SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
SEDIMENTS REMEDIATION ACTION TEAM MEETING
West Coast Grand Hotel at Fifth Avenue
Seattle, Washington
October 29-30, 2002

INTRODUCTION

Nancy Grosso, co-chair of the Remediation Technologies Development Forum's (RTDF's) Sediments Remediation Action Team, welcomed meeting attendees (see Attachment A) and reviewed the conference agenda. She stated that the purpose of the meeting was to update members about ongoing RTDF projects and general business items and discuss the role of ground water, surface water, and sediment interactions in the context of assessment and remediation.

GENERAL SESSION

Update on Anacostia River—Comparative Validation of Innovative Capping Technologies
Danny Reible, Hazardous Substance Research Center-South and Southwest (HSRC-SSW),
Louisiana State University

Danny Reible discussed the latest efforts by the Anacostia Watershed Toxics Alliance (AWTA) to address sediment contamination at the Anacostia River in partnership with potentially responsible parties (PRPs), regulatory agencies, and community groups. The project he described is an examination of the comparative effectiveness of five traditional and innovative capping methods under realistic, well documented, in situ conditions at contaminated sediment sites along the river. The goal of the capping project is to find a cap that will be effective at reducing flux between the benthic community and contaminants, encourage degradative reactivity, and/or enhance the riverbed's sorptive capabilities in order to reduce contaminant levels. The capping technologies are undergoing bench-scale testing at present and will be ready for in situ application in the spring of 2003. Cap technologies include:

• Aquablok for control of seepage and advective contaminant transport.
  
• Zero-valent iron to encourage dechlorination and metal reduction.
  
• Phosphate mineral (Apatite) to encourage sorption and reaction of metals.
  
• BionSoil to encourage degradation of organic contaminants.
  
• Natural organic sorbent to encourage sorption-related retardation (reduction in advective-diffusive transport).

Reible also indicated that two phases of site characterization will be conducted. These assessment efforts, which will be conducted in fall 2002 and spring 2003, will be performed to establish a contamination baseline at the demonstration areas. Characterization activities will focus on defining physical, chemical, and biological parameters through a variety of analytical methods.

Reible noted that additional information related to this project can be found at http://www.hsrc-ssw.org. His presentation is included in this report as Attachment B.

Monitored Natural Recovery Criteria and Case Study
John Davis, The Dow Chemical Company
Clay Patmont, Anchor Environmental
Tim Dekker, Limno-Tech, Inc.

John Davis, Clay Patmont, and Tim Dekker provided an update on the Monitored Natural Recovery (MNR) Subgroup's activities since the last (February 2002) RTDF meeting. They reiterated the goals of the subgroup: (1) provide guidance on the technical confirmation of MNR for contaminated sediments by developing a framework for evaluation, identification, and publication of case history examples, and (2) apply the framework to assess the effectiveness of sediment MNR in reducing risk to human health and the environment.

Davis outlined five elements that were selected to assess the presence of MNR. He then explained that the subgroup had refined its framework and developed a sediment MNR assessment template while identifying additional case history examples. The case history sites now include Bellingham Bay, Washington; Eagle Harbor, Washington; Commencement Bay (Sitcum Waterway under-pier), Washington; Spokane River/Lake Coeur d'Alene, Washington/Idaho; Palos Verdes Shelf, California; Lake Hartwell, South Carolina; Morrow Lake, Michigan; and James River, Virginia. Patmont presented detailed information about the applicability of the Bellingham Bay case history to the MNR framework. Dekker outlined the projected path forward for the project: finalizing the MNR evaluation framework, completing the case history template that is now in draft form, documenting a representative set of case histories, developing framework and case study presentations and publications, and developing a Web site that will eventually house the results of the subgroup's efforts. The MNR presentation is provided as Attachment C.

A question was raised about the appropriateness of using Bellingham Bay as a case history example given the fact that there is concern that the contamination (mercury) at this site migrated (via the atmosphere and other routes) to other receptors (e.g., fish) rather than being truly attenuated. In response, Davis, Patmont, and Dekker all emphasized their desire to be informed of additional case histories that might be applicable to the project.

Technical Update: Bioavailability Control and In Situ Stabilization of Contaminated Sediments Using Carbon Sorbents
Upal Ghosh, University of Maryland

Upal Ghosh presented research he has conducted on the propensity for polychlorinated biphenyls (PCBs) and other contaminants to bind to sediment particle types. He proposed a new concept for sediment management based on in situ control of PCB bioavailability through addition of activated carbon. Ghosh said that he has demonstrated that 60 to 85 percent of polycyclic aromatic hydrocarbons (PAHs) and PCBs are found on low-density black carbonaceous sediment particles, despite the fact that 90 to 95 percent of sediment mass lies in heavy particles. His laboratory research has indicated that when PAHs and PCBs are associated with black carbonaceous particles, they are strongly bound and are not bioavailable. To confirm this finding in a benthic system, Ghosh and his team conducted laboratory experiments with amphipods, worms, and clams to determine their ability to bioaccumulate PCBs in carbon amended sediments. The results revealed that these biota bioaccumulated 70 percent less PCBs in carbon amended sediments than in unamended sediments (see Ghosh's presentation, Attachment D); the results imply that if one can find a way to add activated carbon to PCB- or PAH-contaminated sediments one can dramatically reduce the ecological availability of these chemicals. Ghosh said that efforts will be made to determine how to achieve this in a field situation.

