a. “Passive” dewatering, i.e., allowing the sediment to dry. Although requiring little energy input, this would release very large amounts of PCBs to the air. If done in an enclosed and filtered space, it may be acceptable, but it is not preferred.

b. “Mechanical” dewatering, which uses presses, hydrocyclones, belts, or other mechanical means to extract water. These methods can lower water content dramatically in a short time, lowering overall volatilization. It should be noted that the water extracted from the sediment will be contaminated with PCBs and must be treated. Depending on processing time, it is possible that air in contact with sediment must also be captured and filtered.

c. “Active” dewatering; i.e., heating. Volatilization levels from active dewatering are expected to be as high or higher than passive dewatering. Clearwater considers active dewatering unacceptable, unless a process such as post-treated heated stripping is used.

Disposal options such as landfilling which require solid waste necessitate complete dewatering (and the addition of stabilizing or solidifying elements), while other treatment trains can accept partially dewatered sediment. The exposure of workers to wet contaminated sediment must be carefully addressed. Some dewatering technologies, notably belt filters and plate-and-frame presses, are particularly messy, scattering sediments and require frequent cleaning, thereby exposing workers to a potentially large amount of wet sediment. Other methods, like hydrocyclones, are cleaner or require less cleaning. The treatment method chosen drives the choice of dewatering: solidification, the method proposed by EPA, requires complete dewatering, making this a difficult process, almost certainly requiring the use of belt filters or plate-and-frame presses, and therefore exposing workers to contaminated sediments. By contrast, an alternative treatment train (for example, one involving soil washing, solvent extraction, and chemical dechlorination) might require substantially less dewatering. In this case, a single pass through a hydrocyclone might suffice to dewater the sediment, exposing workers to significantly less volatilization during dewatering.

Clearwater recommends that the treatment methods be considered, in part, by the effect they have on requirements for dewatering, and that treatments which require less worker exposure to sediment during dewatering be preferred.

The tendency of PCBs to volatilize from wet sediment along with water indicates that dewatering must be carefully controlled to prevent escape of PCBs; for further recommendations, see “Potential for Volatilization,” below.

In addition to filtering air, EPA must carefully treat and monitor water recovered from wet sediment before returning it to the river.

EPA’s plan calls for two dewatering plants, one in the area of River Section 1 or 2, and another in the Albany area to manage the sediments from River Section 3. Clearwater believes that only one water treatment facility is needed. Material from River Section 3 could be barged up to one location, rather than building a second dewatering facility in the Albany area. The dewatering facility can offer a win-win solution to both the host community and to the project if it is sited in exchange for full remediation of existing upland hazardous waste sites created by NYS DOT, which dredged the original sediment to maintain navigation. An additional rail siding may be needed to accommodate up to 45 boxcar loads of dewatered sediment per day.

Treatment Technologies for Dredged Sediment

The EPA Proposal

The EPA’s Preferred Remedy proposed dewatering and stabilizing dredged sediment, and then transferring it to a landfill for disposal. While landfilling is a relatively low-cost solution, it should be regarded as a long-term temporary solution. “Effective containment will not only require continuous maintenance, but will need to be maintained to ensure the engineered system continues to prevent off-site migration as well as guard against water intrusion.” (Scrudato et al. 1999) The possibility that PCBs could, at some point in the future, become remobilized and reenter the ecosystem is a potential shortcoming of landfilling.

A common public criticism of the EPA’s proposed plan is that it simply shifts the problem around, by moving PCBs from the river to another area of the country. This assessment fails to consider that the current state of the river approximates an unlined landfill with water constantly washing through it. Containment in a double-lined Part 360 landfill or a permitted hazardous waste landfill is far more protective than continuing to allow PCBs to be dispersed into the environment and bioaccumulate in the food chain.

Alternatives to landfilling that are able to decontaminate the large quantity of sediment that will be removed by dredging and then destroy the PCBs are preferable to landfill disposal. Several treatment options are available, but have been dismissed by EPA. In this section, we will consider applicable treatments.

In the initial screening of treatment options, EPA retained for further study a number of treatment technologies (EPA 2000):

• sediment washing
• solvent extraction
• chemical dechlorination
• thermal desorption
• thermal destruction (incineration)
• solidification/stabilization and landfill disposal

However, later in the same document, EPA eliminates the first four of these methods due solely to cost considerations. The final EPA proposed remediation plan proposes solidification/ stabilization and landfill disposal of all contaminated waste. Depending on PCB concentrations, the stabilized sediment will be brought to a Toxic Substances Control Act- or non-TSCA-regulated landfill for disposal.

Shortcomings of the Preferred Remedy

Unfortunately, stabilization and landfilling suffers some disadvantages. Foremost of these is the fact that the dredged PCBs are neither destroyed nor isolated from the contaminated sediment. The Comprehensive Environmental Response, Compensation, and Liability Act establishes offsite transport and disposal of untreated contaminated materials as the least desirable alternative (Renholds 1998). While landfilling removes PCBs from the ecosystem of the Hudson River, it stops short of being a complete solution. Landfilling also requires transporting a huge amount of sediment. Other treatment technologies can isolate PCBs from sediment, potentially lowering transportation costs and even, in some cases, allowing the cleaned sediment to be backfilled into the dredged areas. Finally, the dewatering and solidification steps required for landfilling (in particular, solidification with cement or lime) can increase volatilization of PCBs from the sediment, as has been discussed above.

