Tuesday, April 20, 2010

FIRESIDE CHAT: Comments To Hudson River PCB Dredging Peer Review Panel

Comments To Hudson River PCB Dredging

Peer Review Panel

Robert A. Michaels; PhD, CEP


RAM TRAC Corporation

3100 Rosendale Road

Schenectady, NY 12309-1510



(518) 785-0976

Uriel M. Oko; PhD, PE


Corrosion Services Corporation

11 York Road

Glenmont, NY 12077-3241



(518) 439-7880

19 April 2010


The manner in which Hudson River dredging Phase 2 will be conducted minimally must assure net benefit compared with no-action/no-dredging. Phase 1 contravened this benefit criterion approved by EPA and adopted by GE in Phase 1, suggesting the need to modify dredging, modify the benefit criterion, or modify both in Phase 2. Dredging must be modified to reduce PCB-tainted sediment resuspension from a Phase-1 rate between one and two orders of magnitude greater than anticipated. The benefit criterion itself must be modified to account for the ignored fact that the potential harm of dredging to the river, river ecosystems, and river communities was measured by resuspension alone, but resuspension constitutes just a miniscule fraction of the potential burden of dredging-induced PCB mobilization. Potential airborne risks also must be considered, given failure of EPA to quantify them in a baseline health risk assessment that provided support for the Hudson River dredging remedy. Finally, Phase 1 generated inadequate data on airborne PCB at dredge platforms and shorelines, waterborne PCB near dredge platforms, liquid PCB oils, and PCB in biota. Phase 2 therefore must expand environmental monitoring to generate adequate data to the public domain that can warn of harm in real time, and enable evaluation of the project over decades. We recommend that the Panel consider for Phase 2 such technologies as coffer damming, hydraulic dredging, and dredging within enclosures to minimize PCB mobilization, resuspension, and PCB entry into ecosystems and the air.


Our premise for commenting on Phase 1 implementation of the Hudson River dredging project is informed by our published, peer-reviewed findings that the project was justified by EPA based upon numerous errors, all in the dredging friendly direction, and our conclusion that these consistent errors amounted to pro-dredging bias on the part of EPA (Michaels and Oko, 2007; 5). We evaluated EPA’s baseline Hudson River health risk assessment (HRA), and found it to be biased toward keeping PCBs in sediments that would be removed by dredging. The HRA did this by systematically misquantifying parameters, most essentially underestimating PCB movement from sediments to water and from water to air. EPA excluded from its analysis all mono- and dichlorinated PCB congeners, which EPA subsequently estimated at one-third of total PCB mass in the river, and also excluded dissolved and colloidal PCB, as well as PCB adsorbed to fine particles, such as clay and silt. GE Site Evaluation and Remediation Program Manager John G. Haggard on 9 April orally reported that half of PCB in the dredging project was found to be dissolved; half was found to exist in particulate form (which would include colloidal PCB).

EPA included silt-adsorbed PCB, but overestimated the rate at which it would settle out of the water column by inappropriately basing the settling rate on Stokes’ Law for more massive spherical particles. Flat clay particles settle more slowly with a longer path length and residence time. Dissolved and colloidal forms of PCB never settle.

EPA’s baseline HRA omitted electrostatic charges on clay particles that separate them, preventing agglomeration and maintaining clay in suspension; EPA also assumed that particles never ‘reflect’ back into the water column after settling, likewise underestimating PCB concentrations in water. Also omitted was PCB codistillation, in which PCBs at low bulk concentrations in river water preferentially distribute to the air-water interface, greatly accelerating PCB transfer from water to air. EPA cited its own empirical data showing more rapid PCB water-to-air transfer but, professing to disbelieve the empirical data, reduced the effect of this data on the HRA: EPA reduced the water-to-air transfer coefficient for PCBs by averaging in modeled PCB transfer coefficients that were orders of magnitude lower than the empirical findings because the modeled values ignored codistillation. Finally, EPA omitted PCB release to the atmosphere from hot water in cooling towers in communities along the Hudson River. Water at cooling tower temperatures may release PCB into the air more than 10 times faster than rates determined from the surface of cold water and multiple orders of magnitude more rapidly than predicted by EPA’s cold-river models.

Together, EPA’s procedures reduced airborne PCB concentrations from above to below de minimis concentrations. This in turn eliminated the requirement for EPA’s HRA to quantify inhalation risks posed by airborne PCBs. The HRA, therefore, ‘considered’ airborne PCBs, but erroneously attributed zero (de minimis) health risk to them.

