SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
WELCOME AND OPENING REMARKS
Mark Lyverse and Bob Maxey, co-chairs of the Non-Aqueous Phase Liquid (NAPL) Cleanup Alliance, welcomed attendees (see Attachment A) to the meeting. Maxey explained that the Alliance was formed five years ago and has worked on several projects, including training modules and a NAPL decision-making framework. He noted that the training deliveries have been well-received and enjoy the full support of his office (Office of Solid Waste) at EPA. Lyverse added that industry and regulators are also beginning to use the Alliance’s decision-making framework. Maxey provided an overview of the agenda and introduced the guest speakers. He thanked the speakers for attending, noting that they would be presenting innovative research on NAPL behavior, management, and recovery, along with case studies. Maxey also thanked Don Ficklen and the Air Force Center for Environmental Excellence (AFCEE) for hosting the meeting and arranging several of the speakers.
RCRA PROGRAM GOALS
Tom Rinehart spoke about EPA’s new goals for corrective action under RCRA and how they relate to NAPL management (see Attachment B). Rinehart addressed two sets of goals: a set of national goals for 2008 and a longer-term goal for the year 2020. For the 2008 goals, Rinehart said, EPA has expanded its set of progress indicators from two to five and will track these indicators for approximately 2,000 high-priority sites nationwide. By 2008, EPA aims to achieve the following results for these five indicators:
The longer-term goal is derived from the 2003 RCRA paper Beyond RCRA: Waste and Materials Management in the Year 2020, which envisioned having existing contamination “largely under control” by 2020. An EPA/state workgroup has suggested that this goal be defined as having construction complete—i.e., all infrastructure to support the active remedy—at 95% of facilities by 2020. This goal could apply to roughly 4,000 sites, including the sites being tracked for the 2008 goals, other facilities currently on the permitting track, and additional sites identified by states and regions.
Rinehart identified several key steps toward meeting these goals. Remediation objectives will continue to be driven by the need to be protective of human health and the environment, as interpreted by the states, with land use a key consideration. EPA will continue to improve guidance, particularly for exposure issues such as vapor intrusion, and will work with industry to develop implementation strategies. Because EPA will expand the baseline list of facilities for the 2020 goal, there may be a need for accelerated remediation through greater use of innovative technologies. EPA will continue to track the five progress indicators, and will work to integrate the 2020 goal into the Agency’s strategic planning. Training will also be a key component of EPA’s effort, as the Agency is developing a RCRA project manager curriculum (a list of recommended training courses, including the training prepared by the NAPL Alliance) and a traveling “Corrective Action Road Show” to provide information about remedy selection. Rinehart said that EPA will seek industry feedback on a pilot “Road Show” scheduled for June 2006.
NAPL SOURCE ZONE DECAY
Doug Downey explained that LNAPL plumes naturally decay through dissolution, volatilization, and biodegradation (see Attachment C). Many remediation efforts have assumed that the rate of this “weathering” is 5% per year, but the scientific basis for this assumption is limited. Weathering rates are influenced by chemical- and site-specific parameters, including plume thickness and the type of aquifer substrate. To support more scientifically defensible LNAPL weathering rates, AFCEE has collected field data from various sites.
AFCEE analyzed several sites with known NAPL releases, temporally rich sampling records, and no remedial activities that would obscure the natural rate of plume decay. Weathering generally appeared to fit a first-order decay curve. For total BTEX, the average weathering rate across sites was 16% per year (11% as a conservative default). For benzene, which typically poses the greatest risk to human health, the average weathering rate was 26% per year (19% as a conservative default). Higher weathering rates for benzene reflect increased water solubility, suggesting that dissolution is the primary driver of weathering. Indeed, AFCEE found that the site parameter most strongly correlated with weathering rates was ground water velocity (which in turn affects dissolution rates). Two sites excluded from the averaging showed very little weathering because they had almost no ground water flow.
