SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
PHYTOREMEDIATION OF ORGANICS ACTION TEAM
TOTAL PETROLEUM HYDROCARBON IN SOIL SUBGROUP
CONFERENCE CALL
May 1, 1998
11:00 a.m. to 12:30 p.m.
On May 1, 1998, the following members of the Phytoremediation of Organics Action Team, Total Petroleum Hydrocarbon (TPH) in Soil Subgroup, participated in a conference call:
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
Jim Brown, Roy F. Weston, Inc.
Pam Davis, EPR
Evelyn Drake, Exxon Research and Engineering
Peter Kulakow, Kansas State University
Ernest Lory, Naval Facilities Engineering Service Center (NFESC)
C.M. (Mike) Reynolds, U.S. Army Cold Regions
Henry Camp, Arthur D. Little (ADL)
Greg Douglas, ADL
Christine Hartnett of Eastern Research Group, Inc. (ERG), also participated.
OVERVIEW
Participants in this conference call discussed in depth the protocol (Attachment A) for soil sampling and analysis. ADL representatives Henry Camp and Greg Douglas were not originally invited to participate, but Phil Sayre contacted them during the meeting for input on several issues. The conversation focused on sample analysis, sample collection, statistician selection, sample storage, and field plot creation. By the end of the call, the RTDF Subgroup decided that investigators at each test site should initiate the following activities, in order:
SAMPLE ANALYSIS
Participants in the RTDF Subgroup agreed that samples must be analyzed for:
Choosing Analytical Laboratories
The RTDF Subgroup has not determined which laboratories will process their samples, but three were identified as possible choices: the ADL laboratory; the Batelle laboratory in Duxbury, Massachusetts; and the Kansas State University (KSU)/Purdue University laboratory.
Participants realized that they may need to choose more than one laboratory to get all required analyses performed. Ideally, they would like to select the laboratories before sampling starts so all samples are sent to the same laboratory. Samples sent to different laboratories would be more difficult to cross-compare because analytical methodologies differ significantly from one laboratory to the next. (Jim Brown noted that his laboratory characterizes oil using a compilation of methods, including a modified 8015, Coast Guard methodology, and Environment Canada methodology. Mike Reynolds stressed that the "nuts and bolts" of methodologies are often different between laboratories.)
Although all participants agreed that it is preferable to send all samples to the same laboratory, Reynolds noted that this could prove problematic for his group. As a government-sponsored organization, his group has limited control over where it can send money. Reynolds said that he thought it would be difficult for his group to give money to the ADL laboratory; for him, it would be easier to give money to the Batelle laboratory. (Reynold's group can transfer money easily to the Navy, and the Navy has a contract with the Batelle laboratory in Duxbury, Massachusetts.)
The ADL Laboratory
Lucinda Jackson noted that she and Evelyn Drake have used ADL extensively and have always been pleased with ADL's work quality. Participants asked various questions about ADL, including whether it is EPA-certified, what extraction method it uses, and how many of the RTDF Subgroup's analyses it can handle. Because no RTDF Subgroup member could answer these questions, Sayre called ADL's Henry Camp and Greg Douglas. The ADL representatives participated in the rest of the meeting and offered helpful input.
Jackson told the ADL representatives that the goal of the RTDF Subgroup's project is to determine the efficacy of plants for degradation of petroleum hydrocarbons in soils at multiple locations and across various climatic conditions. She detailed for the ADL representatives the analyses that are needed, then asked the ADL team members for some information on their capabilities. Douglas stressed that ADL specializes in organic analysis. He said that ADL can perform TPH, PAH, and biomarker analysis. He stressed that the ADL laboratory is well-known for its accuracy and extreme precision in detecting low levels of PAHs and biomarkers. Douglas emphasized the importance of using a laboratory that can provide accurate biomarker analysis because good data helps normalize data and minimize variability. Brown asked Douglas what biomarkers ADL uses for refined products. Douglas said that ADL believes diaterpanes offer the best results because they are prominent in diesel products and extremely resistant to degradation.