Our comments herein are informed by the troubling context of error and pro-dredging bias described above for the decision making process forming the basis for EPA to justify and specify the dredging remedy for Hudson River PCB ‘hotspots’. Consistent with this troubling context, our comments adopt two premises:

--1 that the scope of Peer Review Panel issues excludes consideration of terminating the Hudson River dredging project, thereby abandoning Phase 2; and

--2. that the manner in which Hudson River dredging Phase 2 will be conducted minimally must assure net benefit compared with no-action/no-dredging.

Three major issue areas most essentially are addressed in our comments, and should be considered by the Peer Review Panel:

--1. the possible need to modify dredging in Phase 2,

--2. the possible need to modify the benefit criterion in Phase 2, and

--3. the possible need to modify the monitoring program in Phase 2.

Our comments address technical and policy issues within these three major issue areas.


Our analysis adopts the methods of health risk assessment (HRA), critical evaluation of scientific information sources (for example 1-10), and objective scientific peer review. The latter are not a priori methods, and they are not described in detail here. Rather, they consist of the diverse methods typical of peer review by scientists seeking to remain objective. Most essentially, these methods consist of considering the scientific merit with which numerous methods were selected for use and applied prior to dredging, during dredging Phase 1, and after dredging Phase 1. The scope of our comments therefore includes our own peer review of GE and EPA methods, findings, and conclusions, such as those reported orally in public meetings, and in written public communications on GE (4) and US EPA (8, 9) websites for Hudson River dredging (5, 9, 10), and more formally in GE (3, 4) and EPA (8) draft and final reports published for consideration by the public, specific interested parties, and members of the Hudson River dredging project Peer Review Panel. Members of the Peer Review Panel and other readers of our comments can judge for themselves whether and to what degree we succeeded in applying the methods of HRA and of peer review objectively. We hope that we have done so completely.


Possible Need To Modify Dredging In Phase 2

Mobilization of PCB-tainted sediment

A major discrepancy exists between sediment mobilized in Phase I of GE’s EPA-mandated Hudson River PCB dredging project vs. the much smaller amount of sediment mobilization measured and reported by GE (3-5) and EPA (8-10). Sediment mobilization as reported refers to the amount that is ‘resuspended’ by dredging, and monitored miles downstream of the Phase I dredging area, almost all of which is located near Roger’s Island. As EPA also has reported, however, most dredged sediment falls back to the river bottom in the trench or near the spot from which it was dredged initially. Thus, the preponderance of mobilized sediment remains on the river bottom, still mobile, but unrecorded in GE or EPA sediment mobilization data... hence the ‘sediment mobilization discrepancy’.

The sediment mobilization discrepancy represents more than merely a difference between a measured and an actual parameter value. Rather, it is a fundamental inconsistency in EPA’s past justification of the need to dredge vs. the Agency’s current characterization of the performance of the dredging project as implemented. The need for dredging was justified by the mobility of sediments in PCB ‘hotspots’ requiring, according to EPA, their removal by dredging. Indeed, a small but persistent trickle of buried PCB moving downstream was documented from some 27 PCB ‘hotpsots’.

In contrast, in the new context of actual dredging, EPA dramatically has altered its concept of mobility. Mobility in the dredging project now is quantified by the lightest-weight fractions of PCB-tainted sediments measured resuspended in the water at the Thompson Island Dam about five miles downstream of the preponderance of Phase I dredging, and further downstream at Lock 5 and still further downstream at Waterford (see EPA’s figure, below; click on it to enlarge).

Don’t be confused by EPA’s altered terminology, referring to mobilized sediment as sediment ‘loading’ or ‘resuspension’ estimates. These estimates reflect only near-term mobilization, and ignore the fact that all of the sediment that falls back to the riverbed also is ‘mobilized’, in the original sense of that term as used by EPA to justify dredging. That is, it is mobilized because it can enter riverine ecosystems, and it can re-enter the water column via physical, chemical, and biological processes; examples include, respectively: scouring under turbulent high-flow river conditions, dissolution, and microbial metabolism. Mobilized sediment also can be transported by migrating organisms, such as fish and birds, and it can enter the air in communities directly from river water, or from heated sources (cooling towers) used to control industrial processes, such as in factories and power plants.

Quantifying mobilization of PCB-tainted sediment

We previously estimated that 80 percent of sediment disrupted by dredging in Phase 1 was mobilized rather than transferred successfully to barges. This estimate was based upon the size, shape, and operation of a typical five-cubic-yard dredge bucket, and was approximate at best, in part because not all buckets were of five-cubic-yard capacity, as shown by the table of dredge buckets (see EPA’s figure, below; click on it to enlarge) used in Phase 1 (3). We now have quantified this parameter more reliably by using GE’s ‘bucket files’ (3).