When asked about the projected cost competitiveness of using activated carbon as a commercial technology, Ghosh replied that activated carbon is more expensive than other capping materials, but that there is potential in the availability of re-activated carbon, which costs only 35 to 40 cents per pound amounting to a remediation cost of 10 dollars per cubic meter of sediment. He has not determined the cost of carbon deployment, this will be part of future research efforts. Ghosh was also asked about the implications of his research for sites that may naturally have more coal-based carbon in sediments. His research, he responded, shows that one can make some a priori predictions about the bioavailability of PAHs or PCBs at a site if one knows carbon concentrations in sediment and the nature of carbon at the site.

Update on the Sediment Remediation Decision Framework
Nancy Grosso, DuPont Corporate Remediation

Nancy Grosso provided an update on efforts to produce joint-agency guidance—promulgated by such entities as the U.S. Environmental Protection Agency (EPA), the U.S. Army Corps of Engineers, the National Oceanic and Atmospheric Administration, the U.S. Navy, and the RTDF—for evaluating and managing contaminated sediment sites. The guidance is being developed as a detailed framework. It focuses on the uniqueness of risk and risk management of sediment sites, defines efficient sequencing of efforts, and provides a management "roadmap" for readers seeking guidance. It will be a Web-based document, she said, with 50 to 60 pages of text. Links will be included to direct users to more detailed information and guidance. Grosso noted that individual "chapters" have already been written and are being compiled/edited by Mike Palermo for an approximate delivery date of mid-November. Comments on specific chapters will be welcomed once the most recent version is available. Grosso's presentation is included as Attachment E.

GROUND-WATER–SURFACE-WATER WORKSHOP
    View Groundwater-Surface Water Workshop - Summary of Discussion (439K/21pp/PDF)

Introductions and Focus of the Workshop
Nancy Grosso, DuPont Corporate Remediation

Grosso summarized the focus and goals of the workshop; her presentation is included as Attachment F. She said that the workshop would focus on ground-water (GW), surface-water (SW), and sediment interaction in the transition zone. The goal of the workshop, she said, is to understand important aspects of this interaction (making it possible to develop conceptual models that can be used for assessment and remediation), to start developing pragmatic tiered methods for assessment, to identify areas of consensus among disciplines, and to consider innovative remedial strategies that might effectively deal with impact in the transition zone.

The workshop would focus, Grosso said, on addressing the effects of contaminated GW plumes, or contaminated SW, on "clean" sediments. (For contaminated sediments, an outside-in approach may be more appropriate: "What are the other potential inputs (contaminant sources) to the system?")

Grosso posed the following questions to workshop participants:

• Conceptual models:
  
  

  —	Are there one or more conceptual models that can be used to test hypotheses about GW on the environment?
  —	What are the significant physical, chemical, and biological parameters that should be considered?
  —	What role does natural attenuation play in GW–SW interaction?
  —	Are there environments in which GW–SW interaction is crucial? Environments in which it is insignificant?

• Assessment:
  
  

  —	Is there a hierarchy of screening methods to streamline the assessment process?
  —	Can a plume de minimis zone be defined in relation to the SW size/habitat?
  —	How important are GW "hot spots"?
  —	How do you separate GW impacts from other sources?
  —	Can you quickly determine the relative significance of GW compared to other watershed contributions?

Presentations were organized into three panels based on environmental setting: (1) estuarine and marine, (2) riverine, and (3) lacustrine.

Overview of Processes

Overview of GW–SW Conceptual Models—A Conceptual Model for GW, SW, Sediment, and Organism Dynamics
Bruce Duncan, U.S. EPA

Bruce Duncan began by mentioning how important it is to develop good, representative conceptual site models that takes GW–SW interactions into consideration. He presented and contrasted several types of conceptual models, each of which was progressively more dynamic than the previous. He noted that many traditional models do not show pathways of GW to sediment, sediment to GW, or sediment to SW. Nor do they show the dynamics between GW and SW. Duncan proposed developing an integrative (i.e., ecological and hydrological), flexible (i.e., site-specific depending on interactions between GW, SW, sediment, biota), dynamic (e.g., accounting for movement of SW up and down, changes in GW levels) conceptual model that better represents actual sites. He suggested that attendees:

• Continue to improve conceptual models (to "ease the pain" of understanding complex dynamic systems).
  
• Enhance "box" models to include the implied dynamics.
  
