Wednesday, August 12, 2009

FIRESIDE CHAT: Health Risks and Hudson River PCB Dredging

further info:

In recent days the U.S. Environmental Protection Agency (EPA) twice ordered the General Electric Company (GE) to cease dredging Hudson River sediments at Fort Edward after measured PCB levels in water and air exceeded acceptable levels. On Monday EPA ordered GE to resume dredging after PCB levels had returned to within acceptable limits. EPA insisted that dredging has been safe, but EPA’s sampling stations are remote. Specifically, they are further from the dredge buckets than are employees working on dredging platforms, and they are further than residents are at some riverfront homes. This raises the question of whether the Hudson River dredging project as currently implemented has exposed these employees and residents, protected by neither respirators nor masks, to unacceptable PCB levels.

I can offer two answers to this question. The first is ‘apparently yes’, and the second is that documentation is problematic because the nearest EPA sampling stations are five miles downstream for water and a difficult-to-determine distance (perhaps hundreds of yards) for air. Acceptability for water samples is defined by the applicable drinking water quality standard (500 ppt) which, EPA asserts, will protect municipal drinking water supplies located downstream.

This might or might not be true. The Maximum Contaminant Level (MCL) established under the Federal Safe Drinking Water Act is, indeed, 500 ppt. MCLs however, are set as enforceable limits that are as close as possible to health goals. As a known cancer-causing agent (carcinogen) in animals, and as a probable human carcinogen, PCBs are subject to an MCL Goal (MCLG) of zero ppt under the Safe Drinking Water Act. That is, EPA has the onus of enforcement when PCB levels reach 500 ppt, and of prohibiting projects that would generate PCB levels even below 500 ppt.

Unquestionably, PCB levels in water are significantly (many times) higher at the point where they are mobilized from sediments at dredge locations, before being diluted with river water during five miles of transport downstream to EPA’s water sampling stations. EPA’s data collection protocols exclude water sampling at dredge locations, so we cannot know for sure just how diluted PCBs are five miles downstream, or how many times more concentrated they are at dredge locations. A reasonable, order-of-magnitude guess would seem to be that PCBs would be transported in turbulent flow, undergoing at least a 10-fold dilution, and possibly 100-fold or maybe even 1000-fold, depending upon river flow volume and flow rate. This might not pose a problem for people drinking water nearby, if they do not drink Hudson River water.

PCBs, however, enter the air from the water, especially when they are introduced near the surface of the water. EPA quantified the relationship between PCBs in cold river water and in air a meter above the river surface. Specifically, for each ug/L (that is, for each 1000 ppt) of PCB in river water, PCB levels in air (in ug/M3) were reported to be: a minimum of 0.02, a median of 0.09, a mean of 0.15, and a maximum of 0.40 ug/M3. These values are explained and cited in a fully peer-reviewed article by myself and co-author Dr. Uriel Oko, published in the Cambridge University Press journal Environmental Practice. This article is available for download in its entirety, for free, at

River water five miles downstream of dredge platforms, where PCB levels in water exceeded 500 ppt and forced cessation of dredging, reasonably might be expected to produce airborne levels of PCB that exceed half of the values associated with 1000 ppt, reported by EPA (above): a minimum of 0.01, a median of 0.05, a mean of 0.08, and a maximum of 0.20 ug/M3. Among the four values, as EPA knows, the mean value is the most significant, being the airborne concentration that captures the variability of PCB levels in time and over space in and above river surface waters. Indeed, the reported mean value of 0.08 ug/M3 in air a meter above the river is equal to EPA’s Level of Concern (LOC) for airborne PCB in residential settings, as reported in EPA’s Quality Assurance Project Protocol (QAPP) for the Hudson River dredging project.

Upstream, at dredging platforms where levels of PCB in water reasonably might be expected to be 10 to 100 or maybe even 1000 times higher than five miles downstream, resulting airborne levels on average might reasonably be expected to be 10 to 100 or maybe even 1000 times higher than EPA’s LOC. EPA also has set a stop-dredging standard of 0.11 ug/M3 for airborne PCB in residential settings, and of 0.26 ug/M3 for airborne PCB in commercial or industrial (occupational) settings (as also reported in the QAPP). Both of these standards would be exceeded if airborne PCB is 10 times higher at dredge sites than at water sampling sites five miles downstream where PCB levels reach 500 ppt in water. Indeed, a 10-fold increase of the air levels over downstream sampling points would amount to 0.8 ug/M3, exceeding the residential standard of 0.11 ug/M3 by a factor of more than seven-fold, and the occupational standard of 0.26 ug/M3 by a factor of more than three-fold.