Maximal protection of both human health and the ecosystem can best be accomplished by removing and destroying the PCBs when possible. Despite the increased cost, Clearwater strongly recommends that EPA reconsider the other treatment technologies discussed below.

Applicable Treatments

These treatment processes have been discussed elsewhere; for a complete analysis, see especially Hirschhorn 1994 (also EPA 2000; ATSDR 1993). Except for the first listed, all of the processes listed below take place ex situ.

• In situ Bioremediation. Both anaerobic and aerobic bacteria are known to degrade or dechlorinate PCBs in the laboratory, although this is more difficult to observe in contaminated waters. In 1991, General Electric tested bioremediation in six caissons sunk into the river bottom (Renholds 1998). To encourage mixing of the bacteria with contaminated sediment, each cell used either an impeller, resuspending the sediment, or an automatic rake to mix sediment more slowly. Oxygen, nitrogen, and phosphorus were pumped into the caissons. Some dechlorination was measured, but the results were not encouraging. Bioremediation appears to be most effective at high concentrations, reducing its applicability to in situ remediation projects. Expanding this type of process to the riverbed, with little or no control over temperature, oxygen content, nutrients, and bacteria levels, is not technically feasible, as even the Potentially Responsible Party concluded after years of extensive experimentation.

• Ex situ Bioremediation provides greater control over the bacterial degradation of PCBs. To increase efficiency, a sequence of anaerobic and aerobic bacteria are used in a confined reactor. Although some laboratory work has been done along these lines, there is very little field data available, and the technology is not well understood (ATSDR 1993).

• Soil washing involves separating out the finer-grained particles in sediments, to which organic contaminants like PCBs bind preferentially. This produces a smaller volume of more highly concentrated PCB-contaminated sludge; typical volume reduction range from 70–90%. It is ineffective in soils with a large fraction of fine grains. Surfactants and other agents may be added to desorb contaminants from the sediment grains. Sediment washing has been demonstrated at high volumes and can be considered a proven technology applicable to the Hudson River; in an EPA demonstration project at Saginaw Bay, washed sediment had a PCB concentration of only 0.2 ppm (Hirschhorn 1994). It is possible that washed sediment could be returned to the river, greatly reducing transportation costs. Little or no dewatering would be required prior to sediment washing.

• Solvent extraction. Solvents like propane, kerosene, or alcohols can be used to dissolve contaminants from sediment. The used solvent stream is then treated, resulting in a concentrated PCB waste, and the solvents can be reused. Multiple cycles are usually necessary for efficient extraction; the entire system is enclosed. This method has proven effective on a large scale. Unfortunately, reduction of PCB concentration to levels low enough for sediment to be returned to the river would appear to be relatively rare (Hirschhorn 1994). The use of large quantities of flammable solvents can pose safety issues.

• Chemical destruction/dechlorination: Alkali Metal Polyethylene Glycolate (APEG). There are several different reagents which can be combined with PCBs and heated, resulting in removal of chlorine atoms from PCBs. APEG can result in partial dechlorination. Chemical costs can be very high for large volumes. APEG has been demonstrated at large volumes and can be considered a proven technology.

• Chemical destruction/dechlorination: Base-Catalyzed Decomposition (BCD). Mixing and heating of contaminated waste with a base, a catalyst, and a hydrocarbon results in stripping of chlorines from biphenyl molecules. This reaction can proceed to non-detectable levels. Emissions must be controlled and chemical byproducts separated out. Chemical costs are lower than for APEG technologies. BCD has been tested at field scales and appears to be a proven technology. Both BCD and APEG can be used to treat sludge without complete dewatering.

• Thermal desorption requires heating of contaminated sludge to volatilize, but not destroy, contaminants, which are then recaptured. Efficient thermal desorption requires low water content. Dioxins can be formed if oxygen is present. One EPA demonstration of this technology reduced PCB concentrations to below 1 ppm without detectable production of dioxins or furans (ATSDR 1993). Proven.

• Incineration. Very high temperature incineration can completely destroy PCBs and other contaminants. However, dioxins and furans can be formed as byproducts of incomplete combustion, and extremely high efficiencies (above 99.9999%) are required. EPA has eliminated incineration due to permitting difficulties, public opposition, and prohibitively high cost for off-site treatment. Clearwater does not consider incineration to be a viable treatment option for PCBs, although the non-chlorinated byproducts of chemical dechlorination processes can be safely destroyed using this method.

Treatment Alternatives

The processes listed above can be combined for effective removal and destruction of PCBs from contaminated sediment or sludge. For example, soil washing, solvent extraction, or both might be used to concentrate PCBs, which can then be destroyed with BCD reactions; combinations of BCD with thermal desorption have been demonstrated at large scales. This type of treatment train has the tremendous advantage of completely destroying PCBs, which is greatly preferable to landfilling. With efficient processing, it may be possible to return treated soil to the river, reducing transportation costs as well as the need to purchase clean backfill.

