SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
PHYTOREMEDIATION OF ORGANICS ACTION TEAM
TOTAL PETROLEUM HYDROCARBONS IN SOIL SUBGROUP
CONFERENCE CALL

April 7, 1998
11:30 a.m. to 1:00 p.m.

On Tuesday, April 7, 1998, members of the Remediation Technologies Development Forum (RTDF) Phytoremediation of Organics Action Team, Total Petroleum Hydrocarbons (TPH) in Soil Subgroup, held a conference call. The following members participated:

Lucinda Jackson, Chevron Corporation (RTDF Action Team Co-Chair and Subgroup Co-Chair)
Phil Sayre, U.S. Environmental Protection Agency (EPA) (Subgroup Co-Chair)
Steve Rock, EPA (RTDF Action Team Co-Chair)
Jim Brown, ReAct
Dawn Carroll, EPA, Technology Innovation Office
Evelyn Drake, Exxon Research and Engineering
Felix Flechas, EPA, Region 8
Eric Foot, Battelle
Andrea Leason, Battelle
Royal Nadeau, EPA, Environmental Response Team, New Jersey
Mike Reynolds, U.S. Army Cold Regions
David Tsao, Amoco Research Center
Nish Vasavada, Eastman Chemical Company
Charlene Owens, Exxon Product Research

Also present was Kristen Honey of Eastern Research Group, Inc. (ERG).

INTRODUCTIONS

Lucinda Jackson, Co-Chair for the TPH in Soil Subgroup, began the conference call by saying that the primary agenda for the call was to discuss general protocol issues applicable to RTDF TPH field tests. Jackson emphasized that the protocol (Attachment A) remains open to changes and the RTDF Action Team has been receiving numerous suggestions and helpful feedback. As highlighted in the previously distributed agenda, other topics to be discussed during the conference call included:

• Petroleum Environmental Research Foundation developments
• regulatory perspectives about phytoremediation
• field testing site selections
• field testing resource commitments
• future conference calls and meetings

Participants in the conference call introduced themselves. Ernest Lory (Navy Facilities Engineering Service Center), Kathy Banks (Perdue University), and Peter Kulakow (Kansas State University) were not present but have remained in contact with the Action Team.

PETROLEUM ENVIRONMENTAL RESEARCH FORUM (PERF) UPDATE

Jackson began the update by summarizing discussions from a conference call held on April 6, 1998. The PERF group discussed how to best join efforts with the RTDF Subgroup. To date, three companies (Elf Aquitaine, Chevron, and Exxon) have contributed $25,000 to PERF. These companies are interested in coordinating efforts with the RTDF Subgroup but expressed concern about the handling of proprietary information. Elf Aquitaine, Chevron, and Exxon would like to keep raw data private (e.g., not have it posted on the Web) and requested more information about EPA government publishing timelines and requirements.

Steve Rock of EPA responded that the parties involved in the cooperative research and development agreement (CRADA) (i.e., PERF and the RTDF Subgroup) decide on a publication schedule when they sign the CRADA. EPA cannot "protect" data because they are subject to Freedom of Information (FOI) requests, but the government publishing timeline allows 6 months from the date the last data are analyzed before results need to be made public. The CRADA group will not need to publish sooner and interim reports will not be made public. If the CRADA group does decide it wants to publish results before the end of the study, all changes in the publication schedule must be agreed on by all parties.

Evelyn Drake of Exxon Research and Engineering continued the PERF update. The following individuals and their respective organizations participated in the April 6 PERF conference call:

• Evelyn Drake, Exxon Research and Engineering
• Charlene Owens, Exxon Product Research
• John Fink, ReTech
• Arno Morage, Elf Aquitaine
• David Tsao, Amoco Research Center
• Lucinda Jackson, Chevron Corporation
• Jim Myers, Texaco
• Wayne Fishs, Philips Petroleum Company

Jim Myers and Wayne Fishs may have some sites that could be used for field studies, but both were concerned about making data available to the public. Mike Rightmire of Arco and John Harju from GRI could not participate in the call but remain interested and hope to be included in subsequent PERF activities. Drake said that others had expressed interest in the Phase II project but that PERF has decided not to draw on these individuals until a more detailed project description exists.

PERF participants discussed ways to maximize the cost-effectiveness of the Phase II project to involve at least two sites. The group discussed using only one phytoremediation treatment (whether a mixture of tall fescue and ryegrass species or a native species mix) and an unplanted control area at field sites. They discussed the pros and cons of plugging, slipping, and seeding sites and the subsequent differences in root development under these three sets of conditions. Additionally, the PERF group discussed irrigation conditions and sampling at only one depth (0 to 6 inches).