Downey noted that some sites may have sufficient data to determine a site-specific LNAPL weathering rate—which in turn would allow the prediction of future LNAPL concentrations through modeling. For sites without robust data, however, the weathering rates described above should be a more appropriate default than the old 5% assumption. Downey also noted that weathering rates are relatively higher in thinner parts of the NAPL plume, where there is proportionally greater contact with water.
SOURCE ZONE ATTENUATION
Paul Johnson spoke about his research on source zone natural attenuation (SZNA), which involves the natural decay of NAPL plumes through dissolution, volatilization (soil gas), and biodegradation (see Attachment D). He outlined the types of data needed to answer three key questions:
Johnson presented three case studies to illustrate his approach for evaluating SZNA. At the Guadalupe site in California, plumes of diluent are present below a former oil production field. Electron acceptor data indicate that significant biodegradation is taking place, while mass transport modeling suggests that the gas transport-related loss rate (i.e., volatilization) is greater than the rate of loss through dissolution—particularly in shallow/exposed source areas (SZNA rates vary with depth). At an aluminum smelter in New Zealand where a 300,000-liter diesel spill occurred, modeling suggests that vapor transport is the greatest source of loss; site data correlate with the decay curve suggested by the model. Vadose zone samples confirm that NAPL composition is changing. At the former Williams Air Force Base in Arizona, there is a deep plume of roughly 1,000,000 gallons of aviation gasoline and JP-4 (a jet fuel). Johnson described bench-scale tests to predict future dissolution, noting that dissolution and biodegradation may be comparable for submerged source regions such as the Williams plume. Overall, Johnson explained, it is helpful to have about 10 years of good data to account for normal variability, but the techniques described in this presentation still provide useful information about whether SZNA is occurring and how it may proceed in the future.
NAPL SOURCE ZONE LONGEVITY
Chuck Newell discussed several aspects of NAPL source zone longevity (see Attachment E). Using data from 59 chlorinated VOC (CVOC) sites, he compared four source depletion technologies: enhanced biodegradation, chemical oxidation, surfactants/cosolvents, and thermal enhanced extraction (TEE). Chemical oxidation lowered CVOC concentrations by several orders of magnitude, but concentrations rebounded significantly after treatment. Surfactant/cosolvent treatment also achieved source reduction, and concentrations remained stable after treatment. TEE reduced the source by several orders of magnitude, but data were insufficient to evaluate post-treatment effects. Enhanced biodegradation (i.e., adding an electron donor to reduce CVOCs) achieved source reduction that continued even after treatment was discontinued, although there was relatively less reduction in total CVOCs because biodegradation largely converts from one CVOC to another. To determine the cost-effectiveness of these technologies, Newell consulted various peer-reviewed sources and government agencies. Surfactant/cosolvent technology was the most expensive per cubic yard treated, while enhanced bioremediation was the least.
For comparison, Newell examined trends in source zone decay at 23 untreated CVOC sites. Natural decay had a median half-life of 2 to 6 years, depending on the chemical. Several lines of evidence suggest that this decay occurs in a first-order fashion. Models such as SourceDK (a mass flux model developed for AFCEE by Groundwater Services, Inc.) can predict natural attenuation rates from site data. Because the first-order decay curve has a long “tail,” Newell suggested that partial source reduction by active treatment may not substantially shorten the timeframe to reach a remediation goal. For example, if one aims to reduce source mass by 100%, removing 80% of the source by active treatment might only reduce the remediation timeframe by 17%. Active treatment is still necessary in many cases—for example, pump-and-treat systems might be needed to reduce plume mobility, and excavation or TEE might be justified at sites requiring a rapid cleanup. However, for other sites, source containment and monitored natural attenuation (MNA) may be sufficient. A decision chart can help determine where active source depletion is truly necessary.