• Tailor the complexity of conceptual site models to the sites they represent.

Duncan's presentation is included as Attachment G.

Overview of the Hydrogeology of GW–SW Interactions
Brewster Conant, University of Waterloo

Brewster Conant defined the commonly used terms “hyporheic zone,” “GW–SW interface,” and “transition zone,” concluding that “transition zone” is most appropriate in discussions of the hydrogeology of GW–SW interactions because interactions are not just with rivers, but also with lakes and tidally influenced water bodies. He pointed out some of the features that make the transition zone unique: complex and dynamic hydrology, geology, and biogeochemistry; spatial and temporal variability that make it difficult to determine exact GW flow paths; and the potential for the zone to change the shapes, sizes, and compositions of plumes. Conant discussed the many different types of GW–SW interface interactions, and presented a conceptual model that categorized GW discharges into 5 main types of behavior. He said that the plume, geology, hydraulics, reactions, and time all affect fate and transport of contaminants within the transition zone. He discussed monitoring goals and objectives and tackled the question of how much data is enough. He suggested that monitoring is dictated by the type of questions one is trying to answer and gave a list of 4 questions that were typical ones to ask in these investigations. Finally, he suggested that an understanding of transition zones could have implications for site monitoring, investigations, and remediation. An edited version of the slides that Conant presented are included as Attachment H.

Overview—Biochemical Processes and Contaminant Fate in the GW–SW Interface
E. Erin Mack, DuPont Corporate Remediation

Erin Mack opened her presentation by saying that the GW–SW interface (GWSWI) is the subsurface zone in which the water shares characteristics of the GW and the SW. These characteristics could include redox, water chemistry, biological populations (e.g., microbes, benthic organisms), and contaminant profile. The GWSWI is a zone of gradients and transitions in which environmental, abiotic, and biotic processes can affect contaminant fate and transport. Mack outlined the factors upon which contaminant fate and transport in the GWSWI depend, and detailed how each of the preceding processes may act to change fate and transport. She provided three examples of GWSWIs and concluded that:

• Abiotic and biotic processes in the GWSWI environment can significantly change the chemistry, toxicity, and/or mobility of a contaminant.
  
• Discharge of GW from a contaminant plume into SW does not necessarily translate to discharge of plume contaminants.
  
• Characterization of the GWSWI environment is necessary for accurate determination of risk to human and environmental health and selection of best strategies for management of GW contaminants.

Mack's presentation is provided as Attachment I.

Panel 1: Estuarine and Marine Settings

Coastal Contamination Migration Monitoring
D. Bart Chadwick, U.S. Navy, Space and Naval Warfare Systems Command Center

Bart Chadwick began by showing the results of a recent Navy review, noting that it indicates that there is potential for GW–SW interaction at a large number of coastal landfills and hazardous waste sites. He said that the Navy's typical protocol for addressing coastal contaminant sites with GW–SW interaction involves: (1) performing flow and contaminant detection assessments, (2) evaluating in situ bioassay exposures, and (3) modeling for areas with tidal influence. After evaluating available flow detection and contaminant detection technologies, Chadwick said, he decided to develop a flexible, multi-sensor water sampling probe for screening and mapping GW plumes at the SW interface. The sampling probe, called the Trident:

• Tests for conductivity by detecting contrast in salinity and/or clay content in unconsolidated sediments.
  
• Tests for temperature, detecting GW by thermal contrast with SW.
  
• Uses a pore-water sampler that allows contaminant characterization and detection of other groundwater-specific tracers.

Chadwick provided details about the Trident, including information about field deployment tests, user interface, the conductivity configuration, and laboratory sediment testing. Chadwick said that he has also developed an UltraSeep meter—a modular, state-of-the-art seepage meter that measures GW and contaminant discharges at the SW interface directly using:

• An ultrasonic flowmeter, which provides direct measurement of groundwater flow.
  
• A water sampler—a low-flow peristaltic pump with a sample selector valve and a sample-bag array.
  
• On-board sensors for temperature and conductivity.
  
• A controller that stores data and controls sampling events.

Through the laboratory and field deployment tests, Chadwick demonstrated that the Trident probe and UltraSeep meter were successful for rapid screening of GWSWI sites under a variety of conditions and at a range of sites. Chadwick concluded that, together with traditional shoreside sampling and modeling, these tools can improve assessments of GWSWIs. He closed by describing some studies that will be conducted to further explore the use of sampling tools. For example, an effort will be initiated to determine whether passive diffusion samplers can be used as an alternative sampling device in areas where pore-water samplers are not effective (i.e., areas with high proportions of fine-grained sediments). In addition, in situ bioassessment techniques are being developed to support risk assessment at GWSWI sites. Chadwick's presentation is included as Attachment J.