Unfortunately, costs could be significantly higher than landfilling. EPA has estimated their costs as ranging from $100 to $250 per ton. The cost of stabilization and landfilling is about $145/ton for TSCA disposal, and $80/ton for non-TSCA disposal; the overall weighted cost of the Preferred Remedy is about $108/ton.

Cost of alternative treatments could be lowered by using treatment trains that concentrate PCBs, allowing a smaller volume to be destroyed. Sediment washing or solvent extraction, in particular, lend themselves to lowering volume and reducing costs; several sediment washing processes are available at full scale at a cost in the range of $75 per ton (EPA 2000, Table 4-9). If these processes were used to reduce sediment volume by 90%, the resulting highly concentrated sludge could be treated with a more expensive destruction process (such as a thermal desorption/BCD combination) with a per-ton cost as high as $300 and still have an overall cost approximating that of stabilization/landfilling. While sediment washing is not always applicable, the general process of concentrating and then destroying PCBs appears to be a reasonable solution for even very large volumes of sediment.

Finally, it should be noted that any technology which involves heating PCBs, particularly thermal desorption, must be carefully examined for dioxin formation.

Clearwater strongly recommends that EPA reopen the subject of treatment technologies, and that processes which completely destroy or dechlorinate PCBs be given priority consideration as the most complete solution to the problem.

Feasibility Study Modeling Considerations

The HUDTOX model, while representing a best estimate of future levels of PCBs in the sediments and water of the Upper Hudson, also overestimates the benefits of the No Action (NA) and Monitored Natural Attenuation (MNA) remedial alternatives. This could result in a more favorable comparison of NA or MNA when compared to active remedies than is warranted, when in fact, NA or MNA may not yield acceptable levels within an appropriate time frame (US EPA 2000, Appendix D). The consequences of limitations in the models are many, including an overly optimistic depiction of the rate of “natural recovery” for the Hudson River.

Specifically, the model overestimates the rate of recent declines in surface layer sediment PCB concentrations (in the model it declines too fast) and doesn’t include PCBs from depths greater than 5 cm. (e.g. up to 10 cm) that are mobilized into the food chain by benthic organisms. It appears that a leading factor in this exclusion is the fact that GE-provided surficial sediment data didn’t extend below 5 cm, although it is clear that PCB contaminated sediments do (US EPA 2000, Appendix D).

HUDTOX predicts a reduction of PCB flow over the Troy dam by 42% under the Preferred Alternative when compared with Monitored Natural Attenuation (EPA 2000b). If the model does overestimate river recovery, as seems likely, future PCB levels under MNA would be higher than predicted. This means that true reductions in PCB flow for active remediation compared with MNA will probably be much higher than predictions have estimated. In general, this tendency of the model to overestimate natural recovery understates the need for dredging.

The fact that both the HUDTOX and FISHRAND models optimistically depict the benefits of the No Action and Monitored Natural Attenuation alternatives in favor of the Potentially Responsible Party (PRP) for this site, and that this is a direct consequence of a lack of proper calibration data due to limitations in data generated by the PRP, appears to be a serious conflict of interest. Certainly, the PRP has benefited from such a modeling bias in terms of public perception and awareness of the relative merits of each remedial alternative, especially MNA and No Action.

In addition, the model’s emphasis of surface area over total PCB mass is evidently responsible for EPA’s decision to propose more stringent remediation in River Section 1 than in River Sections 2 and 3. In fact, this represents a bias towards less cost-effective remedies (in terms of PCB mass removed), since the model effectively does not treat masses of PCBs left buried under only 5 cm of sediment.

While Clearwater values the output of the HUDTOX model, and understands the importance of surface area in resuspending PCBs from river sediments, we believe that it is prudent to remove as much PCB mass as possible. As the above analysis has show, the highest efficiency for removal exists in River Section 2 and, to a lesser extent, in River Section 3. Dredging of River Section 1 is important to prevent resuspension of PCBs from the dilute sediment there; but dredging of River Sections 2 and 3 is critical to remove the largest possible mass of PCBs.

Additional treatment of the limitations of HUDTOX may be found in Scenic Hudson’s public comment, filed under separate cover.

Analysis of Remedial Alternatives

Specification of a remedial plan requires a delicate balance of public health benefits and concerns, wildlife habitat benefits and destruction, cost, and technical feasibility. For purposes of planning, EPA has divided the Upper Hudson into three sections. River Section 1 extends from the site of the Fort Edward Dam to the Thompson Island Dam, encompassing the Thompson Island Pool; Section 2 extends to the Northumberland Dam; and Section 3 ends at the Federal Dam at Troy.

An analysis of dredging requirements and benefits by river section and by remediation criterion is instructive; we provide this matrix, essentially a restatement of Feasibility Study Tables 3-4, 6-2, and 6-3, in Table 2. Table 3 shows a summary of this data for several permutations of remediation alternatives, and compares them with the Preferred Remedy (approximated in this comparison by 3/10/10, which removes slightly more sediment than 3/10/Select).

The first conclusion we draw from this data is the dramatic benefit associated with dredging River Section 2. Full-section dredging of this section would recover approximately 35,000 kg of PCBs (a lower bound), as compared with about 15,000 kg in River Section 1, and 11,000 kg in River Section 3 (at 3 g/m², the most rigorous remediation plan evaluated by EPA for that section). In terms of mass of PCBs recovered, River Section 2 appears most productive, but more can be recovered from Section 3 without significant reduction in recovery efficiency.

