PUBLIC HEALTH ASSESSMENT
TERRY CREEK DREDGE SPOIL AREAS/HERCULES OUTFALL SITE
BRUNSWICK, GLYNN COUNTY, GEORGIA
Public comments on Terry Creek Health assessment, Set #I:
I-1. Comment: The Summary is very confusing in the many terms used to describe the contaminants of concern at the site. There are over 670 chemicals of concern that comprise the mixture of polychlorinated camphenes manufactured by Hercules under the brand name of toxaphene. Rather than discussing "components of technical toxaphene," "recent toxaphene," "breakdown products" of toxaphene, and "a chlorinated-hydrocarbon pesticide," simply state that Hercules manufactured polychlorinated camphenes (PCC) for use as a pesticide under the brand name of toxaphene that is a mixture of over 670 chemicals, and these chemicals and their breakdown products are the chemicals of concern at the site. Thereafter, the contaminant of concern should be referred to as PCC. Furthermore, the concern at the site and in the PHA is the toxicity of PCC and the breakdown products, rather than the gas chromatography matching the peaks of the manufactured PCC, so it can be called toxaphene.
Response: USEPA used the term "chlorinated camphene containing 67 percent-69 percent chlorine" to describe the insecticide toxaphene when the tolerance for toxaphene residues was established in 11/25/1971. [FR 36 (228):22556.] Because the tolerances for pesticide residues considered the possible residues of pesticide chemicals or their conversion products, polychlorinated camphene (PCC) is used, as much as possible, in this health assessment as the name for the residues of toxaphene mixture and its conversion products.
I-2. Comment: Page 2, A. Site Description and History.
The statement, "It is a mixture of at least 177 individual compounds...," should be changed to the ATSDR Toxicological Profile description of, "... is a manufactured insecticide containing over 670 chemicals."
Response: Theoretically, over 679 compounds are possible in this mixture. About 177 are significant members, which can be seen on the most modern machine for detecting the chemical compounds in the PCC mixtures.
I-3. Comment: Page 3, A. Site Description and History.
The statement, "Contaminants were found as far as 3 miles away in the indigenous oysters at St. Simons Sound," should be changed to, "Contaminants were found as far as 7 miles away in the indigenous oysters in Jekyll Sound".
Response: Phillip A. Butler (1973) reported that 64 out of 65 samples of oysters collected during 1967-72 from St. Simons Sound contained detectable levels of PCC, with the maximum concentration at 7.5 ppm. At Jekyll Island, 37 out of 62 samples were contaminated with PCC up to 3.5 ppm. The maximum concentration was further reduced to 1 ppm at Satilla River24. The statement in the health assessment has been changed.
I-4. Comment: Page 3, A. Site Description and History.
The discussion of ATSDR's review of nine yard samples ranging from 0.055 to 1.83 ppm for the Georgia Department of Natural Resources concluded that these levels did not pose a short- or long- term health threat to the residents. These conclusions are suspect for the following reasons: (1) There is no discussion of the quantity of PCC the population was exposed to before there were over two half-lives of decay, (2) There is no discussion of the quantification of the PCC that still matches the chromatogram for "toxaphene" and the PCC that is not quantified, (3) There is no discussion of the cancer risk being above the 1 in 1,000,000 benchmark of 0.6 ppm, and ( 4) The sampling conducted in 1994 that found levels in the yards up to 14.0 ppm (2810 Hopkins) are not discussed or analyzed, nor is the 14.0 ppm (Stonewall St.) detection discussed on page 19 of the PHA. This section appears to be written on "pick and choose" data selection.
Response: The concentration of PCC in soil at 1.8 ppm should cause no non-cancer health effects. It is also only 3 times above the conservative screening guideline of 0.6 ppm for cancer risk at 1 in 1,000,000. Nevertheless, ATSDR in recent years found large uncertainty in the reported PCC level in soil samples for a general area (e.g., school yard). Reported concentrations from government agencies and community groups differed significantly (e.g., about 20 vs 330 ppm). The reported concentrations could have varied due to several factors: the skills of the chemists, exact locations for the samples, sampling design, the number of samples taken, the efficiency of PCC extraction from soil samples, the procedures of chemical analysis, and the method for calculating the concentration. The task for determining the PCC is not a simple skill. Data quality differed among data sources. Usually, the end users did not receive detailed reports from the analytical chemists. ATSDR will update the health assessment when reliable new information becomes available.