Exit Concentration in Ground Water Discharging Into Tidal Waters
Farukh Mohsen, Gannett Fleming

Farukh Mohsen discussed the process and results of developing a mathematical model to predict the exit concentration of GW discharging into tidal waters. He began by saying that many industrial facilities are located close to tidal waters and that one must understand exit concentrations of contaminants before one can understand/predict contaminant concentrations in the mixing (GW–SW interaction) zone. He said that he has developed a predictive equation based on the following input parameters: head, time, tidal period, storage coefficient, transmissivity, flow velocity, and effective porosity. He ran versions of the model that incorporated tidal and non-tidal variations into the parameters, and tested the model against field measurements at a sample site. The results showed that measured concentrations compared significantly better to numerical predictions when tidal parameters were included in the model. Mohsen concluded that tides generally lower exit concentrations and that transport in a tidal aquifer is sensitive to regional gradient, hydraulic conductivity, and tidal amplitude. Mohsen's presentation is included as Attachment K.

Natural Attenuation of Chlorinated Aliphatics in Wetlands: Linking Hydrology, Geochemistry, and Microbiology
Michelle Lorah, U.S. Geological Survey

Michelle Lorah presented information about her experience evaluating natural attenuation in a wetland system that was contaminated with chlorinated solvents. She began by noting that, in the context of sediments remediation, there are several challenges particular to wetlands: sensitive ecosystems, inaccessibility and permitting, large number of fate processes, steep biogeochemical gradients, high spatial variability, high seasonal variability, and high microbial diversity.

Knowing that chlorinated solvents reductively dechlorinate in anaerobic environments, Lorah installed nested piezometers at a wetlands site at Aberdeen Proving Ground and determined that anaerobic degradation was occurring. Recognizing that near land surface, it is difficult to determine how much degradation is from anaerobic degradation or other factors, Lorah used passive diffusion samplers (peepers) and a novel scanning voltametric analytical system with microelectrodes to identify wetland geochemistry. She found high degrees of spatial and vertical variability in the wetlands; at one transect there were significant differences in degradation products between 1 foot and a few inches below the peat. She also found prominent seasonal variability in degradation products between summer and winter. Geochemical analysis revealed that volatile organic compounds (VOCs) were highest in summer, cyclic changes seemed related to changing water levels, and natural attenuation was still efficient throughout the year. But different sediments produced different daughter compounds. Lorah looked into possible microbial explanations for the variability, including the potential presence of iron-reducing bacteria, spatial or temporal heterogeneity of microbial diversity, and a possible vegetative growth cycle influence on microbial biomass. None of the examined factors explained the seasonal variations, leading Lorah to conclude that a complicated interlinking between chemistry, hydrology, and climate was at work.

Despite the fact that, in all locations, VOCs were degrading before reaching land surface, they were still being detected in SW. Lorah discussed work performed to investigate the presence of seeps recontaminating the SW. She used aerial and ground thermal infrared imaging (TIR) surveys to identify seeps while documenting the effectiveness of high-resolution TIR imaging in seep delineation. To confirm the TIR results, she used passive diffusion samplers to collect representative samples of shallow GW and co-located SW samples, which she analyzed for VOCs and methane. Lorah found that many of the seeps were contaminated. She concluded that wetlands processes are very multidisciplinary: they must be thought of in the context of chemistry, hydrology, vegetation, and microbiology. Her presentation is included as Attachment L.

Panel 2: Riverine Settings

GW–SW Interactions Can Affect the Bioavailability of Sediment-Associated Contaminants to Benthic Invertebrates
Marc Greenberg, U.S. EPA

Marc Greenberg conducted an integrated in situ assessment to determine how GW–SW interactions affect contaminant bioavailability. He began by conducting hydrologic measurements at river sites where, in numerous places, concentrations of metals and VOCs exceed GW and sediment quality criteria. He used two different types of mesh-screened in situ chambers—one that is placed on top of the sediment bed to capture exposures at the sediment/water interface with a collection of organisms, and one that is inserted 2 centimeters into the sediment bed and allows for direct interaction of organisms with the sediment and water—to assess the effects of GW, SW, and sediment on benthic invertebrates. Greenberg sampled pore water and hydraulic heads with mini-piezometers and also sampled the interior exposure chamber water using syringes. He concluded that mini-piezometer data provided a unique in situ characterization approach to documenting GW–SW conditions, that data from mini-piezometers improved interpretation of exposure-effects relationships, that downwelling was shown to reduce exposure in one system while upwelling conditions were shown to increase exposure and effects when sediments and GW were contaminated, and that integrated approaches are essential in a holistic assessment of sediment toxicity.