INNOVATIVE LNAPL RECOVERY TECHNIQUES
Patrick Haas discussed several challenges associated with site characterization and free product recovery, using technical data from numerous field tests to illustrate innovative strategies and “lessons learned” (see Attachment F). He noted that while many regulations are based on the amount of free product that forms in a well, this is not an accurate measure of the mobility of the NAPL plume or the potential for recovery. Field data suggest that LNAPL is readily broken into a scattered (residual) phase that cannot easily be recovered—for example, a temporary water table shift may smear the product and cause a permanent decrease in recovery rates, even if product is still able to form in a well. There are other complexities to NAPL behavior as well, including capillary fringe effects that may render the NAPL plume stable.
Haas suggested a battery of low-cost field tests to determine the feasibility of recovery and select the best treatment technology for a site. For example, repeated baildown tests will show whether a product is sufficiently mobile to re-form in a well after withdrawal—a good indication of recoverability. Baildown tests can also determine the feasibility of lowering the water table for enhanced recovery (many technologies are more effective with a low water table). In addition to the traditional skimming and vapor monitoring tests, laser-induced fluorescence (LIF) can indicate the depth and composition of the NAPL, while aeration tests can indicate the potential for biodegradation in the smear zone.
Case study data show that where the product is mobile, vacuum-enhanced liquid recovery is effective, particularly using a liquid ring pump. However, a better option might be to use vacuum-enhanced dual-drop technology, in which one tube collects the LNAPL and the other collects water and soil gas (but has a PVC shield to prevent NAPL entry). The result is less NAPL in the extracted water, which leads to lower treatment costs. Conversely, in cases where the NAPL plume is not mobile, biodegradation may be the most appropriate treatment. Overall, Haas noted that a battery of field tests can help determine the “maximum extent practicable” for removal. He suggested that a good short-term remediation objective would be to determine whether the NAPL is mobile and, if so, remove it until it is no longer mobile.
LAND USE AND REMEDIATION STRATEGIES FOR LNAPL SITES
In 1998, the Air Force produced a “Handbook for Remediation of Petroleum-Contaminated Sites.” Downey presented an overview of the Air Force’s remediation strategy, which involves five key steps: (1) matching the remedy and available land use controls to eliminate potential exposure, (2) promoting risk-based standards, (3) site characterization, (4) accounting for natural attenuation, and (5) promoting cost-effective remedies (see Attachment G). Downey then presented case studies to illustrate the strategy in action for various land use scenarios:
LNAPL CHARACTERIZATION AND RECOVERY AT DIEGO GARCIA
Jerry Hansen described remediation activities at Diego Garcia, an Air Force base located on a small island in the Indian Ocean (see Attachment H). Three large plumes of JP-5 (jet fuel) were discovered during the 1990s; two of these spills threatened production wells. AFCEE investigated the plumes, determined through repeated baildown testing that the fuel was recoverable, and decided that bioslurping would be the most effective remediation strategy. Many other Air Force sites have also chosen this simple but aggressive technology. At Diego Garcia, bioslurping systems removed over 100,000 gallons of fuel from the spill sites. Bioslurping was followed by air injection (bioventing) to enhance biodegradation of residual NAPL. Active remediation is now ending and spill areas are being monitored to ensure that natural attenuation is taking place, including natural “bioventing” driven by tidal fluctuations in the water table. The project has been successful, with recovery wells no longer showing any NAPL.
Throughout the project, the Air Force found ways to minimize cost without diminishing recovery. For example, because recovery was expected to take only a few years, the Air Force constructed simple systems—e.g., PVC piping on the surface—rather than complex, permanent systems. Where possible they used existing infrastructure, including an existing oil/water separator and tie-ins to existing power supplies. They carried out bioventing by connecting air blowers to the network of PVC pipes and wells that had already been installed for bioslurping. A portion of the system was located on the aircraft ramp but was designed to be easily dismantled in case the ramp was needed—which it was after September 11 th. This part of the recovery system was subsequently replaced with a trailer-mounted bioslurping unit.
On a more general note, Ficklen explained that AFCEE is using pilot sites to evaluate a wide range of technologies. He noted that AFCEE has found TEE to be largely ineffective. Haas pointed out that vacuum systems also have limitations: for example, while a vacuum system can recover product as deep as 200 feet, it is less energy-efficient than a mechanical pumping system for anything below 60 feet.