I-5. Comments: Page 4, B. Site Visit.
This section should be rewritten or deleted.
The discussion of the Riverside area is based on erroneous information. Back River does not flow south. The estuary is tidal and, by definition, flows in both directions. Therefore, the conclusion that the stretch north of the confluence of Terry Creek and Back River is likely to be unaffected is in error. The Skidaway Institute did a study of tidal flow with dye in Terry Creek and the surrounding area in the 1970's. I strongly suggest that this study be consulted.
The docks in the Riverside area are not all "new." Photos from 1977 show docks existed in the Riverside area.
Response: The statement was revised according to the 1974 Skidaway study. The plant effluent, assuming it followed the same general flow as the dye, was carried to St. Simons Sound by the receding tide, diluted considerably, and returned with the flooding tide5. It can then be detected over a much wider area than at the time of the first high tide. The concentration of the dye at a private dock in the Riverside development was comparable to the concentration at the toll bridge on the Torras Causeway over Back River and the concentration at the dock at the Georgia Department of Natural Resources. The concentrations in these stations were much lower than the dye concentration at the junction of Terry Creek and Back River.
Similar to the dye that reached the Riverside development, PCC was found in fish near an upstream branch of Back River (Quadrant #2) near the Riverside area in 1972. The PCC concentration at that location were 10.4 ppm in Fundulus killifish and 35.6 ppm in mullet muscle at that time (30).
I-6. Comment: Page 5, B. Site Visit.
The statement, "A vegetated dike surrounds the dredge spoil," is not supported by Figure 4 of the PHA or by personal visits to the main dredge spoil area. My experience at the site has been a very poorly vegetated dike, if any vegetation at all, with very friable soils exhibiting numerous areas of severe erosion.
Response: This statement was revised. Salt from salty seawater accumulated on top of the dike after the water evaporated. The high concentration of salt on top of the dike probably discouraged the growth of some kinds of plants.
I-7. Comment Page 5, C. Demographics, Land Use, and Natural
In paragraph one, the statement that, "Small boats are used in the creeks on an infrequent basis,", should be replaced with, "Small boats are used in the creeks on a frequent basis." I have no idea where you got this statement. The tidal creeks in Glynn Country are used extensively for commercial crabbing, bait shrimp trawling, recreational and commercial shrimp cast netting, recreational fishing, recreational crabbing (particularly around the site), and pleasure boating/canoeing/sailing.
The statement, "Marsh vegetation is primarily smooth cordgrass and needle rush," should be replaced with, "The marsh vegetation is primarily Spartina alterniflora."
Contrary to the contention that Riverside Subdivision and the Terry Creek Mobile Home Park are near contamination from dumping and dredging of waste, both have been shown to be contaminated. Unknown is whether you can state with certainty that it is from dredging activities. This section needs clarification to reflect the contamination and uncertainties.
When the survey was conducted, sex, and age of the "fishing" population surveyed would be helpful. You would not expect the fishing population to be present from 8:00 a.m. to 4:00 p.m. during the week. Without more information about the survey, the results should be suspect. This is far below the national average for fishing participation. (Strange. All those docks and boats, and only 16% do some fishing?)
Response: The statement on the frequency of the boating activity was revised. An ATSDR staff member (S.C. Tsai) made an unofficial stop at the site on the Memorial Day Holiday (5/29/01) and saw two boats on Back River at the bridge of F.J. Torras Causeway. There were about thirty people fishing on the old bridge and at the park near the intersection of the causeway and Highway 17.
The scientific name of the smooth cordgrass, Spartina alterniflora, is added after the common name. For the survey of fishing population, the age and sex factors were added in the wording in the text.
I-8. Comment: Page 6, D. Health Outcome Data.