Greenberg then used these data to develop a bioaccumulation model that predicts how much contaminant will accumulate in the tissues of benthic species that are exposed to contaminated sediments. He evaluated the model by comparing his predictions to in situ bioaccumulation data that had been collected at sites containing contaminated sediments. The results showed that the model adequately reproduced the simulated steady-state concentrations observed in laboratory sediment exposures and predicted bioaccumulation data that were within an order of magnitude of field data, and in many cases were within a factor of 4 of the data. Qualitative descriptions of upwelling and downwelling in the model—represented by a term describing the fractional flow of pore water—indicated they are important determining factors in bioavailability. The model simulations also suggested that in situ rates of feeding should be measured, as feeding rate was a sensitive parameter. This report includes Greenberg's presentation as Attachment M. For more information, see Greenberg, M.S., et al. 2002. Environ. Toxicol. Chem. 21(2):289-297.

GW–SW Case Studies in New Jersey and Other Locations
John Pardue, Louisiana State University

John Pardue said that identifying the location and nature of contaminant discharges is a common problem at SW and wetland sites. He noted that the scale of changes in biogeochemical conditions changes rapidly as the GWSWI is approached. Among the traditional tools for identifying GW discharge zones are nested piezometers, dialysis samplers, passive vapor samplers, seepage meters, and thermal imagery. But for a site where there was circumstantial evidence that an 800-foot creek was the discharge point for VOCs—despite the fact that all stream grab samples had been non-detect for VOCs—dialysis samplers were selected as a tool for demonstrating that contaminants were not affecting the stream. In order to determine the effect of the plume on the creek, Pardue and colleagues used dialysis samplers for testing so that measured contaminant concentrations could be interpreted as pore-water samples. He built dialysis samplers for the Marvin Jonas Transfer Site in New Jersey, which he inserted most of the way into the stream bed and retrieved 2 weeks later for analysis. Through this assessment, he discovered two plumes downstream of, and in addition to, the expected plume location. Pardue noted that regulators were pleased with the reliability of the assessment, but are now considering how to proceed with the site: they are comparing contaminant concentrations in the de minimis zone to SW quality criteria.

Pardue proceeded to discuss the possibility of treating the GWSWI. He suggested that a constructed wetland could be used to treat both chlorinated and non-chlorinated VOCs, maximizing biodegradation and minimizing volatilization year-round. A wetland could be constructed as an alternative discharge point for a GW plume within the site boundary, either passively intercepting the plume or serving as a component of a pump-and-treat system. A site adjacent to Chesapeake Bay with unexploded ordnance concerns is being considered as a candidate for this type of remediation. Pardue's presentation is included as Attachment N.

The Potential for a GW Plume To Contaminate a River: the Role of the Streambed and Near-River Zone
Brewster Conant, University of Waterloo

Brewster Conant discussed a multi-stage approach that was used to investigate the complexity of GWSWI at a site in Angus, Ontario, Canada, where a drycleaning business had released tetrachloroethylene (PCE). The release resulted, Conant said, in a PCE plume that stretched approximately 195 meters and impacted the nearby Pine River. Conant performed soil coring to understand the geologic composition of the area in and surrounding the plume and used ground-penetrating radar (GPR) to further understand the geological conditions under the river itself. The water flux through the streambed was calculated using a new method that involves measuring streambed temperatures and hydraulic information obtained from mini-piezometers (as described by Conant in a paper in submission to Ground Water). A Waterloo profiler, a mini-profiler, and multilevel samplers were used to determine the distribution of PCE and its 7 anaerobic degradation products in the riverbed, in both plan view and in cross-sectional transects.

Conant was able to develop a conceptual model/map of the below-surface geology, plume composition, and areas of recharge and discharge into and out of the river. He then sought to quantify the PCE and degradation products and redox conditions within the plume and the riverbed. He showed that very different patterns of PCE mass flux and cis-dichloroethylene (cDCE) mass flux occurred in the streambed. Despite the number of different degradation products found in the streambed, only PCE concentrations were found in SW. Putting all these data together, Conant was able to determine that the GWSWI system is heterogeneous on a fine scale; the transition zone changed the size, shape and composition of the plume; the PCE contamination was transformed, not mineralized; and the mass discharge was low overall but high enough locally to result in SW contamination.

Conant found that the flux-based conceptual model provided an appropriate framework for interpreting water quality, which could be of use when designing investigations or interpreting data from other sites. Conant pointed out that contamination in SW was rare, but consistent with PCE mass discharge patterns in the streambed. He expressed concern that streambed interstitial water and sediment concentrations were rather high and represent a potential hazard to aquatic life. He also noted that contaminants sorbed to the sediments can represent a long-term source of contamination even after the source of the GW plume is remediated. This research was the focus of Conant’s dissertation, titled “A PCE plume discharging to a river: Investigations of flux, geochemistry, and biodegradation in the streambed,” University of Waterloo, Waterloo, Ontario, Canada, 2001. Publications regarding this work are in preparation.