For other analyses (i.e., microbial and soil), Douglas recommended other laboratories. Douglas said the RTDF Subgroup could send all its samples to ADL and ADL would aliquot the samples, run the organic analyses in house, and ship the balance of the samples to other laboratories for microbial and soil analyses. Douglas said that ADL would be able to commit to the project for the 3 to 5 years Jackson cited. Jackson asked Douglas what laboratories Douglas planned to use for the microbial and soil analyses. Douglas said he would speak to Roger Prince (also of ADL) for some ideas. Jackson told Douglas that the RTDF Subgroup might opt to send ADL enough samples for the organic analyses and additional samples directly to the laboratories performing the other analyses. Douglas agreed that the Subgroup would find that alternative cheaper than sending all samples to ADL.
Douglas asked for a written outline specifying the scope of the project, the number of samples, and the required turnaround times. Drake agreed to fax a flowchart outline to him from Munich. Douglas said that he would prepare a cost estimate by May 8, 1998. Steve Rock planned to obtain a copy of the estimate by the next conference call. Douglas also agreed to fax a copy of ADL's biomarker analysis protocol to Sayre. (Douglas told participants that the protocol can be found in the Federal Register and in Batelle's 1994 Bioremediation Conference under the author name Douglas et al.)
The Batelle Laboratory in Duxbury, Massachusetts
According to Reynolds, ADL and Batelle are very similar. Reynolds said he is certain the two laboratories use the same methodologies and knows that the two have shared many of the same personnel throughout the years. For Reynolds, Batelle offers an advantage over ADL because it is easier for his group to send money to Batelle (see preceding). Rock noted that his group also has a contract with the Batelle laboratory in Columbus, Ohio; he suspected that it would be easy to transfer money between the Columbus, Ohio, and Duxbury, Massachusetts, laboratories. Rock agreed to forward an outline of the RTDF's scope of work to Batelle and to request a cost estimate.
KSU/Purdue University
As a Hazardous Substances Research Center (HSRC) laboratory, KSU is of particular interest to the Subgroup because the Technology Innovation Office's (TIO's) Walt Kovalick has indicated that he might be able to contribute some funds to an HSRC laboratory. (Sayre warned that Kovalick probably will not contribute a great deal of money, however.) Rock asked Peter Kulakow whether KSU could do the analyses. Kulakow said it was possible but that he must investigate the matter more thoroughly. He said he would contact Kathy Banks and Paul Schwab at Purdue University to see if they would be able to help. Between the two universities, Kulakow said that they would be able to perform all the analyses except PAH. (Kulakow noted that Paul Schwab is setting up a PAH laboratory but doubts Purdue will be able to offer the precision of ADL.) The RTDF Subgroup asked Kulakow to investigate the possibility of doing the microbial, TPH, and soil analyses.
Kulakow asked about the source of the funding. Rock said the funding would come from the RTDF Subgroup and maybe from Walt Kovalick.
Funding for Sample Analysis
During the April 24, 1998, conference call, Sayre noted that he was hoping to secure additional funding when he spoke to the deputy administrator for the Office of Solid Waste and Emergency Response (OSWER). Sayre met with this person in late April but was unable to request funding in the allotted time.
Methods Used to Obtain Bacterial Counts
The participants briefly discussed the best approach for obtaining bacterial counts. Rock noted that he
favors the Fatty Acid Methyl Ester (FAME) methodology over a plate-counting methodology because EPA
has a laboratory technician who is very skilled in this methodology. Rock noted that the technician runs
FAME using whole soil isolates. Rock acknowledged that petroleum interference could emerge as a
problem using this method. Reynolds said that his group has found FAME desirable rather than isolates on whole plates.
SAMPLE COLLECTION
Sample Size
Jackson asked the group whether a sample size had been determined. Sayre noted that the amount of sample required would be determined by the number of endpoints to be analyzed. (He noted that endpoints are listed on page 2 of the protocol.)
Splitting Samples
Jackson asked participants to discuss how many bags of soil must be collected at each sample location. Rock said that at least two bags would be required: one for microbial analysis and another for chemical analysis. Whether additional splitting would be required for chemical analysis would be determined by the number of laboratories involved. For example, if the Subgroup finds a laboratory that can perform all the chemical analyses, it could send one bag to it. If a number of labs conduct the chemical analyses, the sample would have to be split again. Jackson suggested collecting three bags per sample location: one for microbial analysis, one for PAH analysis, and one for all remaining. Kulakow agreed that this practice would speed results.