During the PERF conference call, it was pointed out that establishing a dense root system may take more than the 3-year project duration. At such sites, Drake and other PERF members hoped that PERF and the RTDF Subgroup would not terminate study efforts after 3 years but would continue longer term sampling for the benefit of all parties.

MERGING PERF AND RTDF EFFORTS

Drake said that PERF members are very close to clearly stating their project definition and all recognize the benefits of working with the RTDF Subgroup. For example, the benefits from standardizing the protocol have already been seen.

The discussion on joint PERF/RTDF efforts prompted a discussion about CRADA legalities. It appears from previous bioremediation consortiums (e.g., Oak Ridge National Laboratory) that the CRADA group can use the standard PERF contract for PERF members and add Memorandums of Understanding (MOUs) for any non-PERF members. There was concern that a MOU may not allow EPA to provide in-kind services as it would if it joined with a CRADA. Steve Rock responded that there is still some "fuzziness" on exactly how EPA will interact formally with these groups. EPA's participation appears to involve two different roles: (1) as a partner in the overall coordination of the CRADA and (2) as a site participant. To help clarify contractual uncertainties, Phil Sayre is going to provide Rock with some example CRADAs.

Drake mentioned that, during the PERF conference call, she described the PERF group as functioning as a subgroup of the RTDF Phytoremediation of Organics Subgroup. Under this scenario, PERF (and PERF funding) would join the RTDF project at a field test site or two. Drake expressed the concern that information obtained from PERF efforts and resources could be dissemination in the future to groups who did not contribute resources. A precedent exists, however, for closed steering committee meetings that are comprised of only those individuals/organizations who contribute money or in-kind resources (e.g., laboratory time or labor equivalent to the annual partner contributions [$25,000 for PERF]). It is expected that within the next couple of months some RTDF conference calls would be exclusively for those members participating in the CRADA (e.g., PERF members and EPA) and that the CRADA group would function as a contained subset of the RTDF Subgroup and discuss ongoing developments in field study projects. Information discussed in these closed meetings would be kept private within the steering committee. It was recommended that interested individuals contact Dave Ellis with DuPont (302-892-7445) for additional information.

DISCUSSION OF GENERAL PROTOCOL APPLICABLE TO RTDF TPH FIELD TESTS

Purpose

Jackson began the discussion by explaining that the goal of developing a standardized protocol is to create a complete package of information that will assist in promoting the acceptance of phytoremediation technology. The RTDF Subgroup must obtain data to assist regulators in determining whether or not phytoremediation is a viable treatment. Specifically, the group must address whether or not petroleum hydrocarbons chemically degrade during the bioremediation process.

Treatments

Jackson said that the PERF group has several good suggestions on standardizing a field study protocol. For example, the PERF group mentioned that the protocol might include a short germination test in the field before sites are planted in order to ensure that treatment species can grow in local conditions (e.g., soil type, contamination, salt). In Phase I of the PERF project, screening of plants in gas and sludge soil occurred in small, heavily watered greenhouse pots. Many species survived and grew well in the greenhouse germination test, so they were selected for additional phytoremediation experiments. Once these species were planted in larger pots, however, they all died. Because of these results, the PERF group realized that depending on the degree of salt contamination, they would have to flush the soil for species to survive and that a small pot, intensely watered screening protocol may not be representative of the impact of salt (and possibly other kinds of contaminants) in the field. The RTDF Subgroup supported the idea of including a short-term, field germination test in the protocol prior to the treatment phase of field tests.

The PERF group and RTDF Subgroup both agree that an appropriate perennial plant species (e.g., tall fescue or ryegrass species) should be used as a standard in all phytoremediation projects. Tall fescue and ryegrass were suggested as standard treatment species because both are widely adaptive and can be planted successfully at most locations. Several PERF tests have shown that tall fescue and ryegrass grow in contaminated soil. The presence of salt, however, was not included in the PERF protocol and may be something that the RTDF Subgroup should consider. Even though the PERF study is limited to only two treatments, the PERF group discussed the use of Bermuda grass during their recent conference call because it is fairly salt tolerant.

The RTDF Subgroup discussed using native grasses as an additional, optional treatment at each site. After recognizing that various sites and treatment technologies have used nongrass species (e.g., poplar and willow trees), the group decided that the optional treatment protocol should be flexible and open to include any plant or tree species.