The last sentence states that no computerized or tabulated health outcome data have been identified. The school system does keep data on learning disabilities that can be queried on an address basis. Since PCC is a known endocrine disrupter, this data on transgenerational effects would be helpful and could be compared to county-wide data. A mile or two radius around the site would be a good study area and sufficient in size for meaningful results. In addition, other school systems have similar data bases for comparison on the county level.
Response: These suggestions will be taken into consideration when ATSDR prepares the updated version of this health assessment. Endocrine disrupters were documented in the toxicological profiles for DDT, PCBs, and dioxins. The 1996 toxicological profile for toxaphene, however, lists no hormonal effects in its developmental effects section. ATSDR is aware of some reports on the hormonal effects of PCC, and will consider the inclusion of this topic when sufficient, consistent reports are accumulated in the database and quality review papers become available.
To analyze the health outcome data, a clear cause has to be identified. Although a complete pathway of PCC exists from contaminated fish to local residents, the methodology to identify the fish eaters in a school population is not well defined. The end point of learning disabilities could be affected by confounding factors such as socioeconomic status, besides the ingestion of PCC in contaminated local fish.
I-9. Comment: Page 13, Environmental Contamination and Other
The level of Toxaphene that is acutely toxic to fish is in the parts per trillion range (50 ppt) instead of the parts per billion cited.The half-life of PCC is mentioned as 10 years, but no discussion is offered as to the levels that can reasonably be expected to have existed at the site during deposition of the existing contamination. Air exposure and air transport of PCC is not discussed, even through ATSDR and USEPA consider this one of the most chronic problems associated with PCC. Several things about air transport need to be discussed. First, not all of the 670 peaks, or chemicals, would be expected to evaporate and be transported at the same rate or in the same environmental conditions. Therefore, any measurement for "toxaphene" and the correct peak ratio would be meaningless when trying to detect airborne constituents of the manufactured mixture.
Response: In the flow-through systems for testing acute toxicity of PCC to fish, the 96 hours LC50 values of PCC were 1.1 ppb for sheepshead minnows Cyprinodon variegatus and 530 ppt for pinfish Largodon rhomboides in salt water (Handbook of Environmental Data on Organic Chemicals, Karel Verschueren, ed., 1983). At these exposure levels, 50% of these saltwater fish died in 96 hours. The acute LC50 values for other kinds of fish ranged from 2 ppb for basses to 18 ppb for bluegills. PCC in chronic exposure systems were one to three orders of magnitude more toxic to fish than were acute exposure systems. The chronically toxic effects of PCC were observed at 39 ppt in brook trout, and at 36.7 ppt in fathead minnow21.
The half-life of PCC was about 10 years in an aerobic, surface soil (sandy loam). The primary pathway for the declined PCC concentration could be the vaporization from soil to air. This situation is different from that in the marsh sediments. Those PCC members with high chlorine contents dechlorinated rapidly in the anaerobic sediment into less chlorinated daughter PCC members. The half-life of the degraded PCC mixture in sediment is unknown. Some of the degraded products are stable in the marsh environments. Data are unavailable for their transport and transformation. At the local scale at this site, ATSDR did not receive any numerical data on the concentration of airborne PCC. ATSDR received some qualitative notices from USEPA during the removal action. These preliminary results were for air measurements at different sampling dates. Each report was a single paragraph e-mail notice informing ATSDR that USEPA did not detect any PCC in the air, with an attachment of some pictures taken at the site. ATSDR might be able to find some uses for these notices when more specific information on the detection limits and the detailed description of analytical methodology are made available to ATSDR. From the attached picture, a salt crust on the dried soil was observed. The salt sheet over the soil might suppress the dust formation at the windy site and reduce the potential PCC particulate in the air. On 8/2/2001, ATSDR received an e-mail and learned of the detection limits of 2.5 to 6.6 (or 5.5) ug/cubic meter. These values were revealed in the letter sent from the Remedial Project Manager of this site to Mr. Lindquist of the Lake Superior Binational Forum. These detection limits are the only numerical data ATSDR has about the airborne PCC at this site. Without the information on the analytical methodology, the definition of "toxaphene" in the report, and the correct peak ratio, ATSDR does not know the exact meaning of these detection limits.