APPROACH TO LNAPL TREATMENT AT WILLIAMS AIR FORCE BASE
Javier Santillan discussed LNAPL treatment at the former Williams Air Force Base in Arizona, which is being reused as a college campus. Below the base is a plume of at least 1,000,000 gallons of fuel, with a depth to ground water of approximately 170 feet. This ground water is not expected to be needed for human use and there do not appear to be any direct pathways to receptors. Nonetheless, some stakeholders would like to see active recovery, which could be challenging because the plume is deep, it lies below several impermeable layers, and the water table has recently risen more than 40 feet, possibly smearing the product. The current ROD prescribes pump-and-treat, but thus far pilot pumping systems have recovered very little product. Evidence suggests that substantial biodegradation is occurring.
The Air Force’s strategy for this site reflects a Performance-Based Management (PBM) approach, with cleanup goals based on risk management and future land use. PBM also includes optimizing treatment and long-term monitoring systems, as well as issuing performance-based contracts that transfer risk to contractors and encourage them to choose the most effective treatment technology. In this case, AFCEE and its contractor did not identify any treatment that would achieve ground water standards at Williams within 100 years. They predicted that TEE could remove 40% of the product, but would be expensive and would still leave hundreds of thousands of gallons of fuel. Some stakeholders have expressed concerns that the vapor generated by TEE could enter onsite buildings. Further, a review of other sites indicated that TEE alone is unlikely to achieve closure of a ground water remediation site. A combination of SVE and MNA would take somewhat longer than TEE combined with MNA, but would be much less costly.
Several issues have yet to be resolved at this site. Although plume stability has been inferred, it still must be formally demonstrated. AFCEE and the regulators also need to agree on more definite goals and a timeframe for the remediation, to ensure that an exit strategy is in place. Santillan suggested that under a risk-based approach, the Air Force might not devote substantial resources to this site because there are many other sites with more urgent remediation needs.
COMMERCE CITY SITE CASE STUDY
Greg Fletcher and Chris Pearson described remediation activities at Suncor’s refinery in Commerce City, Colorado, which has been active since the 1930s (previously operated by Conoco). Soil and ground water below the site have been impacted by various hydrocarbons. Historically, the site’s management goal was to prevent oil from seeping into a creek adjacent to the property. From the late 1970s to the early 1990s, Conoco responded to state orders by installing an interceptor trench, a slurry wall, a high-density polyethylene mitigation barrier, an SVE air sparging system, and a large pump-and-treat system, all along the property’s edge. Conoco also diverted the creek away from the contamination.
In 1998, the company and the state began to focus on long-term management strategies. To permanently protect the creek, an additional barrier wall was installed, pump-and-treat was continued, and the creek was restored to its natural channel. The 1998 plan also required the company to recover free product “to the maximum extent practicable” within the refinery. This goal was interpreted to mean reducing LNAPL mobility to a practical endpoint, defined in terms of conductivity/seepage velocity.
Several techniques were used to determine plume mobility and recoverability within the refinery property. To determine the amount of free product available, Suncor’s consultants measured NAPL thickness using ROST, in which aromatics and double-bonded hydrocarbons fluoresce with intensity proportional to saturation. ROST data were converted to a saturation profile using soil cores (porosity) and an API correlation model. Mobility was determined through repeated baildown tests to calculate seepage velocity, which appeared to be below the practical endpoint for recovery. Mobility and saturation values were plugged into models to determine recoverability—which in this case appeared to be relatively small. To test their prediction, Suncor’s consultants performed a waterflooding/extraction experiment on a test plot, achieving results close to the model prediction (1,500 gallons recovered, compared to 800,000 gallons previously recovered by the pump-and-treat system near the creek). These results confirmed that the plume below the refinery is stable. Future efforts will involve SVE in the interior of the property (because stakeholders want some free product removal), continued pump-and-treat for ground water adjacent to the creek (free product has been removed), and phytoremediation along the creek.