GE’s ‘bucket files’ are computer registers recording each closure of a dredge bucket in each delineated five-acre work unit (‘Certification Unit’, or CU) in the Phase I dredging area. Analysis of these GE files reveals that the preponderance of sediment disturbed by dredge buckets was left in mobile form on the river bottom, not placed in waiting barges. We related the number of bucket closures in each five-acre dredging Certification Unit to separately reported information about the volume of sediment placed in barges in each CU. This analysis is summarized in the table below (click on it to enlarge), in which all bucket closures in all CUs are summed.

As the above table shows, Phase I dredging pulled out 286,006 cubic yards of sediment (actually a slurry of sediment and river water), which went into barges (topped by a layer of water). The dredge buckets closed 221,521 times, producing an average load of 1.29 cubic yards per bucketload that was transferred to waiting barges (including the water). The amount of sediment to be rail-shipped to Texas is less, because sediments must be dried before loading. Even using the larger figure as the estimate of barged ‘sediment’, however, the amount barged still is only 26 percent of an assumed five-cubic-yard average bucket capacity, or 32 percent of a four-cubic-yard capacity, or 43 percent of a three-cubic-yard capacity.

That means that the amount of sediment that was mobilized was 74 percent if the average bucket capacity is assumed to be five cubic yards, 68 percent if the average bucket capacity is assumed to be four cubic yards, and 57 percent if the average bucket capacity is assumed to be three cubic yards. Whatever assumption you choose, the preponderance of material was mobilized, not barged. Whatever assumption you choose, still more was mobilized when dredge buckets descended to the river bottom, but failed to close. Some of the mobilized sediment, however, might be dredged again, and a fraction removed from the river, if the sediment falls back to the river bottom within a ‘prism’ slated for future dredging.

We regard our best estimate as 74 percent for the five-cubic-yard capacity bucket typically used, but that estimate excludes sediment that is disrupted by dredge buckets when they crash to the bottom but fail to close due to obstructions such as boulders or construction debris. The initial 80-percent estimate therefore looks pretty good, though we can quantify reliably only a 74-percent estimate, which still is only approximate.

The mobilized fraction of sediment, therefore, amounts to about 211,000 cubic yards, which is approximately 72 million kilograms, assuming a dredged sediment density of about 2.6 compared with water (1). Yet, EPA’s figure (reproduced above) reports 388 kg resuspended sediment load at Thompson Island, 226 kg at Lock 5, and 122 kg at Waterford. Ignoring the obvious double counting, EPA reports 736 kg of mobilized ‘resuspended’ sediment, which is drastically less than 72 million kg. Thus, EPA’s figures exclude nearly all mobilized sediment… EPA simply has ignored the preponderance of sediment mobilization, and PCB mobilization, in evaluating the performance of the dredging project in Phase 1, notwithstanding that the persistent mobility of dissolved, colloidal, and fine-particle-adsorbed PCB constituted for EPA a central rationale for specifying the dredging remedy for the Hudson River PCB Superfund Site.

Downstream PCB deposition

PCB in dredged, redeposited sediments is more mobile than was the case in the original buried state, which ironically is the physical state that EPA’s strategy of capping dredge prisms has sought to restore. From the river bottom, resting sediment piles produced by Phase I dredging gradually (perhaps over years) can and will erode. Some sediment will travel downstream, to be measured by GE in the EPA-mandated ‘resuspension’ monitoring program. Indeed, on 28 March 2010 recently analyzed Hudson River water samples were reported to harbor PCB levels nearly five times higher than the Federal drinking water standard of 500 parts per trillion (ppt). This news should have surprised no one. The episode was caused by scouring of PCB from the river bottom during a ‘high-flow’ event, in which river flow past Thompson Island increased from 5,000 cubic feet per second (CFS) to a peak of 36,000 CFS.

Demonstrably, as shown above, residual sediments that are disturbed by dredging are mobile, along with their PCB load. As explained earlier, if a full dredge bucket averages five cubic yards, then just about a quarter of dredged sediment was transferred to barges. Some of the remainder must flow downstream with the current. The rest falls back to the river bottom, where it exists as loosely agglomerated piles of mud. River currents erode this material gradually back into the water column, from which PCBs may enter the air and ecosystems, including migrating fish and birds.

As the river slows to its normal flow rate, some scoured, resuspended PCB-tainted sediment will settle (redeposit) downstream. As EPA’s figure above shows, the amount of resuspended sediment diminishes from upstream to downstream, documenting (and roughly quantifying) redeposition (roughly, as dilution also must have contributed to the decrease).