Although EPA has proposed the most rigorous dredging in River Section 1, it appears from this data that both River Section 2 and River Section 3 are more productive in terms of PCB mass removed. Table 2 presents a rough efficiency measure of removal in terms of kilograms of PCBs removed per thousand cubic yards of sediment dredged. (Since cost scales with cubic yards removed, this is also related to a measure of cost efficiency). Dredging in River Section 1 is the least efficient, ranging from about 7 to 9 kg/1000 cy, which RS 2 is extremely efficient at 3 and 10 g/m² and slightly less so at 0 g/m². Although EPA has not presented data for River Section to at full-scale remediation (0 g/m²), the relatively high efficiency of this section at 3 g/m² (19 kg of PCB removed per thousand cubic yards of sediment, nearly triple the efficiency of RS1) indicates that expanded remediation in RS3 is likely to be cost-effective, despite the large area. Clearwater recommends that EPA provide data on more rigorous remediation of River Section 3. At a minimum, EPA should consider expanded dredging of hot spots identified at the 3 g/m² level.

This analysis demonstrates the relative cost inefficiency of the Preferred Remedy, partly as a result of the relatively low emphasis on River Section 2. For example, the Preferred Remedy would remove 41,900 kg of PCB at cost of approximately $10,800 per kg. More rigorous dredging of River Sections 2 and 3, for example, would remove an additional 15,400 kg of PCBs (possibly more, since the EPA estimate for RS2 is a lower bound), at an overall cost per kg of about $8,500—a decrease of more than 20% (alternative 3/0/3). (It should be noted that, since complete costs are only calculated for two alternatives by EPA, we have estimated all other costs with a simple interpolation, assuming an overhead cost and an additional cost per cubic yard of sediment dredged and treated. These extremely rough numbers nevertheless allow us to make reasonable estimates of cost efficiency for purposes of comparison.)

Given the high efficiency of dredging in River Sections 2 and 3, we question the limited treatment of the Preferred Remedy in these areas and the relatively inefficient focus on River Section 1. EPA has based selection of its Proposed Remedy largely on data from the HUDTOX model of PCB distribution and movement through the ecosystem. As has been discussed above (“Feasibility Study Modeling and Fisheries”), HUDTOX fails to fully address PCB mass buried below 5 cm, biasing the results towards remediation of areas which, like River Section 1, have less mass but more surface concentration. In fact, this creates a bias towards less cost-effective remedies in terms of PCB mass removed.

Clearwater supports dredging of River Section 1 at either the 3 g/m2 level proposed by the Preferred Remedy, or at a more rigorous level. In addition, Clearwater strongly supports increased dredging in River Section 2 and River Section 3 to remove the largest mass of PCBs possible. Given the increase cost efficiency of expanded dredging in these sections, increased dredging in these areas will greatly increase the overall effectiveness of the project. Although we cannot quantitatively estimate either the reduction in PCB flow into the Lower Hudson or the reduction in fish PCB concentrations without a model, it is clear that a more stringent remediation than the Preferred Remedy would lower both values faster than predicted.

In summary, the Preferred Remedy is neither comprehensive enough, nor cost-effective enough. It has been made clear, from the foregoing analysis, that EPA’s 3/10/Select remedy is too timorous both in terms of PCB mass removed and cost-effectiveness. Clearwater strongly urges EPA to adopt a 3+/0/3+ remedial standard.

Clearwater believes that 3+/0/3+ best meets the tests of cost-effectiveness and human health protection. In support of this statement, we offer the following:

a) A 3/0/3 option provides the best return on investment, measured by cost per kilogram of PCBs, and may be as much as $2,000 per kilogram less expensive than the 3/10/Select option (Table 2).

b) 3+/0/3+ removes almost as much PCB mass as 0/0/3, and more than 3/0/3, more cost-effectively, while reducing the amount of material that must be dredged by almost 500,000 cubic yards.

c) 3+/0/3+ simply involves extending the removal operations beyond a strict perimeter defined by a 3 ppm contamination level. Three-plus is a pragmatic standard that will remove more PCBs, ease the navigational burden on the dredge operators, and as a result may save money.

d) The most recent NYS DEC fish data indicate that fish at river mile 168 (Stillwater) have lipid burdens of PCBs that average eight times greater than those of fish at river mile 11 (near the George Washington Bridge). It follows that the EPA remediation hypotheses for time to safe consumption of fish at weekly or monthly intervals may be reduced significantly (if not directly by a factor of eight) for fish caught in the tidal estuary. Hence, where monthly fish meals may be safe 26 years post-remediation at Stillwater in the 0/0/3 scenario, thousands of subsistence anglers in the estuarine Hudson may be able to safely eat their catch weekly or better within a few years after remediation at the recommended 3+/0/3+ standard.

Clearwater believes strongly that these elements present a compelling argument in favor of more-stringent remediation, and we believe that in addition to protecting public health, 3+/0/3+ is the most cost-effective scenario.

Volatilization

The evidence for volatilization indicates that this is currently a route by which a substantial mass of PCBs is entering the global ecosystem from the Hudson River. It is also a small but involuntary route of exposure for all residents of the Hudson Valley.