THE ALLIANCE’S TRAINING PROGRAM
The NAPL Cleanup Alliance is developing two training modules that describe technical aspects of NAPL distribution and mobility, correct common misconceptions about NAPL, and explore NAPL management issues. An update on the status of the training modules was presented during the meeting.
Module 1 (NAPL Basics)
Module 1 is complete and has already been used in training sessions. Several additional sessions are scheduled for the coming months, as listed on the RTDF Web site. Maxey reported that the training deliveries have been well-received and they have the full support of EPA/OSW. He urged Alliance members to contact him if they know of additional interest. He noted that Rick Greiner is hoping to set up an additional training in Montana and others are asking for a repeat of the popular Internet training. Maxey encouraged industry representatives to invite consultants or regulators to the training sessions.
Module 2 (NAPL Management)
In response to comments received at the last Alliance meeting, Harley Hopkins worked with Maxey, Ellen Rubin, and Jim Higginbotham to shorten Module 2 to a one-hour “overview.” Hopkins explained that the purpose of Module 2 is to present the NAPL management framework, explain what is known about managing NAPL behavior, and describe some of the management tools that are available.
Hopkins reviewed Module 2 slide-by-slide to give Alliance members an opportunity to comment (see Attachment I). Several of the suggestions from the group pertained to the presentation as a whole. These “global” suggestions include the following items, which the group agreed to unless otherwise noted:
Members of the Alliance also suggested a number of specific changes to the slides. The following changes were agreed to by the whole group unless otherwise noted:
CHEVRONTEXACO CASPER PROJECT UPDATE
Brian Smith provided an update on the Casper refinery site, which the Alliance has followed as a case study for several years. ChevronTexaco and TriHydro have mapped benzene mass and vapor diffusion efficiency (vapor flux based on geology and moisture content). Testing and modeling indicate that the NAPL plume is stable, with pore entry pressure impeding further migration. The only movement appears to be near extraction wells. Preliminary results from a longevity model indicate that benzene depletion depends largely on vapor diffusion efficiency and the hydraulic gradient. Dissolution and biodegradation will likely contribute to the largest source reductions.
After an extensive review of treatment technologies, ChevronTexaco and TriHydro are considering installing a capillary barrier—a wall of permeable material that would allow natural ground water flow but impede any migration of NAPL into the river adjacent to the site. This approach would offer a permanent solution, with the added benefit of achieving total cleanup of the smear zone in the area where impacted soil is excavated to make way for the barrier. Smith and Lyverse noted that the smear zone extends nearly to bedrock in many places, which means a hanging slurry wall probably would not impede the flow of NAPL while simultaneously allowing natural ground water flow conditions to exist.
Through bench-scale testing of candidate materials, ChevronTexaco and TriHydro identified a fine-grained aeolian sand from a nearby deposit that could resist NAPL breakthrough while allowing natural groundwater flow conditions to exist. Modeling results suggest that this barrier material will prevent NAPL migration. The next phase of testing will occur in the field, within a test cell that represents an area of low-viscosity NAPL and higher hydraulic conductivity. The test will involve dewatering, excavation, barrier installation, re-watering, and monitoring for NAPL and dissolved impact break-through for up to 6 months. This test should address several concerns raised by Alliance members, including questions about bioslime development (potentially altering entry pressure), differential cation exchange potentials between aquifer materials, and the barrier’s effects on the overall hydraulic gradient. Kremesec cautioned that when NAPL accumulates next to the barrier, it might lead to soil vapor concerns or a depletion of electron acceptors, which would affect the rate of biodegradation. Fletcher suggested accounting for what might happen if trees grow into the barrier, as the depth to ground water is 8 to 10 feet. Smith acknowledged these concerns.
In addition to finalizing plans for the capillary barrier field test, ongoing efforts at the Casper site include continued study of the composition and longevity of the NAPL plume and revisions to the long-term conceptual site model. Lyverse added that ChevronTexaco is also working on a management strategy for the upgradient plume.