Significance of downstream deposition

Downstream deposition of river sediment involves a wide spectrum of particle sizes. Clamshell dredges capture mostly the coarse particles: sand and rocks that have relatively little surface area. In other words, clamshell dredges preferentially leave behind particles having the greatest surface area, on which the greatest amount of PCBs are adsorbed, while instead preferentially capturing a small fraction of the PCBs that are adsorbed in sand and on rocks. Upon closure of ‘clamshells’, these fine particles preferentially leak into the water column, which carries much of it downstream. The particulates eventually settle. Repeated action of the dredges causes this moving front to become more concentrated with PCBs.

Clamshell buckets, as described, above, tend to exert not only a mobilizing effect on sediment, but also a resorting (or particle ‘classification’) effect, which constitutes one burden of PCB dredging. Moving the pontoon-mounted dredge platform and clamshell downstream effectively constitutes chasing a moving, increasingly concentrated PCB front to previously uncontaminated river areas. Depending on the speed of water flow this can be a few hundred feet downstream or many miles downstream of the dredge.

Perhaps the most damaging to the environment is the pontoon-mounted backhoe/excavator type of dredge, but this is the type used in Phase I. Closing the bucket underwater compresses the sediment, forcing much out of the open top of the dredge bucket, thereby promoting dispersion of fine particulates into the water column. These are slowest to settle, and most susceptible to downstream migration. To prevent such migration, other types of dredges may be used. Those that suck-up the bottom sediment with water to be discharged into hoppers for sedimentation for treatment are the most effective. The process of dredging-induced gradual sediment resuspension followed by downstream flow and redeposition illustrates that, most essentially, four insidious processes are underway.

1. PCB spreading. One insidious process is spreading of PCB from sectors of the river bottom dredged in Phase 1 to undredged sectors downstream. Dredging in Phase 1 was intensively concentrated in a relatively small area (somewhat over half of the full 90 acres included in Phase 1), so dredging affected a high fraction of the river bottom in the Phase 1 area. This cannot be the case in Phase 2. Most of the river bottom on which resuspended PCB-tainted sediments will settle is not scheduled for dredging in Phase 2. As a result, PCB will continue to spread and redeposit over a gradually increasing area of river bottom, nearly all of which is not scheduled for dredging, ever.

2. PCB entry into ecosystems. As the area of river bottom affected by redeposition of mobilized PCB-tainted sediments increases, PCB will enter river ecosystems beyond the Phase 1 area. These will include detritus ecosystems and ecosystems involving the higher trophic levels, from primary producers (phytoplankton and rooted plants) to herbivores (such as carp), primary carnivores (zooplankton and some fish), and secondary and tertiary carnivores (including some large fish, amphibians, reptiles, and mammals).

Already, GE and EPA have reported five-fold increases in PCB concentrations in fish tissue, but this data represents a misleading underestimate of the degree of increase. The data reflect analysis of muscle tissue of fish (filets), whereas PCB levels in organs (especially fish liver) may be two to three orders of magnitude higher. Some people, especially American Indians, consume fish organ meats. Fish also are consumed in their entirety by microbes and by predators other than humans. Thus, the load of PCB entering Hudson River ecosystems from dredging are grossly underestimated by data on PCB concentrations in fish filets.

3. Sufficiency of lesser high-flow events to suspend redeposited PCB. As the area of river bottom affected by downstream movement of mobilized PCB increases, the effectiveness of increased river flow at suspending redeposited PCB-tainted sediment will increase, even if the scouring efficiency remains constant for any flow rate. This is because the scouring efficiency, whatever it is, will affect a larger area of the river bottom. In addition, the scouring efficiency may increase for any given river flow rate, because relatively low-mass particles will be the ones preferentially transported downriver from the Phase 1 area. This also suggests that the duration of river flow-induced exceedances of the 500-ppt Federal Drinking Water Standard for PCB may increase, as the fraction of river bottom contributing to the exceedances increases. Such high-flow events seem to occur with a frequency in the range of once to 10 times per year.

4. Hudson River ecosystems harbor an increasing fraction of PCB burden. The processes described above can be expected to alter the distribution of PCB originally confined to 27 ‘hotspots’ slated for dredging to a wider, and widening, river sector. This sector will continue to expand and, along with it, the fraction of Hudson River ecosystem area affected by PCB will increase. These riverine ecosystems will relate to mobilized PCB-tainted sediments originating from dredged ‘hotspots’ as a source and sink.