Panel 1 and 2 Discussion

Grosso asked the group what they would consider appropriate "next steps" if faced with the Ontario site that Conant described above. She commented that with GW–SW interaction sites, one is often determining the ways in which the water is affecting upland areas—with this site, one wants to understand the impact of an upland GW plume on an SW body. Peter Adriaens noted that he is confronting a similar site (contaminated with chlorinated solvents) at Lake Huron. The Department of Environmental Quality is advising his group to put a reactive barrier in place at the GWSWI. For this particular site, some biological impacts (e.g., deformed frogs) have already been documented. Conant pointed out that his group does not know what the ecological impacts are for the Ontario site. He said that he would therefore like to use in situ samplers (like the ones Greenberg described) to determine whether or not benthic systems are being impacted by the release of the contaminant plume in the river. Alternatively, he suggested, one could focus on remediation for the hot spots only (e.g., through selective pump-and-treat), thereby remedying the problem on a refined scale. Conant did comment that the presence of seeps at the ground surface does make this particular site a bit different from some others, in that humans (especially children at the river bank) may access the site and be exposed to the contamination.

Duncan noted that he would like, or would have liked, to see an ecologist work with Conant on the Ontario site. He concurred with Conant's point that the impacts have not really been defined yet. If there are ecological endpoints (e.g., a particular spawning habitat near upwelling areas, benthic macroinvertebrates), one needs to prioritize them to see what is of most concern. Duncan also pointed out that the depth at which one samples could potentially have a dramatic effect on the results of the sampling. Given Greenberg's in situ experiments in which SW samples were taken only 4 to 6 centimeters above sediment, Duncan said, it seems that where one takes GW or sediment samples could be critically important.

Duncan asked Conant to elaborate on using a "tiered approach" and explain how he would do things differently if he were to assess the site again. Conant replied that the questions one is trying to answer (such as those listed in his first talk) are critical for determining the approach that should be used. If the site assessment is meant to determine the distribution of contaminants, one can begin with diffusion samplers to find hot spots. Following this, one could use temperature measurements to find areas of high or low discharge and start getting at the question of high contaminant mass discharge locations. Installation of piezometers could then be targeted to investigate locations of thermal anomalies. Finally, one could use slug tests to understand the hydraulic conductivity at the site and come up with quantitative estimate of discharge. With those pieces of information, one would have a minimal amount of quantitative data, but a good sense of what was going on at the site. That would represent the first stage in the tiered approach: once completed, one could probably approach a regulator with a solid grounding on the spatial resolution, knowing where the variability is in the system and where contaminants are coming out. From there, it might be possible to target a subset of the different types of flow and contaminant conditions observed at the site, allowing one to zero in on the issues that require additional exploration quickly. Conant pointed out that it would rarely be necessary to conduct all of the different field measurements and analyses that he performed at the Ontario site. The tiered approach centers around the fact that not all assessments have to be done at the beginning on such a fine scale.

Duncan also brought up what he saw as a data gap in researchers' understanding of how volatiles and solvents move through aqueous systems. He said that there is substantial knowledge about the processes that occur in zones where microorganisms actively break down contaminants. We really do not know much, however, Duncan said, about interactions between benthic communities and microbial communities. He asked attendees if anyone knew of research being done to understand these microscale communities.

Recap of Day 1 of the GW–SW Workshop
Nancy Grosso, DuPont Corporate Remediation

Grosso provided a comprehensive summary of the major findings and conclusions discussed during the first day of the workshop. She pointed out that the major take-home message had been: conceptualization of the site-specific problem(s) is critical. In order to effectively and efficiently formulate a concept of the site, one needs to determine the following:

• What is the question?
  
• What/whom are we trying to protect?
  
• What is the appropriate scale of evaluation (meter scale, river reach, watershed)?

Throughout the first day of presentations, presenters and attendees suggested considering the following when trying to formulate a site problem:

• Is the habitat unique?
  
• What does your habitat map show you about areas to prioritize for evaluation?
  
• Are contaminants persistent, bioaccumulative, and toxic (PBT) chemicals?
  
• What are other potential sources in addition to GW?
  
• What is the incremental risk that the GW plume adds to sediment contamination?
  
• Where will resources be best spent to improve the quality of the system?
  
• What is the best assessment approach: broad-scale to small-scale or small-scale with expansion?

Grosso's presentation is included as Attachment O.

Panel 3: Lacustrine Setting and Characterization Tools

Natural Attenuation at a High Energy GWSWI Site—St. Joseph, Michigan
Peter Adriaens, University of Michigan

Adriaens discussed research that he performed at a Superfund site. The goal, he said, was to determine if natural biogeochemical processes could be relied upon to mitigate vinyl chloride fluxes into nearby Lake Michigan. Adriaens said that he and his team used geo-probe materials to conduct GW analyses on- and off-shore. Then they used GPR along land and water transects to detect subsurface features (e.g., the water table, the presence of minerals that exhibit magnetic properties, the depth of wave attenuation). From the GPR results, Adriaens team was able to determine that magnetite minerals below the surface were dechlorinating vinyl chloride through microbial iron reduction. They developed a conceptual model of the plume, showing that it was gradually decreasing in depth as it proceeded further under Lake Michigan. Hypothesizing that this was due to geochemical changes, they proceeded to determine the maximum possible effect of wave action on oxygen and redox profiles; this turned out to be minimal. Noting temporal variations in contaminant levels, Adriaens then performed laboratory experiments to find the extent to which biological activity influenced the reduction of vinyl chloride in the plume. Methanotroph activity on vinyl chloride was responsible for only 0.1 percent of the mere 7 percent of oxidative attenuation that was occurring. Adriaens said that the final conclusions about the site and its natural attenuation potential were:

• The contaminant plume has migrated to the lake and discharges into Lake Michigan between the shoreline and 100 meters from shore.
  