The values estimated above represent airborne levels at dredge sites occurring when downstream PCB levels in water reach the 500-ppt stop-dredging threshold, but the problem is worse than that. When values are more typical, say about a third of the stop-dredging threshold, air levels still would be predicted to exceed the airborne acceptability limit at the dredge buckets by somewhat smaller factors (one third of the factors estimated above). That is, airborne PCB levels apparently exceed EPA acceptability limits pervasively, not just occasionally during brief intervals when levels are high enough to cause EPA to order cessation of dredging.

This problem is only the tip of the iceberg, however, because it also assumes that downstream PCB levels in water were diluted only by a factor of 10-fold compared with levels at the source, at dredging platforms. In turbulent river flow, dredged PCB would mix into a much larger general river flow over a distance of five miles, probably producing dilution factors more in the range of 100-fold or maybe even 1000-fold. In that case, airborne PCB levels at dredge sites, where unprotected employees are working, would be 10-fold or maybe even 100-fold higher.

The factors estimated above represent gross exceedances of acceptability limits. The acceptability limits are not my limits; they are EPA’s limits. The problem, however, might be worse even than the exceedance factors estimated above. This is because the original EPA data relating airborne levels of PCB over a cold river to the source PCB levels in the river water do not account for dredge buckets in the Hudson River billowing sediments to the river surface as they close, and lifting sediments and water above the surface, where leakage drops sediments and water violently onto the river surface. This process must produce abundant droplets (aerosols) of various sizes, all containing PCBs, whereas air samples above river water in EPA’s original data set did not include this source of PCB. So, for occupationally exposed individuals, the relationship of PCB levels in water to levels in air must be much worse, exposing them to inhalable PCBs that are volatilized as well as to PCBs that are aerosolized. That unhealthy process happens yet again when the dredges swing their retained load over the barges, and drop them again, this time in their entirety, producing yet another burst of PCB entry into the air in the form of vapors and aerosols.

The obvious question is: why are such troubling airborne PCB levels not captured in EPA air samples to either side of the river at dredge sites? The answer is either that I am wrong, which I don’t believe, or that EPA’s sampling process is flawed in a way that causes it to miss the levels to which employees on dredge platforms actually are exposed. One major problem with EPA’s airborne sampling procedure is that samples are taken at significant distance from the dredge platform, maybe hundreds of yards away, whereas employees are right there, at zero distance away. The sampling distance is not discernible from maps of air sampling locations that EPA gave to me when I visited the Hudson River Field Office recently.

Remote sampling generally is not protective. If you work in a radioactive area, you must wear a personal radioactivity monitor that reports your personal exposure. If you work in a coal mine, you wear a personal airborne particle monitor to report your personal exposure to airborne particles (coal dust). In general, monitors are placed where they reliably capture the levels of toxic substances to which employees are potentially exposed. This is not EPA’s practice in the Hudson River PCB dredging project… and one must wonder why it is not.

A second major sampling problem is related to the first. That is, by the time air reaches the sampling units, much of the PCB has dropped out, because aerosols are heavier than vapors, and do not carry as far in air currents (wind). So, EPA samplers are capturing only a fraction of PCBs that dredge employees’ respiratory systems must be capturing. By resuming dredging with fewer dredge units, EPA will succeed at reducing PCB levels in water five miles downstream, and air levels at sampling units… but that ‘engineering control’ measure will not reduce airborne PCB levels to which employees of dredge platforms will be exposed, though it will reduce the number of exposed employees.

A third major sampling problem is that wind speed and wind direction are variable. Only a fraction of PCBs entering the air at the location of dredge platforms and employees working on them will reach the sampling units. Of course, dilution in air will have occurred even when the air blows directly toward the samplers. Air samples are 24-hour averages, however, virtually assuring that wind direction will shift, steering airborne PCBs away from sampling units.

I draw three qualitative conclusions from the above observations. The first is that airborne PCB levels to which employees and some riverfront residents are exposed via inhalation appear to exceed EPA acceptability limits significantly. The second is that the Hudson River PCB dredging project is being conducted in a manner that fails to capture and record such exceedances, if they are occurring. The third is that the burden of proof now rests with EPA to show that its pre-project predictions of no unacceptable health risks associated with PCB dredging are correct, in view of evidence clearly suggesting that they are incorrect based upon data from actual project experience elucidated here.

Copyright © 2009 by The Center for Health Risk Assessment and Management, a Division of RAM TRAC Corporation