This phenomenon was documented in an extreme example involving the chlorinated pesticide DDT, which closely resembles PCBs both structurally and dynamically (though PCBs are somewhat more water-soluble and less long-lived in ecosystems). In Clear Lake, California DDT (and its breakdown products DDE and DDD) were monitored in water, and not found, as they are nearly completely water-insoluble. This negative finding, however, proved to be misleading, as the biota of Clear Lake proved to be a major sink for DDT, which exhibited an extreme tendency to bioconcentrate, bioaccumulate, and biomagnify.

By these processes DDT (and related chlorinated hydrocarbons) exerted significant effects from which global ecosystems are just now recovering. For example, the bald eagle and other birds of prey (notably peregrine falcons) were threatened with extinction because DDT thinned their eggshells, causing drastic declines in their populations. Similarly, dredging-induced mobilization of PCB-tainted sediments reasonably may be expected to reintroduce PCB into receptive ecosystems, thereby setting back the clock for river recovery by four decades, during which PCB may be reburied as they were before being dredged this past year in Phase 1.

Possible Need To Modify the Benefit Criterion In Phase 2

Contravention of the benefit criterion

The EPA figure reproduced above depicts contravention of the benefit criterion. This figure has been updated to reflect the totality of Phase 1 dredging (though the updated figure is not reproduced here). GE and EPA used the mid-Phase-1 data to attempt dredging modifications in real time to restore adherence to the benefit criterion. These efforts succeeded at reducing PCB resuspension, but not to a degree sufficient to restore adherence to the benefit criterion. These facts have been widely acknowledged, and will not be amplified herein.

Flaws of the benefit criterion

As described above, the preponderance of dredge-mobilized PCB will not move downstream during the planned six-year period of the dredging project, or for many years beyond. EPA’s criterion of ‘benefit’ of the dredging project therefore is illogical. The ‘benefit criterion’ allows no increase in downstream transport of resuspended sediment during the planned six-year duration of the project. Instead, under the benefit criterion, downstream transport may be increased by dredging in the short term only in amounts that will be offset by future decreases over the longer term, meaning the project duration. This benefit criterion is illogical because it assumes incorrectly that potential dredging impacts consist of nothing more than downstream transport of resuspended sediment over a six- or seven-year period, whereas this appears to constitute a relatively minor contributor to total potential dredging impacts.

The potential harm of dredging to the Hudson River, river ecosystems, and river communities is drastically greater than suggested by resuspension alone. Resuspension and downstream transport of resuspended sediment together constitute just a miniscule fraction of the potential burden of dredging-induced PCB mobilization. Maintaining a stringent resuspension performance standard, even if it could be met, still allows harm from entry of PCB from tainted sediments deposited on the river bottom into river ecosystems including migrating birds and fish, and into the air that is breathed by people living in river communities.

Possible Need To Modify the Monitoring Program In Phase 2

Liquid PCB oils

EPA reported at a public meeting in the Hudson River town of Fort Edward, New York on 19 August that GE clamshell dredges recently had started to encounter liquid PCB oils in dredging ‘prisms’. We attended that meeting. EPA's revelation constituted a virtually silent bombshell, as apparently none of the many media representatives in attendance took special note of it. It is worthy of note, however, because pure PCB oils must be viewed in an entirely different, more ominous context than river sediments harboring PCBs in the parts-per-million (ppm) concentration range. Pure liquid PCB oil concentrations can be expressed in ppm if necessary: 1,000,000 ppm. Mobilizing PCB oils via dredging is commensurately more serious than mobilizing sediments bearing PCBs in the low ppm range as originally anticipated.

EPA later acknowledged that the project recently had discovered “sheens” of liquid PCB oil on the river surface, “an indication, the EPA said, that the river floor contained not only contaminated silt, but more potent pockets of pure PCB oil – and that the dredging is releasing the oil into the river water.” EPA, however, placed a relatively benign spin on this news: “It’s not really affecting dredging, that’s why we’ve taken those mitigation measures to, you know, to counteract the sheens.” Counteracting microscopically thin PCB ‘sheens’ visible on the river surface is very different from the real challenge: counteracting remnant pools of originally-disposed liquid PCB oils of unknown, potentially large volume in trenches beneath.

PCB sheens on the river surface constitute evidence of the presence of liquid (near pure) PCBs, and constitute yet another source of airborne PCBs not addressed in any EPA assessment of risks to health potentially posed by sediment dredging at PCB hotspots in the Hudson River. These PCB sheens cannot have originated from desorption from sediments, because a sheen is a continuous monomolecular layer of liquid PCB. Their observation, therefore, raises the question of whether they might be expected to arise from underlying sediments bearing PCBs in merely the ppm range, or whether they must originate from more massive pools of liquid PCB oils at the river bottom as suggested by EPA.