• SW activity oxidizes the shallow and deep portions of the GWSWI, but this effect is greatest in the shallow zone.
  
• Contaminants are transformed via oxidative processes in the shallow zone and reductive processes in the deep zone.
  
• Aerobic natural attenuation is insufficient to prevent the flow of vinyl chloride into Lake Michigan.

Adriaens' presentation is included as Attachment P.

Application of Integrated Models To Evaluate Sediment Cap Effectiveness
Bob Lien, U.S. EPA

Bob Lien described a modeling approach that he has used to evaluate sediment cap effectiveness. He said that GFLOW 2000, an analytic element model, is used as part of the process. This model, he said, solves conjunctive steady-state GW and SW flow; allows displays of binary base maps for streams, lakes, roads, legal boundaries, etc.; and represents streams and lakes with strings of line-sinks, assigning each a head that is set equal to the water level in the stream or lake. Lien explained the process of inputting GFLOW 2000 step by step. (See his presentation, Attachment Q.) After inputting the GFLOW 2000 model, Lien said, he integrated a one-dimensional (1-D) fate and transport model that considers 1-D advection and dispersion through the liquid phase, sorption to the solid phase, and biological degradation in order to describe the fate and transport of a pollutant in contaminated sediment overlain by a clean cap. The 1-D fate and transport model uses a variety of baseline assumptions about water flow, velocity, advection, dispersion, transformations, and boundary conditions to predict a spatial concentration profile over a given number of years. By integrating the two models, Lien said, he demonstrated the sensitivity of GW discharge in sediment cap performance, illustrated the need to carefully monitor the GW–SW interaction at capping sites, and showed that one needs knowledge of regional hydrologic interactions to evaluate local sediment cap effectiveness correctly.

Panel 3 Discussion

Grosso thanked all the speakers for delivering high-quality and relevant presentations. She opened the floor for discussion about the GWSWI and its challenges. Discussions revolved around the following topics: (1) applying the tools and approaches that were presented during the workshop at larger-scale sites, (2) choosing the scale that should guide site assessment and characterization, and (3) examining GW contributions to sediment sites.

Patmont said that the workshop had done a good job evaluating GW–SW interactions from the perspective of examining an upland site and determining its impacts on the water. But at many sites, he noted, contaminants are in the river and subsequent upland contamination is a concern. He suggested that lessons learned from the latter perspective were really the question of the day. For example, if one is addressing a large urban river with a large navigational channel and trying to characterize sediment quality, numerous questions come to mind: What level of effort is appropriate? What tools should one use? What is an adequate level of characterization? Applicability of the methods discussed during the workshop on a grander scale is a large task that needs addressing.

Duncan replied by reiterating the importance of developing a conceptual model that integrates what is known about receptors (including how they are realistically going to be using the site) early in the process. He advocated having the type(s) of organism present at a site guide the scale of the assessment, suggesting that identifying receptors first is very different from trying to characterize an entire site on a large scale, then minimizing one's scope to the areas of concern. He said it is not efficient to send hydrogeologists to characterize a site completely, then have ecologists conduct the ecological assessment. It is enough to determine what is happening on a broader scale, focusing on general areas or zones of the SW body (e.g., a recharge zone) and also identifying areas of heterogeneity that might be of concern.

Someone commented that when one really gets involved in any site, it ends up being the little pockets of detail that really tell the story. As much as that may be acceptable for a small site, it is a daunting prospect for a larger-scale site. Duncan suggested approaching characterization by deciding what you are willing to miss, coupled with what you do know (e.g., source, well data already in hand), then using the hierarchy of available tools that will be cost- and ecologically effective for characterization. One person questioned whether a typical sampling density of a few hundred feet, generally considered acceptable in Washington, is still appropriate for characterization using this approach. Duncan replied that the site team needs to put the ecological habitat map and the hydrogeologic map together to determine where there are exposures of concern. This will inform their choice of areas on which to focus.

