Panel discussion: Remedy effectiveness: What works, what doesn't?
As contaminated sediment sites in freshwater and marine environments come under increasing scrutiny, and the importance of these sites with regard to ecological and human health is better understood, the number of contaminated sediment sites requiring risk characterization and management is increasing. Risk reduction must be the long-term goal of all sediment management practices. Risk management strategies in aquatic environments focus on mitigating potential exposure pathways for contaminants that may pose an ecological or human health risk over time. Sediment management practices include dredging (“environmental dredging”), capping, monitored natural recovery, and combinations of these generic response actions. A Remedy Effectiveness Panel was selected to identify and discuss the key issues associated with the performance of these generic response actions for contaminated sediments. The panel convened in New Orleans on 26 January 2005 at Battelle's Third International Conference on the Remediation of Contaminated Sediments and made 3 presentations on response actions to conference attendees followed by an open dialog between attendees and panel members. This article introduces the 3 generic response actions, identifies key topics from the open dialog and presents opinions expressed, and provides a summary and observations.
This paper is among 9 peer-reviewed papers published as part of a special series, Finding Achievable Risk Reduction Solutions for Contaminated Sediments. Portions of this paper were presented by the author at the Third International Conference on Remediation of Contaminated Sediments held in New Orleans, Louisiana, USA in January 2005.
Sediments often are the ultimate receptors of chemical and biological contaminants resulting from agricultural, urban, industrial, and recreational activities. Thus, many river and estuarine sediments are exposed to some level of contamination. Characterizing and remediating contaminated sediments is complex, and the complexity is compounded by underwater conditions, sensitive habitats, water currents, and access difficulties. Moreover, the inconsistency and uncertainty to date in establishing appropriate site-specific cleanup goals (which areas to remediate, to what extent and depth, for what ultimate purpose) have further complicated the matter. An appropriate site-specific cleanup goal may not be obvious: unacceptable human exposure to contaminants in sediment may occur only indirectly, through consumption of fish with elevated contaminant levels. Thus, to understand which sediment to clean up and to what contaminant level, one must 1st understand how the contaminants have moved from the sediment into the fish and from where in the sediment.
Contaminant source control or elimination is an essential 1st step in any contaminated sediment management strategy. After source control, sediment management strategies typically employ 1 of 3 generic response actions: dredging, capping, or monitored natural recovery (MNR). Sediment risk management strategies should seek to balance 2 primary goals: (1) minimizing contaminant risk to human health and the environment and (2) minimizing cost (Apitz et al. 2005). A 3rd goal is becoming recognized: minimizing the risks associated with the response action itself, such as habitat destruction and/or modification, as well as injury to workers or the public during remediation. New cost-effective management (remediation) strategies for contaminated sediments need to be developed, and the extent to which existing strategies can cost-effectively manage contaminated sediments needs to be better understood.
In the United States, about 6 million yards3 of contaminated sediment have been removed and disposed through the implementation of 71 major environmental remediation projects. Virtually all of this remediation by removal occurred after 1990, so remediation of contaminated sediments is a relatively new discipline. Disposal in the great majority of cases has been to landfills, either to dedicated onsite landfills or to offsite commercial landfills. Some disposal, with considerable cost savings versus landfilling, has been to confined aquatic disposal (CAD) cells. Total cost for these 71 environmental remediation projects, including removal and disposal, has averaged $160/yard3 (MCSS 2004).
Capping, either alone or in combination with removal and/or MNR, is planned or has been implemented at about 40 sediment remediation projects in the United States, whereas MNR as a primary remedy, or in combination, is a component of about 28 projects in the United States (MCSS 2004). Both capping and MNR tend to be “a harder sell” as the remedy of choice for regulatory agencies and the public because the contaminants are left in place. As a result, greater emphasis is placed on demonstrating effectiveness than is typically done at a removal site.
The Remedy Effectiveness Panel was selected to identify and discuss key issues associated with the performance of contaminated sediment response actions using dredging, capping, or MNR. The panel convened in New Orleans on 26 January 2005 at Battelle's Third International Conference on the Remediation of Contaminated Sediments, and individual panel members made presentations to conference attendees for each of the 3 response actions, followed by an open dialog between attendees and panel members. This article summarizes key aspects of each of these 3 response actions, the key topics identified in the open dialog, and observations and opinions expressed.