Can bottom sediments form surface sheens via some upward PCB migration process? Three factors suggest not:

--1. PCBs chemically bound to sediments for decades are unlikely to become unbound, especially as liquid PCB oil;

--2. If they did become unbound, they also would have to become concentrated from the low-ppm range at the river bottom to near purity at the surface, which is unlikely to occur in a flowing, turbulent river, and

–3. PCB oils are denser than water, so they would be expected to sink, not to rise from the bottom, unless physically forced upward, or lifted; just a small fraction of more water-soluble PCB congeners might reach saturation in the water, and may appear as a sheen on the river surface.

PCB liquids include more than 200 types, or congeners, varying in their degree and pattern of chlorination (with from one to 10 chlorine atoms per PCB molecule). Each congener has a unique density, but bulk density of commercial PCBs of the types disposed to the Hudson River (known as Aroclors) are reported to have a density of about 1.5 grams per milliliter (g/ml; 1, 7), which is 1.5 times the density of water.

The most probable origin of PCB liquids forming surface sheens, in our view and apparently also in EPA's view, is that they are being massively disrupted from pools of liquid PCB oils formed in sediment low points (depressions) following original disposal from land-based facilities or ships. These bottom pools gradually might have become covered over with debris and sediments. The onset of dredging may be forcing the liquids upward toward the surface as dredge jaws expose them and close around them. This process is visible for sediments (which have tightly-bound PCBs that would not be expected to form sheens), and reasonably would apply as well to liquid pools (which would be expected to form surface sheens).

Sediments and PCB oils that are not forced upward by closing dredge jaws might be retained within dredge buckets, and physically lifted. These materials would be subject to leakage during their ascent to the surface and beyond. These dual processes of disruption by squishing and by lifting reasonably would be expected to generate PCB liquids and PCB sheens at the surface, as PCBs are chemically attracted to surfaces, including to the air-water interface in rivers.

EPA’s report of encountering PCB liquids seems, at least in retrospect, unsurprising given the history of past PCB disposal into the river in the form of liquid PCB oils from land-based facilities and from ships. Given this history, why did EPA adopt the clamshell method of dredging, thereby failing to prepare for this seemingly expectable eventuality? Its sobering actuality casts further doubt on the wisdom of EPA having specified clamshell dredging rather than vacuum dredging, or no dredging, for remediation of the Hudson River PCB Superfund Site.

One objection to the liquid-pool hypothesis that must be addressed is the fact that PCB concentrations in downstream water samples have not revealed PCB oils or concentrations high enough to suggest their presence upstream. The heavier-than water density of PCBs, however, would be expected to cause them to hug the river bottom as they move downstream. They would not be expected to register in surface or mid-depth water samples taken five miles downstream.

EPA and the Peer Review Panel now appear to be faced with the quandary of whether to:

--1. mobilize PCB oils by continuing clamshell dredging, even though the dredge buckets cannot retain the oils efficiently,

--2. allow the PCB oils to be mobilized by river currents washing over the now-exposed pools,

--3. institute vacuum dredging, or

--4. stabilize the oils by covering them over again.

We wish we had the answer to this quandary, but for now suffice it to bring this issue to the light of public scrutiny… which EPA’s silent bombshell on 19 August did not seem to do.

Monitoring for waterborne PCB

Waterborne PCB in the Hudson River, including water in dredge buckets and barges, constitutes the primary source of airborne PCB. Indeed, EPA studied the relationship of waterborne PCB in a cold river to airborne PCB a meter above the river surface. That study produced the EPA Level of Concern (LOC) for airborne PCB of 0.08 ug/M3, associated with EPA’s Maximum Contaminant Level (MCL) for PCB of 500 ppt in water. Generated primarily by dredging, waterborne PCB therefore should be measured where dredging occurs… but that was not done. Instead of sampling at Rogers Island, water was sampled, and waterborne PCB measured, near Thompson Island, about five miles downstream of Rogers Island. As a result, waterborne PCB concentrations at dredging sites are anyone’s guess, specifically determined by anyone’s guess about the applicable dilution factor that occurs as dredge-mobilized material travels five miles downstream.