A manager of a large, complex sediment contamination site, where there are lots of inputs, outfalls, historical spills, etc., asked about the incremental risk of GW contamination to sediments. He suggested that the question of whether or not it is worth going after or cleaning up GW had not been answered. Duncan replied that it really comes down to the cost of assessment versus the cost of remediation. He advocates sampling concentrations at receptor exposure points (which, in fact, is fairly inexpensive) rather than using models. Typical models extrapolate well data to a river, attempting to incorporate degradation and other factors. Because of models' inherent uncertainties, regulatory agencies are likely to object to them if they are based on insufficiently conservative assumptions. This makes for higher costs and more time spent during remediation—site managers should weigh those costs against the cost and time involved in exposure point sampling.

Mohsen said that this workshop had explored the necessity of determining where the variability lies in a system, but had not talked enough about what one does next. He noted that capping was really the only type of remediation that was discussed during the presentations, and asked why attendees were not talking about upstream solutions (i.e., attacking remediation where there is less variability). Grosso replied that upstream approaches tend to be more traditional solutions, but that it was important for this workshop to address GW discharge into the SW body and how conceptual models, assessment, and remediation could be approached from the SW side. Conant pointed out that with some upstream solutions, if GW plumes have already loaded the sediments with contaminants for decades, one is really just halting the addition of additional mass into the system, not necessarily providing a remedial solution for those contaminated sediments.

A final comment was made about the long-term implications of examining or not examining the potential GW contributions to a site. One attendee noted that it is important to consider GW–SW interactions in terms of their contaminant contribution to a site and also in terms of post-remediation. If a site team is going to spend money on cleanup, they certainly do not want to risk re-contamination because they have not thought about the contribution of the GW. In that type of situation, there is more at stake than incremental risk.

WRAP UP

Next Steps

Grosso asked for volunteers to put together a document summarizing consensus understandings of GW–SW interactions that could be gleaned from the presentations and discussions of this workshop. The document will be based on the minutes of the meeting. It will address Grosso's original set of questions ("Introductions and Focus of the Workshop," above) and will discuss some of the case examples presented, but it will not be more than 10 pages or so long. It was suggested that someone with experience in urban rivers contribute to the document, because those systems are different from smaller stream research projects: when working with urban rivers, one must answer questions about the importance of GW input to the system and find approaches that are reasonable to regulators. Attendees volunteered to work on sections of the document. The document will be distributed to RTDF members for review. (See "Action Items", below).

Future Meetings and Workshops

Grosso said that the RTDF Action Team will hold additional meetings and workshops in the future. She asked for input on future workshop topics, and presented a list of topics that had been proposed during a spring 2002 poll. Meeting attendees responded favorably to the topics that Grosso presented. Grosso said that some people have expressed strong interest in holding a workshop that focuses on evaluating uncertainty in remediation, assessment, and risk assessment. Grosso said that it might be appropriate to have the next RTDF Action Team meeting focus on remediation issues since the last few have had an assessment focus. She also suggested holding a conference that would provide updates on in situ treatment technologies. When an in situ workshop was held 3 years ago, she said, most technologies were in planning/bench-scale stages. Some of the technologies have since progressed to the stage of field pilots and projects. It would be interesting to learn more about these projects. Finally Grosso told the attendees that there is interest in holding an RTDF workshop on dredging. The potential organizers recently realized that the state of the science and engineering on dredging is not well documented. They are specifically interested in the magnitude of solid and associated chemical releases to media that results from dredging, not only at the site, but also during the transport and disposal of dredge materials. She said that there has been discussion about developing predictive models to realistically predict the effects of potential dredging releases when evaluating remediation alternatives.

Miscellaneous

Kenneth Wittle announced that the Interstate Technology Regulatory Council (ITRC) now has a sediments working group and that the group could use more regulatory participation. He asked state representatives who are not yet participating to consider doing so. He also said that he had volunteered, through another working group, to try to summarize available treatment technologies for sediments; he asked anyone with suggestions or information in that area to contact him. (See Attachment A for contact information.)

ACTION ITEMS

• Attendees volunteered to write a summary of the consensus items that were identified about GW–SW interaction issues during the workshop. John Edwards will document issues related to larger urban rivers. Ralph Stahl (in abstentia) and Bruce Duncan agreed to discuss tiered approaches and the ecological viewpoint. Farukh Mohsen, Nancy Grosso, Bob Maxey, and E. Erin Mack also volunteered to contribute to the document.
  
• Grosso will poll the Sediments Remediation Action Team to determine interest in a topic for the next Action Team meeting.

ATTACHMENT A: Attendee List

SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
SEDIMENTS REMEDIATION ACTION TEAM MEETING
West Coast Grand Hotel at Fifth Avenue
Seattle, Washington
October 29-30, 2002

ATTACHMENTS B THROUGH Q

SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
SEDIMENTS REMEDIATION ACTION TEAM MEETING

West Coast Grand Hotel at Fifth Avenue
Seattle, Washington
October 29-30, 2002

Attachments B through Q are available on the Internet. To view these attachments, visit the RTDF home page at http://www.rtdf.org, click on the "Sediments Remediation Action Team" button, then click on the "Team Meetings" button. The attachments will be available as part of the October 2002 meeting summary.