The RTDF Subgroup also discussed a protocol for an unplanted control area. They determined that if field studies use a limited number of treatments, a weed-free, unplanted control would be more useful to have than an unweeded (i.e., natural vegetation) control. Drake explained that the reasons for preferring a weed-free control are based on current understanding of the mechanisms of bioremediation (Laura Carreira's suggestions aside). Bioremediation processes appear to occur because plant roots facilitate soil aeration and promote aerobic biodegradation by using an enzyme exudate or special microbial communities in the root region. If this bioremediation mechanism holds, control areas without weeds will contrast more strongly with planted sites than would control areas with weeds.

Currently, it is common practice to control vegetation with weedkillers. It is not known if weedkiller depresses microbial populations promoting bioremediation. Drake explained that PERF refinery sites frequently use weedkiller to prevent fire hazards and reduce rodent populations. The RTDF Subgroup, therefore, decided that they probably should make some recommendations on what to use for weed control. Additionally, the RTDF Subgroup would like to run some microbial analyses of unweeded control sites to learn more about the impacts of weedkiller in the plant root region.

During protocol discussions, several participants raised issues encountered at their sites. Steve Rock explained that at his Ohio site contamination goes down 3 to 5 feet, with the heaviest contamination occurring in the top 1 or 2 feet. Because grasses are generally limited to the top 6 inches, Rock was thinking of planting one or more tree plots. Rock said that this site is large by experimental standards (5.5 acres, about 5 acres of available land) and might be a good location to plant grasses (1 or 2 species that everybody uses) as well as native tree species. Drake responded that tall fescue and ryegrass species are capable of sending roots down to depths of 3 to 4 feet, but this root development is time- and condition-dependent. It can take grasses 5 years or more to develop dense root systems, depending on numerous conditions, including the degree of filling, soil, moisture, fertilizer, and initial tilling practices. In general, the group supported Rock's idea of comparing tree-site treatment verses a grass-site treatment. According to Drake, similar comparisons may be made at Richmond field sites, where treatments are going to include not only the standard protocol treatments and the unplanted control but some treatments with other species and experiments with soil cover. The soil cover experiments will bury plants under a clean soil cap and measure differences in survivability, growth, and phytoremediation between plots with a soil cover and those without a soil cover. These soil cover experiments are not only efforts to maximize phytoremediation but are also potentially going to reduce the infiltration of contaminants into ground water. It was agreed that no matter what optional treatments are applied to individual sites, each field study location will ensure that it incorporates the standard treatments outlined in the protocol.

Locations

The RTDF Subgroup discussed whether they want to incorporate certain site types or characteristics into their field study (i.e., should they try to target certain climatic areas?). Drake suggested that if the RTDF Subgroup chooses its field study sites to reflect the distribution and characteristics of sites that need cleaning, phytoremediation increases its chances of being used commercially. Unfortunately, EPA does not have a list of U.S. sites with surface soil contamination that may benefit from phytoremediation. It was agreed that the best way to select field study locations may be to find locations where: (1) people want to offer use of a site, (2) researchers want to work on the site, and (3) funding exists for the site. Additionally, the RTDF Subgroup decided that field study sites should reflect the diversity of refinery locations found in the United States (i.e., East Coast/West Coast/Gulf Coast; cold/warm; wet/dry; marshy areas/nonmarshy areas).

Soil Sampling

Previously the RTDF Subgroup had talked about limiting costs by limiting the number of soil samples they analyze. The group supported the idea of collecting and analyzing surface soil samples (0 to 6 inches), as well as collecting deeper samples during the initial stages of the study and archiving them. Then, if plant roots do extend below the surface level during the course of the study, the deeper samples could be taken from the archive and analyzed to provide comparison values. It was agreed that archiving samples would only work if the process does not alter soil sample properties and if all archived samples are archived following the same storage protocols. Drake mentioned that PERF archives samples at liquid nitrogen temperature and successfully preserves microbial characteristics under these conditions. The RTDF Subgroup agreed that there is a need to standardize archiving in the field study protocol.

Sayre reiterated Peter Kulakow's suggestion (from a previous meeting) to use a composite soil sample from a single depth and to do an average across microbial parameters. In general, the group supported the use of composite sampling in the protocol.

Sampling Analysis

The RTDF Subgroup agreed that the first analysis should be a general soil analysis (e.g., color, temperature, organic matter, and soil type). Subsequent, laboratory analyses should examine available nutrients, TPHs, and polycyclic aromatic hydrocarbons (PAH).