Removal of contaminated sediment for purposes of remediation is called “environmental dredging” to distinguish it from conventional navigational or maintenance dredging. Environmental dredging (and excavation) involves physically removing contaminated sediment from a water body, either submerged (dredging) or after water has been diverted or drained (excavation), to mitigate contaminant exposure pathways. Environmental dredging is intended to remove sediment contaminated above a specified action level, while minimizing the spread of contaminants to the surrounding environment (NRC 1997). The key issues for environmental dredging include, but are not limited to, the following:
Dredge equipment and productivity;
Accuracy of sediment removal;
Efficient management of rocks, vegetation, and debris in the areas targeted;
Contaminant mass removal versus achieving a specific target concentration;
Residual contamination remaining after removal;
Resuspension of contaminants during dredging;
Impacts to habitats, as well as methods of restoration;
Effects of sloughing of disturbed sediments;
Management of dredged materials; and
Sediment capping involves the placement of a subaqeuous covering or cap of clean material over an area of contaminated sediment. Sediment capping mitigates potential contaminant exposure pathways by physically isolating the contaminated sediment from the aquatic environment and by preventing resuspension and transport to other areas. Sediment capping techniques are also applicable to CAD cells, which are becoming the selected disposal option for contaminated sediments at an increasing number of projects. Confined aquatic disposal cells may involve (1) the use of naturally occurring bottom depressions, (2) sites from previous mining operations such as beach renourishment borrow sites, or (3) new dredging operations expressly conducted to create the cell. The key issues for sediment capping include, but are not limited to, the following:
Control of contaminant sources,
Selection of suitable and readily available cap materials,
Evaluation of hydrologic impacts (floods, ice scour) on cap stability,
Stability of the underlying sediment bed,
Adequate water depths to support anticipated uses (navigation, flood control),
Presence of and potential impacts from groundwater,
Compatibility of the cap with infrastructure (piers, pilings, buried cables), and
Habitat impacts and restoration.
The use of CAD cells as a capping alternative has an inherent benefit, level bottom capping. With level bottom capping, contaminated sediments are confined to a smaller footprint, contaminant diffusion pathways are lengthened, and the contaminated sediments are kept farther removed from physical processes that can mobilize the contaminants. The increasing selection of CADs as the preferred response action for sediment projects results from CADs providing an acceptable compromise solution when cost, logistics, regulatory acceptance, environmental risk, and perception are compared with other alternatives. In addition to these benefits, CADs also have the advantage of being constructible with readily available equipment. Confined aquatic disposal cells, as with any alternative, have constraints and disadvantages that need to be considered, including such factors as impacts to aquifers and subsurface geological incompatibilities.
MONITORED NATURAL RECOVERY
MNR is a risk reduction strategy that relies on ongoing, naturally-occurring processes to contain, destroy, or reduce the bioavailability or toxicity of contaminants in sediment. As with dredging and capping, the focus of MNR should be mitigation of contaminant exposure pathways. NRC (1997, p. 91) defines MNR as follows: “Natural recovery involves leaving the contaminated sediments in place and allowing the ongoing aquatic processes to contain, destroy, or otherwise reduce the bioavailability of the contaminants. Although no action is required to initiate or continue the process, natural recovery is considered the result of a deliberate, thoughtful decision.” A deliberate, thoughtful decision can only be made following careful site assessment and characterization.
Monitored natural recovery relies primarily on source control and ongoing sedimentation with increasingly clean sediments, as well as dispersion processes, to reduce surface sediment concentrations over time. Sediment remediation using MNR relies on multiple lines of evidence to reduce uncertainty and demonstrate long-term sediment and ecological recovery and risk reduction. Monitored natural recovery lines of evidence typically focus on demonstrating contaminant burial and surface sediment recovery, reduced contaminant mobility, chemical and biological transformation, and sediment (hence, contaminant) stability. Because MNR relies on relatively slower natural processes, it is likely to be most applicable to sites, or portions of sites, where risk is not immediate or substantial. Monitored natural recovery remedies rely on the following physical, chemical, and biological processes:
Contaminant burial and natural deposition with clean sediments over time;
Reduced contaminant mobility from sorption, precipitation, and other binding processes;
Chemical or biological transformations to less toxic forms; and
Dispersion of particle-bound contaminants that leads to contaminant concentration reductions.
KEY TOPICS FROM THE OPEN DIALOG
Seven primary topics were identified by attendees and were discussed with the panel members in open discussion. These topics are an excellent representation of the issues confronting industry, regulators, and the public regarding remedy effectiveness for contaminated sediments. The topics and the summarized comments of the attendees are presented in a logical sequence (not necessarily the sequence in which they were discussed). The topics are as follows:
Additional data collection,
Meaning of effectiveness,
Uncertainty versus success,
What cleanup levels are achievable,
Performance-based contracting, and
Need for more response tools
The panel emphasized the importance of completing source control before implementing response actions. One panel member cautioned that “you think you have good source control, and evidence pops up otherwise; it's not as easy as it sounds; it's very important to turn off the faucet.”