NEW NAPL ALLIANCE PROJECT: TECHNOLOGY CASE STUDIES
Several members of the Alliance noted that there appears to be a wide body of literature reviewing innovative remediation technologies—including analyses presented by the speakers at this meeting—but that these studies are scattered among different venues. Kremesec and Adamski suggested that the Alliance could serve a useful function by compiling and reviewing these studies. Adamski, Ficklen, and Hansen agreed to serve on a Case Study Subgroup, which will develop a methodology and criteria for review and then determine the level of effort that will be required on the part of the Alliance to apply this methodology. Other members of the Alliance agreed to pass applicable case studies on to the Case Study Subgroup. Members suggested the following sources of technology analyses:
Adamski suggested that the Subgroup focus on large sites; Lyverse added that it would be particularly interesting to see whether sites have been able to meet goals and return to drinking water standards. Zabcik urged the group to take a broader view as well, and consider the value of the energy/resources that a treatment technology will use. For example, he suggested that in some cases, the social cost of the energy used for TEE (and the greenhouse gases released) might exceed the social benefit of the remediation.
FUTURE OF THE NAPL ALLIANCE
Funding for the NAPL Alliance currently comes from EPA’s Technology Innovation Program. Rubin announced that as of April 2006, this funding will no longer be available. In addition, Rubin explained that, while evaluating and providing information on new technologies for site characterization and remediation (principally through field projects) was a major commitment in the Memorandum of Understanding signed by Alliance Core Team members, such information has not been forthcoming. While the decision-making framework and the training projects have been good, they are not the products that were envisioned when EPA decided to fund the Alliance. Many participants expressed interest in finding a way to continue meeting, and noted that the products developed by the group can potentially deliver large benefits from a small investment. Some members suggested advocating with EPA that there is still work to be done; others asked if the Alliance might be able to continue if it produced more concrete deliverables and focused more on technologies, to counter EPA’s perception that the Alliance may have strayed from its original mission. Rubin explained that, at this point, largely because of budget concerns, she did not believe continuation was an option. She suggested, however, that the Alliance might look for other venues in which to continue its work. Rinehart suggested that the group might be able to switch its focus toward corrective action; however, he indicated that funding would be an issue in his office as well. A budget of roughly $70,000 to $80,000 per year, which includes administrative and meeting support, is needed to sustain the Alliance. (The RTDF Web site is funded separately under EPA’s Clu-In program, and EPA expects to continue its support for it.) Hopkins noted, however, that the RTDF program has more than just monetary value: it also provides a credible forum where all members receive equal standing. The group will not be able to continue to use the “RTDF” name once EPA funding ends. Thus, participants moved on to discuss potential funding sources and a number of possible venues that would offer the group a credible forum for industry-government collaboration. The following options were suggested:
Overall, the members of the Alliance expressed interest in exploring ways to continue to meet. Core Team members agreed to brainstorm ideas via telephone and email. Rubin will contact ITRC to determine the potential for establishing a NAPL team that could continue the Alliance’s work, and will prepare a list of mechanisms for industry to work with EPA on research and technology issues.
Attachments A through I
Attachments A through I are available on the Internet. To view these attachments, visit the RTDF home page at http://www.rtdf.org, select the “NAPL Cleanup Alliance” button, then select the “Alliance Meetings” button. The attachments will be available as part of the February 2006 meeting summary.
Attachment A: Final Attendee List (PDF, 20 KB)
Attachment B: RCRA Program Goals (PDF, 205 KB)
Attachment C: NAPL Source Zone Decay (PDF, 285 KB)
Attachment D: Source Zone Attenuation (PDF, 16 MB)
Attachment E: NAPL Source Zone Longevity (PDF, 1.8 MB)
Attachment F: Innovative LNAPL Recovery Techniques (PDF, 2.2 MB)
Attachment G: Land Use and Remediation Strategies for LNAPL Sites (PDF, 647 KB)
Attachment H: LNAPL Characterization and Recovery at Diego Garcia (PDF, 1.2 MB)
Attachment I: Draft Module 2 Training (PDF, 1.1 MB)