Drake suggested using biomarkers to characterize bioremediation at field sites with highly variable hydrocarbon concentrations (0.5 to 3% weight hydrocarbon). Use of biomarkers causes field sites to need fewer samples to demonstrate statistically significant differences, especially when high variability exists within a site.

Drake explained the biomarking process. GCMS extracts and analyzes oil/hydrocarbon compounds derived from either microbial or plant processes. Because of their chemical structures, these oils degrade very slowly. Each oil source is imprinted with a hydrocarbon "species" that serves as fingerprinting markers which do not change over time due to the chemical stability of the oil. These oil-hydrocarbon fingerprints have been used as markers to identify oil sources. This technology is commonly used by competing oil companies over disputed oil fields. With biomarkers, researchers can determine the ratio of constituents of interest (which can either be sums of hydrocarbon types or individual polynuclear aromatic compounds) to the very stable oil species, then look at the change in the relative ratio to measure treatment progress over time. In its Phase 1 project, PERF examined two different hydrocarbon contaminants in all samples using GCMS analysis. GCMS analysis adds approximately $200 to $300 dollars per sample, but the number of samples needed decreases due to the dramatic increase in analysis sensitivity. Drake is going to forward papers further detailing the biomarking process to Co-Chair Sayre. Sayre may post these papers on the RTDF Subgroup Web site.

Although the group briefly discussed developing a protocol that can be used across laboratories, there was greater support among group members for a protocol that requires everyone to use the same laboratory for analysis. Arthur D. Little, Inc., (ADL) was recommended as a contractor because the PERF group and several individuals had had high-quality PAH analysis done by ADL. ADL has a very sensitive GCMS technique that includes a cleanup step, which is approximately 100 to 1,000 times greater than the detection limits of conventional analytical lab GCMS analysis. PERF found the best way to contract for sample analysis was to archive samples and send them all for analysis at the same time (preferably once a year, as long as the effects of storage do not interfere with interpretations). Analyzing samples at the same time increases analysis uniformity and decreases analytical variability among field samples from various dates. Additionally, various field sites at different locations could archive their samples and subsequently consolidate them into a single contract. Because contracting for more samples lowers the per sample cost, consolidating the analysis of multiple project samples into a single contract could save significant amounts of money.

Some members of the group expressed concern about contracting with ADL because Rock was unable to find a contract vehicle between EPA and ADL. (He did say, however, that such a contract vehicle may exist.) Rock clarified that EPA contract vehicles should not affect protocol because the CRADA group does not need to go through the EPA. Rather, the CRADA group can decide where to send their samples, which labs to contract with, and what to do with their resources.

To keep project costs down, the group discussed the possibility of combining TPH and GCMS analyses on soil samples. Samples could be taken and divided, with one part analyzed for TPH and another part sent for GCMS analysis. Numerous TPH analyses could be performed at a relatively low cost and then these could be compared with the signature of the one GCMS sample to infer site characteristics. Rock had mentioned during previous group discussions that EPA might be able to cover some of the costs of sample analysis, but no EPA funding sources currently exist to cover project costs.

Plant Sampling

It was agreed that the protocol should include a plant analysis. It is important to look at plant bioaccumulation of contaminants and address the question of whether or not phytoremediation mobilizes contaminants through the plant into the biosphere (i.e., "What happens when geese eat the crop?"). In general, plants are not believed to bioaccumulate hydrocarbons, but the RTDF Subgroup agreed that they need more proof demonstrating that phytoremediation actually breaks down contaminant compounds. One of the major criticisms from EPA's Hazardous Waste Program is that advocating phytoremediation creates a monoculture that may concentrate contaminants. If this perspective persists, other parts of the agency may stop supporting EPA's phytoremediation activities. The group, therefore, wants to be able to defend the concept that phytoremediation is a viable technology and not just a method of bioaccumulating contaminants.

Rock said that some research (including PERF's lab projects and some literature) indicates that plants do not readily bioaccumulate hydrocarbons. Other researchers (possibly Kansas State or Utah University researchers), however, have apparently found that hydrocarbons are mobilized into plants. During the course of the discussion, it was mentioned that hydrocarbons (TCE and some volatile fractions) appear to accumulate in poplar trees. Ryan Miya, a graduate student from Berkeley, reportedly found relatively low concentrations of hydrocarbons in grasses. Moreover, the hydrocarbons present in Miya's grasses were usually chemically transformed before being taken up by the plant.