While the Proposed Remedy will lower the volatilization rate, it is important to note that an incautious remediation project may itself cause substantial volatilization during dredging. Clearwater believes that volatilization must be controlled to ensure that inhalation does not become an important pathway during the remediation process. Here we will try to estimate the amount of PCBs that could be volatilized during remediation, to calculate the potential hazards of this effect, and to make suggestions for the design of the remediation to mitigate the volatilization.

In doing so, we address several specific concerns. The first is that uncontrolled volatilization could increase airborne PCB levels enough to pose a hazard to worker safety, and even, in extreme cases, to neighboring communities. Secondly, even if PCB levels in air remain low, volatilization represents a pathway by which a large amount of PCBs reenter the ecosystem, contaminating otherwise pristine areas. It is the mandate of EPA to remove toxins from the ecosystem wherever possible, and Clearwater strongly believes that PCBs cannot be permitted to escape during the remediation process.

A number of papers by Scrudato, Chiarenzelli et al. have studied the effects of volatilization of small amounts of PCBs in a laboratory setting (Chiarenzelli et al. 1997; Chiarenzelli et al. 1997b; Scrudato et al. 1999). From 20% to 65% of the total PCB mass of a small sample was lost to volatilization over the time it took the samples to dry. Wetter sediment volatilized more PCBs, presumably due to the dipole-dipole effect of water with individual PCB molecules. While this shows the potential importance of volatilization, it does not provide us with a specific yardstick for estimating PCB release. The tiny samples (a fraction of a gram) used in the studies dry much faster than our large barge loads of contaminated sediment; in addition, the very large surface-area-to-mass ratio of the small samples means that a huge fraction of PCBs exist at or near the air-sediment boundary, where most volatilization takes place. The large volumes involved in the remediation project have a very low surface-area-to-mass ratio, exposing a small fraction of PCBs to the air, and will dry very slowly. In addition, a layer of water over the PCB-contaminated sediment will help slow this process, since the dipole-dipole interaction is strongest with a single-molecule-thick layer of water. It should be noted, however, that even one percent volatilization over the lifetime of the project would result in the release of approximately 500 kg of PCBs into the environment — an unacceptably high level.

Background levels of PCBs in outdoor air have been measured by a number of studies, usually averaging below one ng/m³, with levels sometimes higher in industrial cities. The indoor level is almost always higher, ranging from a few ng/m³ to a few hundred ng/m³ in some cases. One study found a mean of 9 ng/m³ in Birmingham (UK) (Currado and Harrad 1998); Vorhees et al. (1997) measured a range of 5.2 – 51 ng/m³, with a mean of 10 ng/m³, in their “reference” homes.

Several studies have measured increased PCB levels in air near contaminated sites. At a small lake in Sweden, which was remediated and the sediment landfilled, PCB levels were measured at a number of locations outdoors; the highest levels, with a mean of 5.9 ng/m³, were found in the area where the sediment was dewatered and landfilled (Bremle and Larsson 1998). PCB levels were found to correlate with ambient temperature as well as recent dredging activity. Measurements of PCB levels taken near New Bedford Harbor during dredging found concentrations from 0.4–53 ng/m³ (outdoor; mean, 4.94.6 ng/m³) and 7.9–61 ng/m³ (indoor; mean 181.8 ng/m³) in areas nearest the harbor (Vorhees et al. 1997). An ATSDR measurement of indoor air levels in Pittsfield, Mass. found concentrations of PCB of 21 ng/m³ (in the basement) and 7.24 ng/m³ (in the living room) (ATSDR 1999).

In most cases, indoor air levels appear to be driven largely by indoor sources (for example, old fluorescent light ballasts and electrical appliances) (Currado and Harrad 1998). Even the significant problems with volatilization experienced in New Bedford Harbor, which caused an eightfold increase in outdoor PCB concentrations, caused less than twofold increase in indoor levels.

Finally, PCB levels in produce showed some increase during dredging of the New Bedford Harbor. Tomatoes grown during dredging, at a site just 500 yards from the hot spot, showed a 6- to 8-fold increase in total PCB concentrations; tomatoes from a site five miles away showed a 2- to 3-fold increase. The increased concentrations of lower-chlorinated congeners supports the idea that these PCBs came through the air. Material dredged at this site was stored in a Confined Disposal Facility (a lined and covered holding pond), and covered with black plastic sheeting until a treatment option was agreed upon several years later. The large surface area of the CDF and heating of the black plastic covering both contributed to the volatilization of PCBs from the holding pond (David Dickerson, EPA, personal communication). Clearly, long-term storage of contaminated sediment can greatly increase risks from volatilization.

Health Effects of Volatilization During Remediation

We have seen (see “Health Effects,” above) that inhalation may be a substantial contribution to PCB body burdens, although not at dangerous levels. Should we be concerned about inhalation as a route of exposure during remediation?

Clearly, workers who are exposed to high levels of PCBs in enclosed work areas for several hours per day are at the greatest risk. Clearwater strongly recommends that all enclosed work areas be kept under negative pressure, drawing air away from work areas, and the air filtered.

