TAG reports: Volume 3, Number 2; June,2001
Brunswick Wood Preserving Supplemental Investigation Reports:
Over the past few months, the Glynn Environmental Coalition (GEC) has received several reports and responses from the Environmental Protection Agency (EPA) on the cleanup at the Brunswick Wood Preserving Site (BWP site) in Glynn County, Georgia. The principal report is the Final Supplemental Sampling Investigation Report (the “supplement”). The responses to community questions are referenced as the May 2001 EPA written response.
The BWP site is a former wood treatment facility contaminated with arsenic, chromium, PCB, dioxin, carcinogenic hydrocarbons (PAH), and many other chemicals. The site has undergone some remediation by soil removal through Federal and State emergency actions. However, many tons of highly contaminated soils remain, and contaminated groundwater is moving off-site as underground plumes. The EPA is entirely responsible under “Superfund” for cleaning up this site.
The GEC asked us to address six questions in regard to the
most recent reports. GEC questions are restated in bold text below. Some replies
also relied on the Final Report Phase III Remedial Investigation, December 2000,
(the “RI”) and on various EPA Region 4 Updates as referenced in the text.
1. Concerning the June 2000
sampling of sediments in Burnett Creek.
a.
What
was the depth of the samples? (Was it a 3-foot core that was mixed? Has
the EPA provided the report?)
|
|
b.
Were the samples representative of what a person might be exposed to by
walking in the creek?
c.
Could actual levels of contamination in the surface sediments be much
higher than were reported by the EPA?
d.
What
chemicals in Burnett Creek should we be concerned about, why, and do we have
sufficient information about each?
Replies to questions 1.a. and b.: We cannot answer the GEC’s questions with the information EPA has provided. Studies for this site are both poorly designed and poorly reported. Sediment sampling data at the BWP site is probably not reproducible (the basic goal of all professional laboratory and field research is “reproducibility”—the ability to perform the same operation and achieve the same result). The Final Supplemental Report describes eleven “grab” samples taken from Burnett Creek in November of 2000 from the bridge on Old Jesup Road downstream to the bridge on New Jesup Highway. The only sampling description was that ten “…samples were collected during low tide from a canoe along the bank of the creeks using stainless steel scoops.” One sample was collected while “wading.” No descriptors, such as where in the stream, or on which stream bank, or how deep, were provided.
Another ten samples were collected in July of 2000 and referenced in the RI. Eight are from an area beginning about 500 feet downstream from the former outfall pipe at the intersection of Burnett Creek and Perry Lane Road down to the bridge on Old Jesup Road. One was from further downstream, and another, from a tributary. Little information is provided about these July samples. The description indicates two are “grab” samples; the remaining eight are “comp” samples. We could not find a key to these sample descriptions, but “comp” usually means “composite” samples. However, whether these are area or depth composites cannot be readily determined from the information provided for review. Sample compositing just means sample mixing. Area samples are composited by taking several samples from the same depth level but over a large area-- 5 samples in a 20 square foot area all from the top 6 inches of the ground for example. Or a sample may be composited from depth samples taken at a single spot—mixing a 3-foot deep core for example. It makes a great deal of difference in interpreting the results. Also, it is not known if the “grab” samples are surface samples or not—most often grab samples are from depth. Generally, the sample descriptions supplied are not of any practical value. There are no sample logs that we could locate and no descriptions of depth cores or sample sizes. Of the 21 samples, the “wading” sample could be representative of what a person might be exposed to by walking in the creek.
Reply to questions 1. c. and d.: To answer these questions requires some speculation on our part. Whether or not we have confidence in the data influences any answer. We pulled sediment data out of the reports (Table 4-35 of the RI; Appendix E of the supplement) and plotted select hydrocarbons and dioxins. These compounds are carcinogenic in nature, and affect both people and the environment. The results are shown below in Figures 1 and 2 and on Map 1.
It seemed likely that the closer the sediment samples are taken to the site, the higher the levels of chemicals would be found. However, looking at several site chemicals this trend did not occur. Instead, highest levels of these chemicals occurred in sediments further downstream from the site, about midway between the bridges of Perry Lane Road and Old Jesup Road. Either the study was flawed and underreported some areas, or large amounts of chemicals are moving. These are very different scenarios—like light and dark, actually—and there is no approach from the information provided to even guess which is the true trend. The odd sample numbering suggests different studies developed on a “make- it-up-as-you-go-along” approach, rather than a careful look at environmental trends. The cavalier description of the “canoe-trip” sampling event does not establish a professional approach (the description reads literally that “…samples were collected… from a canoe…” rather than samples were collected from a stream). Further, there is no statistical basis for any decision-making using data such as these. A better approach to sampling would be to take a cross section of six samples at a marked location. One surface sediment sample and then sampling again down a foot deep, and taking these surface and subsurface samples on the left bank, midstream, and right bank, for a complete sampling cross-section at each station at intervals along the stream bed. Then there would be no ambiguity over any trend.