ADDITIONAL DATA COLLECTION
The consensus was that more data are often needed at contaminated sediment sites both to define risk and to design efficient remedies. However, given budget limitations, it was not clear who could pay for the additional data needed to reduce uncertainty. The panel had no obvious solutions in this regard. One panel member offered that “it's a question of paying up front or on the back end; monitoring is not cheap. If one wants to show that MNR is effective, then you're going to be monitoring for a long time. Fish are the focus because they are the risk driver, but fish need to be monitored for a long time.”
An attendee cautioned that obtaining more data is the right approach, but the data must be used effectively. One use is inputting the data to site-specific models. This not only allows informed decision making at a particular site but puts the data in context when using them in comparison to other sites.
MEANING OF EFFECTIVENESS
Attendees had questions regarding what is meant by remedy effectiveness, including
Is effectiveness about whether or not the remedy can meet the design criteria? If that is the definition, then techniques to improve performance must be identified. If US Environmental Protection Agency 5-y reviews demonstrate that we are not achieving what we need to achieve, then what are the uncertainties in the design criteria and what can we do to reduce these uncertainties?
Independent of what the target number is, what engineering techniques can be used to give decision makers a good probability of hitting the site-specific goal?
What defines success? When can an agency or responsible party claim that a remedy is successful? When can it be said that the project is done? Warnings about “never getting off the hook” are prevalent. We need to carefully define the criteria for a successful remedy, identify the uncertainty, and then live with the outcome. At some point, the project needs to be declared “done.”
Panel members offered the following thoughts:
We should be collecting the right data in future projects to better be able to answer these questions. It has been difficult to extract usable data from past projects to help plan future projects.
Goals need to be risk based. Goals were quite jumbled in historical projects and were not necessarily tied to acceptable risk reduction. A good conceptual site model—which identifies contaminant sources, transport mechanisms, exposure pathways, and receptors at the various levels of the food chain—is needed before designing an effective remedy.
As an example, suppose the cleanup goal is 10 ppm of polychlorinated biphenyls (PCBs) and the objective is recovery of the fish population (reduction of PCB concentrations in fish tissue). Two targets apply: (1) design the response action to reduce sediment concentrations below 10 ppm (past projects can provide information about what methods and means were used to achieve a similar goal) and (2) achieve the ultimate target, that of fish recovery. We need predictive tools and need to define and achieve a certain level in sediment to ultimately achieve the desired level in fish. We need to monitor after the response action to determine success and then report the predictive data and monitoring data so that they can be useful for other projects.
What we are trying to do is reduce risk. One panel member suggested fish consumption advisories should be reevaluated and made less restrictive, where defensible. Also, time is a problem. It can take a long time to see improvements in fish. Lake Hartwell in South Carolina (USA) is an example of a Superfund site where PCB contamination was addressed using MNR as the selected remedy. After 15 y, “the fish data are a mess”; interpretation of PCB levels in fish remain difficult because so many variables are present. In this regard, a different metric is needed. The majority of remedy goals are tied to reduction in fish. “Are we setting ourselves up to fail?”
Time is indeed a problem. These response actions are similar to grand experiments. Even by remediating hotspots, it may take a long time to see success. We may not have the answer as to outcome for a long time; we need to let natural recovery operate before significant improvements can be seen. As an analogy, consider the Clean Water Act (1972). Today, we can clearly see that the act has made a difference. It took small incremental steps, and we could not see the difference for the first 10 to 15 y. We may not know we are effective with sediment remediation for quite some time.
Uncertainties are embedded in the remedies, including uncertainties in implementation and in long-term ecological recovery. The latter is more elusive. In 50 y, we'll have hindsight.
UNCERTAINTY VERSUS SUCCESS
Regarding the issue of uncertainty, 1 attendee suggested that as the spatial scale of remediation increases, then the uncertainty with regard to potential success also increases. If uncertainty is plotted on a vertical axis and spatial scale on a horizontal axis, the resultant line has a positive slope. Another attendee reinforced that questions of temporal and spatial scale are quite important: “We need to get more comfortable talking about long temporal scales; we also need to consider spatial variability and lack of understanding about how the system works.” It was posited that the spatial scale may be a favorable variable in that it may be easier to understand the system on a larger rather than smaller scale: the variabilities may even out.