Is there a risk to the public during remediation? The hot spot dredging project in New Bedford Harbor encountered some problems with volatilization, particularly in the area of the Confined Disposal Facility. These were apparently due to the large area of the holding pond, aggravated by the warming effect of the black plastic used to cover the sediment. Air levels in the immediate vicinity of the dredging site had concentrations elevated about eightfold (Vorhees et al. 1997). However, since indoor air concentrations are not strongly correlated with outdoor levels, indoor levels in this area experienced only about an 80% increase in PCB concentration, to about 18 ng/m³. While this elevation is significant, it falls well within the range of background PCB levels measured by Currado and Harrad, and is well below mean indoor concentrations reported by other studies (mean values ranging from 100 to over 450 ng/m³ in homes and offices, reported in Currado and Harrad 1998). Since the mean PCB intake through inhalation was found to be approximately one-tenth of the ATSDR’s recommended chronic MRL, this increase does not represent a health risk even to those residents living nearest the harbor. In addition, because of the lowered PCB water concentrations after dredging, lower volatilization will occur in these areas.

Although not posing a health risk, the experience in New Bedford Harbor does emphasize the importance of containing volatilization during remediation. Clearwater strongly recommends that volatilization be controlled with the goal of no statistically significant elevation of airborne PCB concentration during remediation and treatment.

Stages of Remediation

Any stage during which wet contaminated sediment is exposed to air must be looked at as a potential source of volatilization. Contaminated sediment on tidal flats or washed up on shore has been shown to be a significant source of PCB input to the ecosystem, with particularly high levels in predators living near the river (DEC 2001).

However, volatilization can be decreased by simply closing any system, allowing the air and sediment to come into an equilibrium in which the vast majority of PCBs resides in the sediment. It is the disturbance of this equilibrium by a stream of fresh air (or fresh water) that allows continuous volatilization (or dissolution) of PCBs to occur. Therefore, a closed barge is greatly preferable to an open barge, which allows wind and diffusion to continually carry PCBs away from the sediment; similarly, any closed system will limit volatilization, although large pockets of air might contain significant amounts of PCBs when equilibrium has been reached.

During sediment treatment, care should be taken to limit contact of the wet sediment with outside air as much as possible. This will probably require that processing occur within enclosures in which a negative pressure is applied. An air pump located as far as possible from the entrance and the work areas would ensure that airborne PCBs are kept from workers. Depending on the amount of volatilization and the effectiveness of the filtration, workers may not require personal protection equipment for short exposures, although levels should be carefully monitored at all times. The air pumped in this way should be filtered before being returned to the atmosphere. Enclosure of dewatering and treatment/solidification stages will dramatically lower the levels of PCBs released to the ecosystem.

Transport and dewatering of contaminated sediment carries the highest risk of volatile PCB loss. Mechanical dredging carries increased risk because of the exposure of sediment to air in the clamshell, while hydraulic systems which pump slurry through a pipe have the advantage of keeping sediment enclosed. Open barges would allow continuous volatilization.

Dewatering of sediment can be a difficult and messy process, and probably provides more potential for volatilization than any other step of the remediation process. Volatilization must be controlled carefully during dewatering. Enclosing the dewatering and treatment facilities, and pumping and filtering air from the enclosure, is the only way to ensure that volatilized PCBs do not escape into the local ecosystem. Because volatilization happens most effectively with wet sediment, sediment in which water content has been greatly reduced will volatilize fewer PCBs, although it must still be contained.

As has been noted above (“Sediment Dewatering”), different treatment trains require different levels of dewatering, and consequently expose workers to different amounts of contaminated sediment and air. Clearwater recommends that the treatment methods be chosen, in part, for the effect they have on requirements for dewatering, and that treatments which require less worker exposure to sediment during dewatering be preferred. The filtration and monitoring of air in the work enclosures will keep worker exposure to volatilized PCBs at a minimum; however, a treatment and dewatering process which requires less worker intervention (whether for cleaning, maintenance, etc.) and less worker contact with contaminated sediment is to be greatly preferred. Treatment trains involving extraction and destruction of PCBs, allowing dewatering to be accomplished with hydrocyclones, appear to be the best choices in terms of worker exposure as well as final destruction of PCBs.

EPA has chosen solidification and landfilling as its preference for disposal of PCBs. Clearwater has elaborated (“Treatment Technologies”) the reasons why we consider this to be an insufficiently protective treatment. Solidification carries some additional risks in terms of volatilization, as well. The addition of lime or other cementaceous processes has been shown to encourage massive volatilization (ATSDR 1993; Constant 1995). The need for complete dewatering, probably requiring several cycles of various processes including belt filters, adds risk of volatilization. Finally, although PCBs in solidified sediment are relatively stable, there exists the potential that they can be rewet and escape back into the ecosystem.

Recommendations for Controlling Volatilization

The available evidence indicates that volatilization is currently a major input of PCBs to the ecosystem, but will be lowered as the waterborne PCB levels are reduced by remediation. Volatilization may also be a concern during remediation, and must be carefully managed for the protection of worker safety in the immediate area of dredging and treatment, for the protection of the public from potentially increased levels of airborne PCBs, for the prevention of potentially increased levels of PCBs in local agricultural products and for the prevention of any substantial fraction of PCBs reentering the ecosystem.

Fortunately, a few steps should suffice to ensure that volatilization is kept to an absolute minimum. To this end, Clearwater makes the following recommendations.