The sediment sampling in the RI and supplement would not be
reproducible—an independent researcher could not, using the information
provided, return to the same places and collect the same samples. As described,
the studies barely meet a standard of minimum effort, more typical of a high
school science fair project, rather than field-style studies aimed at
publication in a scientific journal. Overall, it is possible that higher levels
of toxic chemicals do exist that can affect the environment. At least, it cannot
be ruled out by these observations.
2. Concerning the May 7, 2001, Final Supplemental
Sampling Investigation Report.
a.
What
does "... evidence that groundwater to the west of Burnett Creek is
relatively unimpacted" mean (Section 3.2, last sentence of second
paragraph)?
|
|
b.
Has the west side of Burnett Creek been impacted by migration of
contaminated groundwater?
c.
If so, has the EPA provided any information about how fast the
groundwater
contamination is spreading?
d.
Should the Fate and Transport section the EPA has deleted from the RI be
included, and do we know what the fate and transport of chemicals from the site
from both above and below ground?
e.
Were the fish sampled for dioxin of sufficient size for the EPA to come
to the conclusion that there is no risk?
f.
Were the fish species tested the size and species that people eat?
g.
If
not, can higher than reported levels be reasonably expected to exist in fish
people are eating from the creek?
Reply to Question 2. a.: The phrase “relatively unimpacted” is meaningless. Chemicals are present, or they are not; chemical concentrations threaten, or they do not. Chemicals are found in wells on the western side of the creek, therefore the groundwater is impacted. We may debate the extent and relevance of the threat, but using bureaucratic slang to hide threats from the public is unconscionable.
Reply to Question 2. b.: The Final Report Phase III Investigation contains several Figures (4-2 through 4-9) detailing plumes of PAH, pentachlorophenol, and BTEX chemicals moving from the site and terminating at the eastern bank of Burnett Creek. The Executive Summary of the RI states on page ix: “Sampling west of Burnett Creek, however, indicated that the plume has not migrated beyond the creek.”
Note that:
a. Figure 2-1 of the RI shows only two soil sample locations west of the creek. This is not sufficient to draw significant conclusions.
b. Figure 2-2 of the RI shows no monitoring wells west of the creek.
c. RI Figure 2-3 shows two Geoprobe™ sampling locations 205B,C and 204B,C. However, these do not appear to be from the “west bank” of the creek but, instead, are upstream from small tributaries that lead into the creek. Further, 205 is on the very north edge of the plume boundary while 204 is actually out of the southern-most plume boundary. Based on the sample descriptions, these likely are not indicative of west bank chemical plume monitoring.
d. RI Figure 2-4 shows ten residential wells on the west side of the creek; however, only two of these-- “02” and “01”-- are within the path of the off-site plume and only “02” is fairly close to the creek.
e. RI Figure 2-5 shows six samples taken from Burnett Creek along the creek bed south of the site. These samples are designated PP01 through PP06. However, only PP04, PP05 and PP06 appear to be relevant to the path of the plumes, and only to the southernmost edges of the plume. These samples are not on the western shore, but actually within the stream itself.
The text refers us to Table 4 and Appendix F of the RI report for the actual data from these sampling events. A careful search of Table 4 found data for monitoring wells, soil samples, and sediment. Table 4-14 shows 5 of the 9 residential monitoring wells actually had low parts-per-billions of site organic chemicals (05, 06, 07, 08, and 09). However, no surface soils or subsoil sample designations in
Table 4 indicate extensive testing from the west side of Burnett Creek and data for the six creek samples could not be matched with the descriptors on the maps. Further, Appendix F, section “Complete Data” bears the phrase “Reserved” and is empty.
The sole sample that is unambiguously for west bank contamination evaluation is “Location 6” from the supplemental sampling investigation. Only one groundwater sample was taken at 38 feet. Paragraph 2 of section 3.2 contains the statement: “Only trace quantities of PAH’s (fluoranthene, naphthalene, and phenanthrene at 2 and 3 mg/ml) were detected…” Why there are three compounds quoted, but only two analytical data points, is not obvious. In addition, we could not locate a groundwater analytical report in Appendix D for these samples, so this discrepancy remains a mystery. Nor could we locate data for the soil samples taken with this sampling event. If actual laboratory soil analysis was done, we could not match a report to the sample designations on the log provided us for review. It appears that only visual and olfactory (sight and smell) analyses were performed. If so, we can only remark that “sight” analysis is biased and unscientific since most chemicals do not stain earth, and even those that do usually will not at their lower toxic levels. Further, environmental “smell” analysis is only used in the immediate vicinity of leaking underground tanks for highly contaminated soils, otherwise smell is not a valid analytical test, except for some highly trained dogs.