Monitoring for airborne PCB

Just as waterborne PCB concentrations at dredging locations are anyone’s guess, airborne PCB concentrations at dredge sites also are anyone’s guess, because airborne PCB was not routinely measured (and might not have been measured at all) at dredging platforms. Airborne PCB instead was monitored by two portable air samplers set up adjacent to each dredge platform, on opposite shores of the river. These samplers recorded 24-hour average concentrations of airborne PCB. Three problems undermine the usefulness of these monitors. First, personnel engaged in dredging are exposed to aerosols generated at the dredge platform, which reasonably would be expected to fall back to the river surface before reaching monitors on shore, except when wind speed is adequate to transport them as far as the shore, and when wind direction is aimed at the portable monitor. In those cases, of course, only one monitor would record the PCB, as the wind would be blowing away from the monitor on the opposite shore. Second, wind direction varies, so each air monitor would be pumping air originating from 360 degrees, not primarily from the direction of the dredge platform to which it is adjacent. This is illustrated in the figure below (click on it to enlarge), in which red arrows represent air originating from the dredge platform, and the preponderance of black arrows represent air originating from other directions.

The figure above also illustrates how, as the dredge platform moves downstream, the angle from which PCB might impinge on the portable monitor increases. Thus, a time series should show increasing airborne PCB concentrations as the surface area of dredged river increases. This observation cannot be made, however, because GE procedures call for moving each portable air monitor downstream with the dredge platform, so no monitor remains to record evolving airborne PCB concentrations at any former dredge site. Indeed, as the area of dredged river surface increases to its maximum, the concentration of airborne monitors per acre of dredged river declines in direct proportion, because no monitors are added to cover the increased area of dredged river.

This gives rise to the third problem: that PCB volatilization might not reach a steady state until long after the portable air monitors are withdrawn downstream with the dredge platforms. Indeed, PCB release from the river surface might not reach steady state rates until long after Phase I is complete. By then, portable monitors would have been withdrawn, not only to points downstream, but altogether. In short, the monitors are withdrawn before they can characterize the evolution of PCB release to the air over the six-year planned life of the project and beyond.

Indeed, as the acreage of dredged river increases, the number of monitors per acre declines, but the fraction of river surface contributing to air levels over the river and on shore increases… never to be measured because the monitors that could have measured these evolving levels were withdrawn from service. We call this practice ‘hit-and-run dredging’. The Peer Review Panel, therefore, should call for a permanent array of air monitors to capture the evolution of PCB release levels as the fraction of river bottom that has been dredged increases, and as PCB-tainted sediments are re-stratified by physical, chemical, and biological processes of degradation, mobilization, and eventual release to ecosystems and the atmosphere.

One strategy to rectify this egregious situation might be effective: examining meteorology data to determine the fraction of time when wind direction was toward a portable air sampler on shore. If the fraction of time was, say, one percent, then the airborne concentration from the point of origin (the dredge platform) would have been at least 100 times higher than the concentration reported by the sampling unit (“at least,” because the sampling unit would not have recorded PCB-bearing aerosols that had precipitated back to the river surface before reaching shore). We do not know if the required meteorology data are available.

Personnel monitoring for airborne PCB

A major concern associated with PCB dredging relates to risks potentially posed to human health (Michaels and Oko, 2007; 6). Apart from participation of people as ecosystem components at the top of the food chain, we are concerned about airborne PCB, which is derived from PCB released to the water and largely unmeasured. In that regard, we urge the Panel to demand delivery of the results of personal monitoring samples taken on dredge platforms (minus employee identifying information). These samples generated the only data of which we are aware that reflects airborne PCB levels produced by dredging, at the location of dredge platforms. These samplers all operated with sensitivity to the occupational airborne PCB limit of 1,000 ug/M3, which is over 9,000 times higher than EPA’s residential limit of 11 ug/M3… so these samples all had better be negative. EPA indicated that the data belong to GE, and GE indicated that the data belong to its contractors. Information of possibly critical concern to the public clearly has been ‘compartmentalized’ out of public view. Members of the public and of the Peer Review Panel have an opportunity now to demand an end to such information compartmentalization.


Our findings justify conclusions and recommendations, as follows.


--1: Risks to human health potentially posed by airborne pathways of exposure to PCB under dredging and non-dredging scenarios were neither assessed objectively, nor compared.

--2: Water monitoring miles downstream of dredging is inadequate to characterize PCB mobilization.

--3: Air monitoring using portable air samplers on shore is inadequate to quantify either residential or commercial exposure to airborne PCB.

--4: Results of personal airborne exposure monitoring of GE dredging personnel to PCB is undisclosed, and therefore fails to satisfy the legitimate public interest in gaining access to this unique source of data on airborne PCB levels at dredge platforms.

--5: EPA’s emphasis on PCB sheens found during dredging, rather than the possibly substantial sources of these sheens, is misplaced.