• The design phase of the Hudson River remediation project must consider volatilization as a potentially significant byproduct of dredging. Volatilization must be minimized wherever possible.
• All efforts must be made to assure worker safety; in particular, the air in any enclosed areas on site must be tested and filtered, and workers must be provided with appropriate personal protective equipment.
• We recommend that hydraulic dredging be used wherever possible, reducing contact of contaminated sediment with both air and water.
• If mechanical dredging is used, all efforts must be made to restrict contact of the wet sediment with air. Barges must be closed. In cases of long travel times, air should be pumped out and filtered before off-loading.
• We recommend that the dewatering and solidification facilities be built to handle the volume expected from dredging, reducing the need for storage.
• In general, we recommend against storing material. Any stored material should be kept in lined and covered holding ponds and all measures taken to control volatilization. The problems with volatilization from covered holding ponds in New Bedford reinforce the difficulty of storing contaminated sediment.
• Dewatering, solidification, and treatment processes which are liable to release additional PCBs must be employed in enclosed spaces. The application of negative pressure and filtration of the pumped air will ensure that as little PCB is returned to the atmosphere as possible, while maintaining low levels inside for the health of the workers.
• Should it become unfeasible to treat sediment in the port, sealed modular shipping containers, transportable by truck and rail, would lend flexibility to the project.
• We recommend that a series of stations be established to monitor PCB levels in air in a variety of locations in, near and remote from the dredging projects. Total releases of PCBs should be estimated throughout the project.
• EPA must have processes to recover and filter or treat water and air from all storage containers and facilities.

Finally, Clearwater recommends that EPA establish the lowest possible target level for overall volatilization during the lifetime of the project. Even a target level of 0.1% would still represent 50 kg of PCB released to the environment.

Public Participation

EPA is to be commended for holding a series of eleven public meetings from Queensbury to New York City during the public comment period, beginning December 12, 2000 and ending April 17, 2001, at which they provided information and received verbal and written public comment. In addition they actively solicited public comment, including setting up an email site to receive electronic comment, assuring the public that its input would all be considered in their final Record of Decision. Most of EPA’s public meetings were very well attended, with several standing-room only crowds; however, this outreach only included eleven evenings in four months and occurred in ten locations.

Much of the public education on this complex subject was left mainly to environmental groups who went out into the communities up and down the Hudson River providing context and explaining both the intent and content of EPA’s proposed plan, distinguishing environmental dredging from dredging used for navigation or construction, clarifying that sediment is continuously resuspended by natural movement of the river, describing the methods that are proposed to protect drinking water and minimize remobilization of sediments, reporting health effects and environmental impacts, explaining bioaccumulation, defining potential routes of exposure, and so forth. Clearwater staff alone made presentations, sometimes more than once, to more than 50 municipalities, and dozens of civic, community and religious organizations, and schools. Other representatives from the Friends of a Clean Hudson coalition also did public outreach, including a major educational campaign by Scenic Hudson in the Upper Hudson area, and contributions from Riverkeeper, Sierra Club, Arbor Hill Environmental Justice Corp. and others. Clearwater’s award-winning video, The Hudson River PCB Story: A Toxic Heritage, was shown at dozens of schools, universities, municipalities, civic groups and community meetings. Clearwater mailed out more than 500 information packets, many including videotapes and the 30 page summary of EPA’s Proposed Plan and a variety of fact sheets (see samples in Appendix C: Clearwater’s Public Education and Outreach Materials.) Additional information was also posted on websites of EPA, Clearwater, Scenic Hudson, Friends of a Clean Hudson and various other organizations.

As a result of this public education, since the release of EPA’s proposed plan for public review and comment on December 12, 2000 65 villages and towns, including Albany and New York City, and three counties (Rockland, Putnam and Westchester) have passed municipal resolutions or written letters in support of EPA’s plan to actively remove PCBs from targeted hot spots in the upper Hudson River. In addition 173 organizations have also passed organizational resolutions. See Appendix D-1 and D-2.

As representatives from Clearwater and other organizations went from town to town educating municipal officials and the public on the problem and seeking support for remediation, several municipalities and many individuals expressed concern that the public would have little or no input into the decision-making process after the comment period closed on the extended date of April 17, 2001. One municipality described unresponsiveness of EPA to their community’s concerns during the remediation phase of an earlier project. Several have addressed the need for effective, ongoing public participation.

In addition the National Academy of Sciences report, A Risk Management Strategy for PCB-Containing Sediments 2001, by the National Research Council (NRC) recorded similar observations. From their interviews NRC noted that “participants made it clear that regulatory agencies have to communicate with the public during all phases of the decision-making process,” beyond what is required by Superfund regulations during the public comment period. (NRC 2000).

During the public process EPA has been consistently fair and forthcoming. Clearwater has seen no evidence of EPA withholding information or showing favor to anyone, with the possible exception of extending the courtesy to municipal officials to speak first and giving them unlimited time, while limiting the public’s comment time to two minutes per speaker. Much of the concerns about EPA’s relationship with the public expressed in the NRC report have not been observed during the period of public comment. However Clearwater does agree with one important point in the NRC “Lessons Learned About Superfund Community Involvement.” Clearly “the minimal approach required by CERCLA and NEPA” is insufficient to assure mutual trust and cooperation, or to obtain the best possible outcome.