Examining the Remedial Investigation III Figures 4-2 through 4-9 detailing the off-site chemical plumes there are two distinct possibilities to explain the figure plots:
a. The chemical plumes stopped at the creek; or,
b. The investigation stopped at the creek.
Both scenarios will give the exact same plot as shown in RI figures 4-2 through 4-9.
In response to the GEC’s questions, we are unable to confirm statements in the RI that groundwater contamination stops at Burnett Creek’s eastern shore. Specifically, within the Final RI Report, Appendix, and Supplement, we can find no actual sampling support for the conclusion: “Sampling west of Burnett Creek, however, indicated that the plume has not migrated beyond the creek.” We recommend the GEC request the Remedial Project Manager specifically compile the data supporting this conclusion and make the model available.
Reply to
Question 2.c.: EPA has not provided any specific calculation of flow that
would allow determination of contaminant velocity. It appears there may be
sufficient data present in the report to estimate the flow with some accuracy by
computer modeling. It does not appear this has been done. This should be
performed before choosing any remedy. The RI is incomplete without this data.
|
|
Reply to Question
2.d.: In the correspondence between EPA and GEC, dated May 30, 2001, the EPA
acknowledged deleting this section from the draft, but offered no explanation
given as to why the Fate and Transport Section was deleted. There are a number
of criteria any cleanup must meet, including a requirement that a remedy must
“reduce the amount, toxicity or movement” of toxins in the environment. It
would be impossible for anyone, at the EPA or outside, to evaluate for this
requirement without a model of fate and transport. The RI is incomplete without
this data.
Reply to question 2.e., f., and g.: The background of this question starts with an EPA Update, dated March 2001. This EPA Update concluded that seafood from Burnett Creek was safe. Later, an email from the EPA Remedial Project Manager dated April 18, 2001 requested we reference the EPA’s March Update in the GEC newsletter review. Unfortunately, the actual data used in reaching these conclusions were not sent for review and the GEC formally requested the Supplemental Sampling Report. Additionally, EPA revised the reporting of the confusing mullet categories in a recent update, dated May 2001.
The presence of a chemical in a fish, or other water animals, is usually called “bio-accumulation” and occurs from two very different processes: bio-concentration (direct uptake from water) and bio-magnification (increase in chemical concentration from food sources). Over twenty different conditions have been identified which affect bioaccumulation. These include such things as solubility and half-life of the chemical in water to the size and position of the fish’s gills. Several factors do dominate any calculation, namely the age of the animal, its size and growth rate, and the stability of the chemical once it’s in the food chain. At the ends of the scale: an older larger slow-growing fish will show very high levels of a stabile chemical; a small juvenile fast-growing fish will not show the same chemical. Using fish to determine water quality is not an exact science. Nor, for that matter, is risk analysis. Unfortunately, the EPA has made a public risk proclamation regarding the safety of fish in Burnett Creek using these two inexact sciences. Paraphrasing the GEC’s questions into two issues:
a. Is the EPA’s risk assessment published in the March 2001 Update a bona fide risk analysis with the data obtained so far? “No” for many reasons:
a. There seem to be no guidelines for converting dioxin levels in juvenile mullet to health-based risk. Without an accepted consensus scientific model, any conclusions for or against safety are based on the specious reasoning of the person performing the calculations.
b. There are no guidelines for determining how much fish or shellfish tissue is sufficient for a risk analysis. For example: is a pound of whole juvenile tissue equally representative to an ounce of adult fillet? EPA provided us with no rationale for their decisions. This is very unprofessional; we cannot accept conclusions without appropriate scientific references.
c. Although mullet are edible fish, the mullet in the study were not really “keepers.” Size was not representative of fish for human consumption. Table 2-3 on Page 2-7 of the Supplemental Sampling lists just ten fish, averaging no more than 7 inches in length. The average weight was about 2 ounces. Tiny fillets were taken from these little two-ounce fish and all added together for analysis. The remaining carcasses were also mixed together and analyzed. It would have been better to keep the fish intact, since there was just not enough fish for even one study. A fillet usually means taking an oval portion from each side from behind the gills to the tail region and from the top (dorsal) fin down to the belly (ventral). Nowhere is it described how much meat was actually analyzed, but it couldn’t have been much. The skeleton, organs, head and tail would have been in the carcass area and accounted for most of the weight. We are guessing, since the report provided no detail, but based on past experience with these kinds of studies, each fillet from these little fish was probably a sliver a couple of inches long and maybe half an inch thick. It’s doubtful if that was really representative of human consumption.
b. Can anything be concluded from studies so far?
a.