Collecting Samples for Site Characterization
Douglas and Brown emphasized the importance of collecting solid data during the site characterization phase of the program. This data is important because it is used to calculate site variability. Douglas stressed that the site should be tilled and homogenized as much as possible before the site characterization samples are collected. He noted that fewer samples can be taken to assess a site's variability if the site's soil is mixed well.
According to Brown and Douglas, pinpointing the variability at a site is crucial, because the variability determines how many samples must be collected from test plots during T0 and all subsequent sampling events. (Douglas told the group members that they may want to contact Roger Prince to discuss how he addressed a site with high variability.)
Many of the proposed test sites have been characterized. Participants agreed that it would probably be beneficial to recharacterize these sites, if possible. Reynolds said that, for him, it may not be possible. He did say, however, that he would investigate the possibility of sending one person to Barrows, Alaska, to collect samples. Reynolds noted that funding issues challenge his group. Douglas reminded Reynolds that understanding site variability up front can reduce costs down the line. Douglas noted that ADL could simply reprocess the original site characterization samples if the originals had been frozen.
In an attempt to find ways to reduce the cost associated with recharacterization, Ernest Lory recommended sending the characterization samples to a local laboratory. (Lory realized that T0 and subsequent testing need to be sent to laboratories that offer precise results but thought that less precise results might be adequate at the characterization stage. He stresses that the aim at this stage is to get "ballpark" estimates.) While some participants expressed interest in this idea, Brown and Douglas were skeptical. Douglas noted that he had sent samples to some commercial laboratories and that those laboratories failed to approximate ADL's results. Brown said that all data should be analyzed using one methodology. He noted that all efforts to normalize data should be made. He said that less expensive (and less precise) analytical methods could cause a statistician to erroneously recommend too many samples be collected during T0 and all subsequent sampling events.
The participants talked about how many samples would be required to accomplish the characterization.
Their numbers ranged from 10 to 20. Rock noted that the sample locations for the characterization study
must be selected randomly. Jackson asked Douglas whether ADL would be able to turn test results around
for the characterization samples within a week of receiving them. Douglas said that ADL could do this.
GETTING A STATISTICIAN'S INPUT DURING PROTOCOL DEVELOPMENT
A couple of weeks ago, Brown suggested involving a statistician in protocol development. Since then, Brown, Sayre, and Rock have talked to some statisticians. Those statisticians who were asked were most troubled by the:
During the May 1, 1998, conference call, Brown reiterated the importance of involving a statistician early in the protocol design process. The statistician to whom he has spoken to most is Charles Kufs, a former Roy F. Weston employee who is now an independent contractor. Brown described Kufs as a professional geologist and a high-level statistician. Brown has already talked to Kufs about the RTDF Subgroup's protocol. Kufs told Brown that the Subgroup must identify up front the type 1 and 2 errors that are acceptable and then establish the kind of differences that the Subgroup hopes to detect. Once these issues are resolved, Kufs said that a statistician should be able to use site characterization data to determine how many samples must be collected per plot for T0 and other sampling events.
The participants agreed that it is imperative for the RTDF Subgroup to find a statistician who can commit
to the project over the long term. Rock and Sayre have not been able to identify a statistician who would
be appropriate for the project. (Sayre has talked to some statisticians, but none specialize in field plot
design. He has gathered names of some statisticians from the Office of Solid Waste but has not yet
pursued them. The statistician with whom Rock spoke was a mathematician. Rock said the group would
be unlikely to find someone within the Office of Pollution Prevention and Toxics or OSWER who could
commit over the long term.) Brown thought that Kufs might be interested in the project. Sayre worried
that funds would become an issue if the group chose Kufs. Nevertheless, the Subgroup agreed that Kufs
should be contacted. Brown and Rock agreed to talk to Kufs before May 8, 1998. Assuming that ADL is
chosen to participate in the project, Douglas noted that it will be important for the laboratory and the
statistician to communicate to maximize data normalization.
SAMPLE STORAGE
Jackson noted that the group must determine how soil samples should be maintained and stored. She stressed that a decision regarding this matter is fairly urgent, because some RTDF Subgroup members plan to initiate projects in the spring.