In describing affected parties, the NRC report includes “individuals that live by or use the waterway, tribal groups, subsistence or sport fishers..., and commercial businesses who rely on fishing, recreational boating or tourism for ...economic survival.” It stresses that because sediments are mobile, “communities downstream may also be affected by contamination occurring upriver. For example, migratory fish contaminated by PCBs at one site may be eaten by people living hundreds of miles away.” NRC also points out that interests vary considerably among affected parties. (NRC 2001)

NRC distinguishes stakeholders as representatives of organized constituencies or organizations such as labor, environmental groups, government regulators, local government, responsible parties, consumer rights and advocacy groups, education and research institutions, religious organizations and others. NRC raises the environmental justice concern that stakeholders with the most powerful vested interests may dominate the process over less powerful members of the community, and that “sometimes people who are most affected by the contamination ... have the weakest voice in the remedy, and the least influence in the decision-making process.” (NRC 2000)

Among the benefits of inclusive community involvement NRC lists promotion of a democratic process, increased credibility and legitimacy, empowerment of affected communities and inclusion of “essential community-based knowledge, information and insight that is often lacking in expert-driven risk processes” (NRC 2000). All this information will inevitably result in a improved process and an better decision. Further, the degree to which the process encourages participants to collaborate, rather than compete, will further enhance the solution.

Recommendation:

Given the degree to which GE’s massive public relations campaign has confounded the process and confused the public, it is absolutely essential that the remedial design phase, the actual remediation process and follow-up monitoring be open to thorough public scrutiny.

EPA’s final ROD needs to specify a mechanism for ongoing public input during the remedial design phase and throughout remediation, as well as follow-up monitoring. This should include the establishment of mutually agreed upon milestones and the creation of an advisory committee that has active input into the entire remediation process.

GE and other responsible parties should be allowed to be included in the public process if they are cooperating with the ROD, but should not participate if actively creating obstacles to ROD remediation.

1) Hudson River PCB Remediation Advisory Committee: This oversight group would include at a minimum representatives of all major stakeholders:

• Local Municipalities (one per county bordering the Hudson River), plus one from each municipalities that take drinking water from Hudson
• Researchers and health experts
• Environmental experts
• Civic groups
• Local conservation groups
• Labor groups
• Dredging and decontamination industries
• Economists
• Sports and commercial fisheries
• Regulators

As appropriate, the Advisory Committee should form into specific working groups to focus on various aspects of the project.

2) Satellite Offices: Because of the size and geographic range of the project, satellite offices should be established in Fort Edward, Albany and New Paltz. These satellite sites should be staffed with a community outreach expert, and funds for user friendly education materials, videos and displays. They should be full repositories of all project information. Libraries, which are presently the repositories, are often inadequately staffed and provide little of no assistance to users. Citizens are entitled to have the opportunity to interact with a knowledgeable person on site.

3) Community Grants:

Technical Assistance Grants (TAG): As indicated by NRC/NAS, there should be multiple TAG grants proportional to size and cost of project. For example, under current Superfund regulations the same $50,000 TAG was awarded to citizens in the small community of High Falls (Town of Rochester) which has approximately 17 acres of TCE-contaminated groundwater estimated to cost $30 million to remediate as was allocated to the entire 200 mile Hudson River site, with a 40 miles remediation at an estimated cost of approximately $500 million.

Public Education Grants: As the administrator of the Superfund project, EPA’s education role has been limited to holding a handful of informational sessions at public comment meetings and providing literature on the project. To protect the public interest, public education grants are needed to assure that citizens have access to factual information. This is needed to counterbalance the vast resources the PRP is able to put to advertising. In addition, the PRP should be held to a rigorous standard of truth in advertising which has been consistently lacking in GE’s massive advertising campaign.

Environmental Justice Grants: One environmental justice aspect of this case is the fact that in spite of health advisories fish consumption by recreational/subsistence anglers continues. As noted, this is of particular concern for women, ethnic minorities and low-income groups in the Lower Hudson (Golden et al. 2000). Funds to address these specific concerns should also be made available. Closely related to environmental justice concern is the need to increase outreach and education to the general angler population on the dangers of PCBs in fish.

4) Schedule of Ongoing Public Information Meetings:

After the ROD is released, a series of four public information meetings should be held in the Hudson Falls/Ft. Edward area, Albany, the Mid-Hudson and in New York City.

EPA should open the process for creating a scope of work for the remedial design phase to the public.

Bearing in mind that the Hudson River belongs to everyone, the outcome of the remediation process is of great importance to all the people in the Hudson Valley. During the Remedial Design phase the HR PCB Advisory Committee and the public should have active input into:

• Technology Selection
• Request for Proposal (RFP) Development
• Review of Draft RFP
• Contractor Selection

Advisory committee should have active input during the development process at each of these phases and public meetings should be held when recommendations have been drafted, but before final determinations are made, to assure thorough public participation.

During remediation, EPA should respond promptly and fully to any and all concerns from any sources about contractor performance or any part of the remediation process.

After remediation has been completed, post-monitoring data and follow up information should be well publicized, with details available at designated repositories for a minimum of 30 years.