This study found a site chemical, dioxin, in off-site fish and shellfish
tissue. None of the species tested were very far up the food chain (mullet are
grazers - vegetarians); none were very old in terms of growth. Very little
tissue was analyzed. The stations were far down stream. However, dioxin was
found in mullet carcasses, blue crab, and mummichog fish (small minnows). The
real significance of these studies lies in the fact of finding site chemicals
far down the creek. A study done with more scientific basis and less
editorializing may likely find a problem in this creek.
3.
Concerning onsite contamination.
a.
If there is not much contamination in the soil near the ponds on the
southern border as the EPA claims, why is there a visible oily sheen on the
water?
b.
Has the oily substance floating on the ponds been tested?
c.
If so, what is it?
d.
Where
is the oily substance coming from?
Reply to Question
3.a.: Short answer: maybe, depends on the “sheen.”
Long answer: A “sheen” is actually an interference pattern caused by
bending of light as it passes through the insoluble chemical layer floating on
the water. Depending on the chemical this layer may be only one molecule thick
and still produce the characteristic sheen. If it is a monomolecular film, it
may not be toxic and may not show up in a standard sample. At this site, the
sheen is likely from a mixture of hydrocarbon chemicals.
Reply to Question
3.b., c., and d: Surface water and soil samples contain mixtures of several
chemicals that could produce sheen on the water. These types of samples were
analyzed and levels of different chemicals found. Nearly all of the BTEX and
PAH’s can make this sheen and the area contains enough of the mixture to
produce the sheen. It will be interesting to see how EPA handles this area
during any cleanup.
4.
Please evaluate the quality of the data produced by the EPA over the past year. Would
this be acceptable to the EPA if it was a PRP led site?
Reply to question 4:
Most sites are “PRP-led” sites, meaning a chemical company is
responsible for cleaning the site (PRP is short for “potentially responsible
party”). The EPA is directly cleaning up the Brunswick Wood Processing site
since the original owners went bankrupt.
We have worked at PRP-led sites where the planning, testing and cleanups were performed so well our review consisted of the single phrase “good work!” We have also worked at PRP-led sites where the EPA-approved testing plan evoked the word “goofy.” We have not worked on a site, other than BWP, where there was no interaction in the planning phases and such a hostile refusal to provide data to support published conclusions. The drifting direction and poor data quality appears to be a unique characteristic of this EPA-led site.
Some responses from EPA to the Glynn Environmental Coalition in a letter dated May 30, 2001 and signed by the Remedial Project Manager would definitely not have been considered responsive if written by a Potentially Responsible Party. When asked about the boundaries of OU-1 the EPA merely said, “…the scope of OU1 has been expanded…” which is not an answer. When asked if the hydrological implication of closing the outfall pipe to Burnett Creek would be addressed in a future study the EPA merely said “At this time the only field investigation planned is for sampling residential wells…this summer” which also is not an answer. These and other responses seem deliberately and needlessly evasive.
By the way,
the well written, properly executed, and fully documented RI’s that elicited
the “good work!” reviews cost much less than the “goofy” plans, and were
completed faster, required no retesting and led to quick and safe cleanups.
5. Is there sufficient information to complete a Feasibility Study? If not, what information is missing?
Reply to Question 5: We have reviewed several draft BWP Feasibility Studies and parts of studies. None have adequately described the Operable Units, indicated the effects of any cleanup on chemical fate, or provided any insight into how a cleanup would meet the legally mandated CERCLA/SARA cleanup requirements. In response to several questions asked by the GEC on site borders and plume transport, the RPM only provided non-responsive “circular” arguments for the GEC to go back and re-read the RI. The real question should be “can the EPA support a Record of Decision (ROD) with what it has now?” And that reply is an absolute “no.” Too many conclusions without supporting facts, plus loose studies, poorly defined or described borders, and no clear short- or long-term goals characterize the Remedial Investigations at this site. Any Feasibility Plan based on these shaky RI foundations would be biased and untenable. All of the data from the multiple RI’s and small independent studies need to be placed into a single document by someone that understands data and statistics. The inaccuracies in the studies need to be closed using verifiable scientific methods by real engineers. The entire creek sediment and water studies should be repeated by someone actually qualified in aquatic systems. The groundwater studies should be correlated and a genuine model produced. Only then can this site move forward.
6.
Did the EPA answer the questions in the three letters?
Reply to Question 6: No.
Jump to top of page
Return to TAG table of Contents