--6: The dredged matter that is treated and shipped represents a small fraction of the total mobilized in the river by clamshell dredges.

--7: Levels of airborne PCB probably are higher than suggested by available data, and possibly unsafe.

--8: Hudson River PCB dredging is making exposed people into experimental subjects, and riverfront communities into the subjects of epidemiology studies that will continue for generations to come.


--1: Risks to human health potentially posed by airborne pathways of exposure to PCB under dredging and non-dredging scenarios should be assessed objectively and compared via correction of prior EPA assessments before onset of Phase 2.

--2: Phase 2 must expand environmental monitoring to generate adequate data to the public domain that can warn of harm in real time, and enable evaluation of the project over decades.

--3: Water and air monitoring, including personal monitoring, should occur together at dredge platforms.

--4: Results of personal airborne exposure monitoring of GE dredging personnel to PCB should be made available to the public, as is the case with other project data.

--5: Permanent arrays of air and water samplers are needed to confirm or refute EPA safety claims, and to protect public and environmental health.

--6: The sources of PCB sheens, and the composition and volume of the sources, should be investigated and, if possible, their relationship to originally disposed PCB liquids determined. Strategies to manage possibly rich sources of PCB liquids that dredge buckets disrupt and fail to hold must be developed to prevent mobilization of liquid PCB and assure compliance with the benefit criterion.

--7: Technologies should be considered for use in Phase 2, such as coffer damming, hydraulic dredging, and dredging within enclosures to minimize PCB mobilization, resuspension, and PCB entry into ecosystems and the air.

--8: Vacuum dredging where the water, sediment and rocks are captured without dispersing fine particles into the water column may be the best method. The pontoons could be equipped with sedimentation basins in which flocculates that bind fine particulates can be added to the slurry. The slurry then could be filtered on-site to return clean water into the river. Alternatively, the pontoons could move their entire load into the treatment plant, where proper treatment can be applied and the water returned to the river.


01. Gardiner, W.W.; et al. Evaluation of dredged material proposed for ocean disposal from Hudson River, New York. Washington, DC; Department of Defense (Sponsor); Richland, Washington; Pacific Northwest National Laboratory. Technical Report No. PNNL—11342; OSTI ID: 408098; Legacy ID: DE97050267; 320 pages, 1 September 1996;

02. GE QAPP. River PCBs Site, Phase I Remedial Action Monitoring Program, Quality Assurance Project Plan; Final. Prepared for General Electric Company (GE; Albany, New York) by: Anchor QEA, LLC (Liverpool, New York) in conjunction with Environmental Standards, Inc. (Valley Forge, Pennsylvania) and ARCADIS (Syracuse, New York); 344 pages, i.p., May 2009;

03. GE Draft Phase 1 Evaluation. Phase I Evaluation Report: Hudson River PCBs Superfund Site. Draft report prepared for General Electric Company (Albany, New York) by: Anchor QEA (Glens Falls, New York) and Arcadis (Syracuse, New York); 191 pages plus tables, figures, and appendices; Appendix G, Table G-1, 44 pages; Appendix N, Table N-1, 1 page; January 2010;

04. GE Final Phase 1 Evaluation. Phase I Evaluation Report: Hudson River PCBs Superfund Site. Final report prepared for General Electric Company (Albany, New York) by: Anchor QEA (Glens Falls, New York) and Arcadis (Syracuse, New York); 667 pages including tables and figures; March 2010;

05. GE Hudson River Dredging Website. http://www.hudsondredging.com;

06. Michaels, R. A.; and U. M. Oko. Bias in the US Environmental Protection Agency’s Health Risk Assessment Supporting the Decision to Require Dredging PCB-Bearing Sediments the Hudson River. Environmental Practice (Cambridge University Press), 9:96-111, 2007;

07. UN EP. Training manual for the preparation of a national environmentally sound management plan for PCBs and PCB-contaminated equipment in the framework of the implementation of the Basel Convention. Ch√Ętelaine, Switzerland, United Nations Environment Programme; Basel Convention Series/SBC No. 2003/01, ISBN : 92-1-158674-7, 103 pages, March 2003;

08. US EPA. Hudson River PCBs Site EPA Phase 1 Evaluation Report. Prepared for: US Environmental Protection Agency, Region 2; and U. S. Army Corps of Engineers, Kansas City District. Prepared by The Louis Berger Group (Morristown, New Jersey), 272 pages plus appendices, March 2010;

09. US EPA Hudson River Dredging Website. http://www.epa.gov/hudson;

10. US EPA Hudson River Dredging Data. http://www.hudsondredgingdata.com.