Participants agreed that it would be best to ship samples directly to laboratories once those samples are collected. Brown said that laboratories extract samples immediately, then store the extracts indefinitely. He explained that extracts have no holding time. Jackson asked what should be done if samples are collected before the Subgroup agrees on which laboratories to use. She asked whether the group that collected the samples should hold them until the Subgroup decides. Rock said he would likely run into difficulty if asked to hold samples for too long, because he lacks the facilities. The question served to reemphasize how important it is for the group to choose laboratories quickly.
Participants asked the ADL representatives how they ship and store samples. Douglas agreed to send the RTDF Subgroup ADL's shipping protocol, because ADL ships samples all over the country. Douglas recommended cooling the samples to 4 degrees quickly once they are taken and shipping them to the laboratory on blue ice or dry ice. At the laboratory, Douglas recommended extracting the samples and running them immediately (but not quantifying them), freezing the extracts, and storing them for the duration of the program.
Reynolds stressed that the freeze/thaw cycle should be minimized. He recently read a paper that discussed the impacts of freezing and thawing on microorganisms. The paper concluded that fast growers dominate when a soil is thawed. Reynolds pointed out that freezing may have little impact on Alaskan soils because these soils are often frozen. He was unsure how freezing would impact soils from the southeastern and southwestern United States.
Reynolds said that he plans to start collecting samples soon and to run many analyses on "fresh" samples.
He offered to store some of the samples and to test them after a given period to compare results with
"fresh" samples. Reynolds noted that various storage techniques (e.g., freezing or air drying) could be
tested to see which yields the best results. Rock agreed that doing some preliminary testing might be
beneficial. Reynolds said that he will contact Scott Angle of the University of Maryland for storage
technique suggestions. (Angle has extensive experience performing soil bacteria counts.)
SETTING UP FIELD PLOTS
Plot Design
Brown asked participants if they have decided how they will design their plots. Jackson said that she plans
to place her blocks one next to the other, with untouched strips of land separating each. Reynolds said that
plot design will prove to be a large challenge for his group because the sites that he is working on consist of
a number of "hot spots" rather than a site that is uniformly contaminated. He suspects that his plots will
be set up in a "gerrymandering" fashion.
Weed Killer
During the April 24, 1998 conference call, participants asked whether Roundup (a weed killer) would
adversely impact microbial populations or interfere with soil analyses. Between conference calls,
participants followed up on this issue. Reynolds found some information indicating that Roundup will not
cause a problem. Likewise, Jackson talked to a microbiologist and was told that Roundup will not
adversely impact microorganisms. Sayre, on the other hand, was told that Roundup could cause a problem.
This information was provided by a statistician. According to Sayre, the statistician was so concerned
about the potential effects that she went so far as to recommend using a separate control. Jackson said she
does not want to set up a separate test control. Jackson said that hand picking is always an option, but
noted that the practicality of controlling weeds with this technique depends on how much help
investigators have at their site.
MISCELLANEOUS
Rock asked Lory whether he plans to establish some field plots this year. Lory said that he has the soils and the ability to set up plots at Port Hueneme but is researching the funding aspect. He and Kulakow will discuss whether it is appropriate to establish another test plot at Port Hueneme.
Douglas told the group about a site that could serve as a potential candidate for the program. The site is an 80-acre land farm and has already been characterized extensively. The variability has already been established at this site. The site owner has expressed interest in conducting phytoremediation at the site and might be willing to fund a test. Douglas was not able to give the RTDF Subgroup the site owner's name but agreed to contact the owner.
The conference call participants talked briefly about the IBC conference that is scheduled to take place in
Houston, Texas, in June 1998. Assuming that the Subgroup is able to extend contracts in time, Lory
recommended inviting laboratory representatives and the statistician to the meeting.