The RTDF Subgroup discussed various protocol options for plant analysis. One possibility mentioned was running TPH analysis on plants to keep costs down. It was noted, however, that a refinery recently performed extensive TPH analyses in conjunction with independent analyses for "Appendix 9 Constituents" (a "pretty extensive" list of contaminants), and that the refinery did not find correlations between the TPH levels and the contaminants detected. The RTDF Subgroup agreed, therefore, that they should be cautious with the assumption that TPH is a good surrogate for constituent levels present.

To limit costs, the RTDF Subgroup discussed limiting analyses and only looking at highly regulated, key targets (e.g., benzopyrene). Although this type of analysis is expensive, it is part of the GCMS analysis of soil biomarkers. PERF further limited costs by analyzing one composite sample of plant material from eight replicate pots at the end of their study. They found very low levels of constituents in the plant material, indicating that contaminants are not being absorbed into the plant tissue.

The RTDF Subgroup decided that the protocol for analyzing plant tissue ultimately comes down to budget. They do not necessarily need to analyze plant tissue samples at the same time that soil analyses are being performed. It is important, however, that the group obtain results from mature plants (stem and leaf material).

Study and Statistical Design

The RTDF Subgroup agreed that they would standardize the protocol's statistical design with a minimum plot size and number of replications to use. It was mentioned, however, that replicate analyses of a composite sample appeared preferable to analyzing multiple samples without replication.

Growing Conditions

The group agreed that the protocol for growing conditions should allow people to do whatever is necessary to create a good plant establishment. Study sites should be prepared, planted, fertilized, and irrigated based on local needs.

Plant Evaluation

The RTDF decided to add a microbial count and site history description to the plant evaluation protocol. The site history should include information about present contaminants, sources of contaminants, and previous cleanup efforts.

REGULATORY PERSPECTIVE

Felix Flechas of EPA asked the group to consider how phytoremediation may be used technically and practically in conjunction with other regulatory mechanisms. Flechas said in his view, the biggest problem from a regulatory perspective is not land-based refinery contamination but how to keep these contaminants from entering nearby water bodies. Flechas asked the group to consider how plants might be used in the future to attain ground-water protection standards. Plants could be used to prevent plume expansion and migration in addition to treat ground-water plumes. Such phytoremediation uses could protect our watersheds, especially in regions such as the Rocky Mountains, where refineries overlie a relatively shallow depth to ground-water zone and land bordering streams and rivers is a very active biological zone capable of supporting a strong vegetative environment. Big trees can certainly be established in the Rocky Mountains and could go a long way toward keeping contaminants from entering the water cycle, perhaps better than any pumping system devised using mechanical technologies.

Sayre mentioned that if the group places a lot of confidence in biomarker technologies and reduces the number of samples taken in the field, they would need a lot of good literature supporting the use of biomarkers to satisfy questions raised by regulators. Flechas responded that biomarkers generally are used to identify relative changes in degradation or contamination at a site. Regulators will ultimately look at the absolute level of contamination left at the site relative to soil media protection standards, regardless of relative changes in contamination.

Flechas also mentioned that regulatory concern about monocultures generally occurs because of species diversity loss. Monocultures would become a regulatory issue if the group transformed a riparian ecosystem into one treatment species (e.g., tall fescue, perennial ryegrass, or poplar trees). People interested in habitat issues do not like to see nondiversified ecosystems. Additionally, people concerned about habitat generally prefer to see native species (versus an agricultural crop) planted because it creates a better base for that environment to turn into a more diverse habitat.

CLOSING REMARKS AND DATE FOR NEXT CONFERENCE CALL

The RTDF Subgroup did not have sufficient time to fully discuss specific field site locations. Therefore, another conference call was tentatively scheduled for Wednesday, April 15, 1998, from 11:00 a.m. to 1:00 p.m. (EDT). Before the next conference call, Sayre would like people to e-mail him (sayre.phil@epamail.epa.gov) site descriptions, including information on the site owner, location, size, current use characteristics, and contamination present. Sayre would like to have these site descriptions by Monday, April 13, 1998, so he can compile a list of potential field study locations for the next conference call.

Jackson will make another protocol draft incorporating all the RTDF Subgroup's comments and mail it to all participants.

Attachment A

REMEDIATION TECHNOLOGY DEVELOPMENT FORUM
PHYTOREMEDIATION ACTION TEAM
FIELD STUDY PROTOCOL