of dredging in terms of contaminant removal. Detailed site assessments will be neededbefore dredging begins to refine ourknowledge of the current spatial distributions of sediments and contaminants in PCB hot spots. The collection and analysis of high spatial resolution data detailing sediment and PCB distributions through the project area will allow managers to select the best removal technology (for example, hydraulic versus mechanical dredging), access points, and waste management procedures. Such information also is needed for accurately estimating overall project costs.

Any disturbance of contaminated sediments can release both particle-associated and dissolved PCBs. Operations must be designed to minimize these releases. In a well-documented study in the Fox River, Wisconsin (Steuer, 2000), the release of particulate and dissolved contaminants was 2 percent of the total weight of PCBs removed. No particular attempt was made to optimize PCB confinement in this project. We believe that substantial improvements can be realized and that ultimate losses will be less than 2 percent. However, even at the 2 percent loss level, the additional release and downstream transport of PCBs would amount to 180 kilograms per year under the proposed EPA alternative,7 an amount comparable to the current annualrelease from the Thompson Island Pool sediments. In this "worst case," the total amount of PCBs released into the Upper Hudson with dredging could be doubled, relative to not dredging, over the duration of the project period.

Dredging will temporarily destroy habitat in several ways besides changing the substrate: local flows will be altered and submerged aquatic vegetation and marginal wetlands will be lost. However, aquatic vegetation will readily recolonize disturbed areas from upriver sources once dredging is finished. Wetlands can be restored by established techniques with full consideration of the concerns raised recently by the National Research Council (2001b) regarding implementation, monitoring, and selection of success criteria. Fish undoubtedly will be driven from areas of dredging because of bottom disruptions, turbidity, and noise. The stress of displacement and of crowding on established populations elsewhere may increase fish mortality for a period of time. However, fish and aquatic invertebrates typically recolonize abandoned areas rapidly after disturbances have ended. Scheduling

7 This estimate of 180 kilograms of PCBs per year potentially released from the dredging operation was calculated as follows. Approximately 100,000 pounds of PCBs are to be dredged from the Upper Hudson River over a five-year period (equal to about 9,100 kg/year). If 2 percent of this is released, 180 kg/year (9,100 kg/year dredged times 2 percent released) could be released.

of operations to avoid known periods of spawning and migration will be important nonetheless.

Management of waste sediments can greatly disturb adjacent human communities if it is not carefully designed and implemented (National Research Council, 2001a). The plan as presented calls for the wastes to be ultimately transported to an out-of-state Hazardous Waste Landfill (U.S. Environmental Protection Agency 2000a), but operational aspects must be considered. These include the dewatering facility, waste transfer stations, and transport of waste from the dredging site to the processing site by pipeline, barge, or rail. The dewatering facility consists of a settling basin and a filter press to remove interstitial water from dredged sediments. Residual waters will be treated to remove PCBs and returned to the river. Dried sediments may be moved directly or barged to a transfer station for out-of-state rail transport and disposal. If these operations are sited and managed to minimize the number of times sediment is handled, community impacts will be lower than otherwise. Efforts to reduce these impacts will benefit from early and continuing consultation with community representatives.

No data indicate that dredging operations themselves will directly affect public health. Despite claims to the contrary,8 construction projects similar in magnitude to and larger than the proposed Hudson River dredging occur regularly in densely populated areas and are accommodated by the affected communities. Although the entire proposed dredging operation along the upper Hudson will take several years, particular communities will be affected for much shorter times. Economic impacts can be offset by care in planning and scheduling and, when unavoidable, financial compensation. Lighting and noise intrusions often can be reduced to below expectations. Continuous operations (night and day, seven days per week) are most efficient and therefore preferred from an

8 For two contrasting views on the efficacy and impact of environmental dredging, see General Electric (2000) Environmental Dredging: An Evaluation of its Effectiveness in Controlling Risk, General Electric Company Corporate Environmental Program, Albany, New York (http://, and Scenic Hudson (2000) Results of Environmental Cleanups Relevant to the Hudson River (http://www.

operational standpoint, but more accommodating schedules might be adopted in areas of high population density. Innovation and a willingness to compromise will be needed by all.

Was this article helpful?

0 0

Post a comment