The panel noted that earlier sediment remediation projects focused on hotspots, with typically no measure of recovery. As the projects get bigger, they get analyzed more. Uncertainty in recovery is inevitable; to date, it has not been readily measured or predicted. As another example from the panel, consider a small lake with 1 contaminant of concern. For such a situation, it should be relatively easy to evaluate success. However, with a 200-mile stretch of river and risk from a stretch that is only 1 to 3 miles in length, remediation of that stretch may not make the fish any safer to eat. Cleaning up that relatively short stretch may make no measurable difference. The key is determining the correct scale for evaluating risk.
Another attendee voiced a similar concern by noting that the presumption favored to date, hotspot remediation, assumes that a hotspot is a major contributor to the impact; however, this may not necessarily be a correct assumption. It may be that diffuse low-level contamination is causing the impact.
One attendee emphasized that when the target cleanup concentration for a particular contaminant is very low, it is important to remember analytical uncertainty, the ability to see the difference between, say, 1 and 10 ppm. To illustrate, on one project the goal for the average concentration was 10 ppm and the result was 16 ppm. The attendee thought this result was quite good and within the analytical uncertainty; others, however, were upset. Analytical uncertainty must be kept in perspective.
WHAT CLEANUP LEVELS ARE ACHIEVABLE?
Opinions varied on the lowest cleanup levels that are achievable. The following opinions and examples were offered by attendees.
Many targeted cleanup levels are too low to be achieved. Is there a low end number that we know we can't achieve?
We cannot get to 0.25 ppm PCBs with dredging!
Agreed, that it may be hard to achieve 1 ppm PCBs (or other contaminant) without backfill. Then why not cap? Why pick 1 ppm as a goal if you know you can't achieve it?
We can't reliably predict postdredge concentrations. A current project in Puget Sound which is meeting 0.3 ppm PCBs, so you cannot say that we cannot hit a low target. What we need are better tools for predicting what level might be achievable. To say in advance that we cannot achieve a certain goal is a mistake.
“Don't forget about dry excavation. Get rid of the water! For example, at an Air Force Base cleanup (the Loring Air Force Base in Maine, USA), an old and large site, less than 1 ppm PCBs was achieved. You can't say it's not doable.”
At a site with a 6-mile river reach (Fox River in Wisconsin, USA), a target surface weighted average concentration of 0.25 ppm PCBs has been established, averaged over the whole reach. When the nondetects are added in, a lower surface weighted average concentration level can be achieved than might be expected. Another scenario is to use a sand cover. Design the cleanup scheme so that goals are achievable without numerous dredge passes. Try to make the design process more effective and successful, and have flexibility in the design approach. Be reasonable. Set up the approach for success, and build in flexibility.
The panel acknowledged there are differences of opinion on what the lowest cleanup levels are that are achievable on a sediment remediation project. To a large extent, the answer depends on the site and project goal. A panel member noted that of the 71 contaminated sediment remediation projects documented as completed in the MCSS database (2004), only one verified achieving a 1 ppm cleanup level (for PCBs at Loring AFB, a dry excavation project). The same panel member further noted that compliance to date was generally measured with discrete (point) sample results, whereas more recent projects are measuring compliance on a spatial averaging basis. In any event, achieving less than 1 ppm PCBs, or any other contaminant, is difficult and has not been demonstrated on dredging projects.
One member of the panel emphasized that achievable cleanup levels are site-specific, but felt that somewhere between 1 and 10 ppm was achievable. Several panel members emphasized that one of the difficulties is that the lowest achievable cleanup level is not predictable in advance of implementing a removal action. Several panel members suggested it may be more appropriate to move away from the concept of a lowest achievable cleanup level. For future dredging projects, for example, it may be more appropriate not to set an expectation for low residual levels, but, instead, consider limiting the goal of dredging to mass removal, then backfill to provide a clean surface as appropriate. Other possibilities include simply capping sediments with no dredging or using MNR for areas with low level, diffuse contamination.
An attendee raised the concept of performance-based contracting, whereby the contractor would be held to 2 metrics: short term (met specified cleanup levels) and long term (remedy was a success). The attendee noted that the US Department of Defense is moving toward performance-based contracting and the US Navy is considering making payments based on outcomes rather than model results. The attendee acknowledged that more information up front is needed to set the metrics.