ACTION ITEMS
| — | Drake agreed to fax ADL an outline of the RTDF Subgroup's scope of work. Drake got off of the call before all of the communications logistics were ironed out. Rock said that
he would contact Drake and request a copy of the outline as well. After he receives the
outline, he will send it on to Batelle (i.e., Kevin McCarthy [phone number: 781-934-0571]) and KSU (i.e., Kulakow). |
| — | ADL (i.e., Douglas) agreed to prepare a cost estimate by May 8, 1998. Douglas got disconnected from the call before all of the communications logistics were ironed out. Rock agreed to e-mail the ADL representatives (camp.henry@adl.com) to ask for a copy of the estimate. |
| — | Rock agreed to lead the effort to find a statistician. |
| — | Brown agreed to call Kufs to let him know that Rock will be contacting him. Brown agreed to update Kufs on the RTDF Subgroup's progress and to ask him to pull together a cost and time estimate. (Kufs can be reached at 215-657-7469 [phone], 215-657-4156 [fax], or charliekufs@juno.com [e-mail]). |
NEXT CONFERENCE CALL
The next conference call is scheduled for May 8, 1998. The group agreed that it would be beneficial to try
to include Kufs in the conference call, but that it is not necessary to include the ADL representatives in the
next call. Douglas said that he will be available to answer any technical questions if the group wants to
contact him during the call.
Attachment A
Field Study Protocol
REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
PHYTOREMEDIATION ACTION TEAM
FIELD STUDY PROTOCOL
28 April 1998 Version
| TITLE: | PHYTOREMEDIATION OF PETROLEUM HYDROCARBONS IN SOIL |
| PURPOSE: | Determine efficacy of vegetating with agricultural and native plants for
degradation of petroleum hydrocarbons in soil at multiple locations and under
varied climatic conditions. |
| TREATMENTS: |
Note: In a tiered approach, treatment 1 or 2 could be dropped to favor the treatment that provides the best ground cover and dense root zone. For some test sites, an additional salt-tolerant control could be considered. The allelopathic effects of rye on other plants in the item (1) mixture should be considered. Legumes discussed for the item (1) mixture included alfalfa, clover, and birds'-foot trefoil. |
| LOCATIONS: | To be determined. |
| SOIL SAMPLING: | Take soil samples (unspecified amount) in each sampling location at the following times:
0: Before sampling (samples taken at 0 to 6" and 1 to 2') The soil sampling method and storage should meet specified standards. A 1"
diameter GEOPROBE "large bore" sampler is suggested. Place each sample in a
clean plastic sample liner; use hand sampling devices. |
| PLANT SAMPLING: | Take plant shoot samples at time 3 (30 months) for hydrocarbon analysis only.
[protocol not yet specified = Annex 1] |
| SAMPLE STORAGE: | After samples are collected in the field, they should be placed on regular or blue ice and shipped to the laboratory as soon as possible. The laboratory will
analyze the samples and should store them at liquid nitrogen temperatures.
Samples should be collected, handled, shipped, stored, and analyzed in ways that
optimize the soil monitoring endpoints in the next section. |
| SAMPLE ANALYSIS: | Send composite soil samples (after wet sieving) to the Environmental Monitoring
and Analysis Unit of Arthur D. Little, Inc. (ADL), for archiving and analysis of:
|
Note: In a tiered approach to sample analysis, initial soil composite samples are split after wet sieving, and labs other than ADL can do the analyses in 1, 2, 5, and/or 6 (using the same protocols as described by the RTDF) based on archived samples. Also in the tiered approach, ADL archives all samples (or portions of the split samples) and time 0 analyses. Time 1, 2, and 3 analyses may be delayed for later analysis of archived samples, based on the results of TPH analyses. TPH would be measured and, if no significant changes were seen (at 3-year point or earlier), further analyses as noted in items 1 through 5 would not be done. For those sites that show significant TPH reductions in soil, samples could be selected from the archives for full analysis. Finally, TPH analyses could be done at labs other than ADL, using ADL protocols. The bioavailability of TPHs following phytoremediation could be added as a seventh endpoint for analysis.
[Protocol = Annex 5; noted that extraction procedure would vary depending on soil conditions such as pH, etc.] |
|
| PLOT SIZE: | 20' x 20' minimum |
| REPLICATIONS: | 4 |
| STATISTICAL DESIGN: | Randomized complete block [plots selected based on presence of TPHs as analyzed before plot selection]. For soil and plant samples, take 4 random sample cores per plot and make a composite sample. [Details of statistical approach in Annex 6.] |
| GROWING CONDITIONS: | Conduct seed bed preparation, planting technique, planting rate, and
irrigation in site-specific manner to establish good stand growth. Record
procedures.