Several panel members opined that this was not a viable concept for sediment response action contracts. The link between performance goals and recovery goals is currently not well known, and there are complications. A contract could specify dredging to a given depth, as well as how to do it precisely and efficiently or how to achieve a certain residuals level with a certain number of dredge passes. Even if the job is done well, the contractor cannot be held responsible for the effects of residual contamination on overall recovery. That responsibility lies with the designers and engineers. We as responsible parties or implementing agencies can specify what the contractor should accomplish, but they should not hold the contractor responsible for whether that result will lead to the intended recovery in the system.
NEED FOR MORE RESPONSE TOOLS
An attendee noted that more tools were needed to accomplish cleanups than just dredging, capping, MNR, or a hybrid. What other tools are available?
The panel emphasized that it “pretty much came down to these three tools,” or a combination of the three. A few variations are being explored, such as carbon amendments for caps and thin-layer capping. An attendee noted that at the Passaic River and Newark Bay (USA) demonstration project, in situ stabilization using grout would be implemented.
SUMMARY AND OBSERVATIONS
The Remedy Effectiveness Panel identified, and discussed with conference attendees, key issues associated with the 3 generic contaminated sediment response actions: environmental dredging, capping, and MNR. The panel emphasized that sediment risk management strategies (response actions) should seek to balance (1) contaminant risk to human health and the environment, (2) remedial costs, and (3) risks associated with remedy implementation.
Of the 3 response actions, dredging typically is the most expensive but results in the greatest mass removal from the aquatic environment. Ecological benefits of dredging may be compromised by leaving behind residual contaminated sediments on the dredged surface and by downstream transport of contaminated sediments resuspended during dredging. Capping effectively eliminates surficial contaminant concentrations by isolating contaminated sediments, thereby eliminating contaminant availability to benthic and aquatic organisms. Capping is typically much less expensive than is dredging, but because contaminants are left in place, caps require long-term monitoring and have persistent risks of contaminant breakthrough. MNR is most effective for low-risk sites or sectors where surface sediment contaminant concentrations approach surface sediment cleanup goals. MNR is the least expensive of the 3 response actions but is likely to require extensive and prolonged long-term monitoring. Some sites use combinations of the 3 response actions. Ultimately, selection of a response action must balance the advantages, limitations, uncertainties, and costs of each.
The open dialog between the panel and conference attendees resulted in 7 key topics identified that are an excellent representation of the issues confronting industry, regulators, and the public regarding remedy effectiveness for contaminated sediments. These key topics and important observations expressed for each follow:
Source control—Source control is required before implementing remediation.
Additional data collection—More data are needed before selecting a response action, particularly at large complex sites; cost is often limiting.
Meaning of effectiveness—Confusion reigns as to whether success is meeting the cleanup goal or the long-term recovery goal; how can results be better predicted before remedy implementation and when is a project “done”? The panel emphasized (1) collection of the right data is crucial; (2) goals need to be risk-based, and (3) lengthy times for recovery are a problem. We may not know the success of implemented remedies for a long time; perhaps a shorter-term goal or relaxed fish advisories (if technically defensible) are warranted to allow projects to be “done.”
Uncertainty versus success—As spatial scale increases, the uncertainty in remedial outcome increases; small, focused cleanups may produce no measurable benefits. It is important to determine the correct scale when evaluating risk and be cognizant of analytical uncertainty when setting cleanup goals or evaluating chemical residuals.
Achievable cleanup levels—No consensus was reached on lowest achievable cleanup levels; levels are very site specific and difficult to predict in advance; for example, only 1 project among the 71 major sediment remediation projects implemented and documented in the USA has achieved a contaminant cleanup level as low as 1 ppm (for PCBs). A Puget Sound (WA, USA) project reportedly is achieving 0.3 ppm PCBs. The panel noted that compliance to date was generally measured with discrete (point) sample results, whereas more recent projects are measuring compliance on a spatial averaging basis. The panel suggested it may be more appropriate to move away from the concept of a lowest achievable cleanup level; for future dredging projects.
Performance-based contracting—The panel's opinion is that a remediation contractor should not be held responsible for achieving both cleanup goals and long-term recovery goals. Performance goals and recovery goals are currently not well linked or well known. Even if the cleanup job is done well, the contractor cannot be held responsible for the subsequent effects of residual contamination on overall recovery. That responsibility lies with the designers and engineers.
Need for more response tools—No other response tools are currently available except the 3 generic response actions of dredging, capping, and MNR; some variations are being explored, such as carbon amendments for caps and thin-layer capping, as well as in situ grouting.