Fertilization is based on need from initial soil sampling. |
| PLANT EVALUATIONS: | Evaluate plant characteristics at 6 months, 18 months, and 30 months:
|
WEATHER CONDITIONS: |
Record rainfall and average daily temperature (available from local airport) throughout growing season. |
| SITE HISTORY: | Source of contamination Past site uses Previous cleanup attempts Climate |
Incomplete Annexes
| Annex | Status |
| 1. | Protocol for plant shoot sampling |
| 2. | Analysis techniques for soil-borne hydrocarbons (saturated, aromatic, TPH, individual PAHs) |
| 3. | Biomarkers (From ADL/Exxon; validation information on protocol?) |
| 4. | Below |
| 5. | Bioavailability protocol |
| 6. | Statistical approach for field test design/data analysis |
Annex 4: Analysis of Soil Bacteria
Protocols for Bacterial Analyses
Bacteria to monitor:Total hydrocarbon degraders, based on growth on a standard weathered crude oil. Note that this protocol accounts only for bacteria and therefore does not lead to enumeration of other hydrocarbon-degrading taxa, like fungi.
Soil sample size needed and treatment of sample before analysis: Sample size not yet specified. Sample handling should minimize freeze-thaw cycles on samples.
Hydrocarbon mixture to test with: A standard crude hydrocarbon that can be weathered in a standardized way. [Options: an Exxon or ?Amoco standard, an American Petroleum Institute Standard Oil; weathering protocol not yet discussed]
Protocols for bacterial counts: [exact protocol to use and costs not settled]
Kathy Banks' lab at Kansas State University (KSU) and Exxon have developed two apparently similar most probable number (MPN) techniques. Banks' approach that relies on diluting soil 10-9 in a Bushnell Haas broth (mineral salts broth lacking carbon source) and running 5 replicates of each dilution. A tetrazolium (redox) dye is added to each suspension and vortexed. The petroleum substrate is then added to each suspension, and the suspension is capped with an aerated cap. Controls of the dye and hydrocarbon are included to guard against contamination. Changes in dye color, which indicate microbial respiration, are recorded weekly until activity ceases. The number of positive responses at each dilution level are used for final calculations. MPN tables are used to determine the number of degraders in each soil sample.
The Exxon technique, which resembles Banks', uses the Bushnell Haas broth and multiwell plates. In this technique, each well contains 10 ul of weathered crude oil on top of the broth, and growth is assessed visually (sheen broken, growth noted, etc., "sheen-screen" method?). Four replications are used for each dilution. Heterotrophs are also enumerated in a rich medium but no dyes are used. Cloudiness indicates growth.
Microbe Innotec Labs uses an automated MPN technique that employs software developed by Al Venosa (EPA). This software automates the process for determining the number of heterotrophs and TPH degraders. Paul Flaffman of OHM is also familiar with this technique. Costs are estimated to be $68/sample for heterotroph analysis and $195/sample for 48-hour TPH degrader analysis.
The 1980 APHA/AWWA/WPCF reference book Standard Methods for the Analysis of Water and Wastewater, 15th edition, has a Method 907 that uses the same Bushnell and Haas agar and a hydrocarbon supplement to enumerate hydrocarbon degraders in soil. Some later editions of the text do not have the protocol, however.
Timing of Sampling for TPH Degraders, Identifying Number of Degraders Needed for Efficient Degradation
Hydrocarbon degraders often make up 10 percent of heterotrophs at a contaminated site. Usually, there are enough hydrocarbon degraders to affect remediation if time 0 counts are on the order of 104 TPH degraders per gram of soil.
Protocol currently calls for microbial sampling at time 0 (before fertilization) and time 3 (30 months). The time 0 sampling is intended to determine if sufficient degraders are present. The time 3 sampling is intended to see if degraders have increased in number due to rhizosphere effect (and, to a lesser degree, nutrient amendments) in phytoremediated plots. Further sampling between times 0 and 3 would (e.g., at the end of each growing season) would help determine if bacterial numbers have changed due to plant presence. Because TPH degraders tend to increase in numbers in the first few weeks after fertilization, samples should probably be taken 2 months after the last fertilization. All fertilization dates/application rates should be specified.