Response: The statement was revised. The surface water flows east and south during receding tide and comes back in the opposite direction during flooding tide. The water in this tidal basin changes direction in each tidal cycle.
I-11. Comment: Page 15, Sediment.
Change the statement, "In 1997, toxaphene concentrations in sediments of Dupree Creek and Terry Creek ranged from approximately 7.9 to 230 ppm" to "Sediment concentrations were measured up to 33,000 ppm in 1998."
Response: The 1998 data of 0.62 to 33,000 ppm in 31 surface sediment samples at the Outfall Ditch were added to the text. The reference is GeoSyntec Consultants, Site Status Report, Revision 0, Terry Creek Site, Brunswick, Georgia. Prepared for Hercules Incorporated 38. 1998.
I-12. Comment: Page 16, Historical Data (1970s).
Change the first sentence of paragraph two from "St. Simons Sound," to "Jekyll Sound." See comment number 3. Average Fundulus toxaphene levels were substantially higher than "around 50 ppm " between 1970 and 1972. It is truly amazing how high the levels of PCC were in the fish. Several fish species developed that were able to tolerate high levels of PCC (this is in literature). "A mortal meal" was literally possible back then.
Response: Information on PCC contamination in Jekyll Island was added to the text. The average concentrations of PCC in killifish at this site were 62, 40, 10, and 50 ppm as shown in Table 2. The average concentration of PCC in killifish sample of September/October 1972 was 50 ppm with the range of 3.4-217 ppm.
In the toxicological literature, LC50 values of PCC to long-nose killifish Fundulus similis were 28 ppb for 48- hour exposures and 0.9 ppb for 28-day exposure (21). Toxicity data of PCC to killifish at this site are unavailable and the degree of the PCC resistance in local killifish is unknown. PCC resistance in other fish has, however, been documented. The LC50 value for susceptible mosquitofish (Gambusia affinis) was 12 ppb after 48-hour exposure. This LC50 value was 459 ppb for the resistant mosquitofish, which was 38 times more resistant than the susceptible strain (see the 1983 handbook by Karel Verschueren, cited previously).
I-13. Comment Page 16, Recent data (1990s)
Strike the section about the 1995 Hercules fish study or add the following explanation to the PHA. Hercules cites a very flawed study conducted under a Consent Order from the EPD to conclude, "These studies have not identified any adverse ecological impact that can be attributed to the facility effluent." Many studies are available that directly contradict the Hercules contention. Toxaphene exists in food fish above FDA allowable levels. A review of the study cited by Hercules follows: The USEPA had found toxaphene, a banned pesticide, in fish at 19 to 27 parts per million (ppm), which is above the FDA action level, while Hercules found none in their fish testing. The testing of fish conducted by Hercules was in compliance with a Consent Order (EPD-WQ-3129) negotiated because, "The Director (Georgia Environmental Protection Division) alleges that Hercules was in violation of the average daily toxaphene limit for six months, covering a period of time from February 1994 through April 1995." Regarding fish testing, the Consent Order stated, "The primary purpose of this study will be to determine whether toxaphene discharges are affecting the edibility of fish, such as mullet, utilizing the current federal guidelines for permissible toxaphene residues in fish."
The final report for the collection and analysis of fish and crab tissue from Terry and Dupree Creeks revealed the following:
- The species specifically identified for testing--mullet--was not tested.
- Only ONE fish sample was tested.
- The ONE analyzed fish sample had a five-fold increase in detection limits.
- The ONE fish sample was from the sampling station furthest from the plant outfall.
- No fish or crab samples were taken from the sampling station near the plant outfall.
Hercules hired LAW Engineering and Environmental Service, Inc. (LAW) to do the sample collection and final report. Prior to sampling, a final study design and contingency approach for conducting an assessment of fish and crab tissue was presented to the EPD and accepted. In essence, a plan for failure to find mullet had been submitted. Sampling was done with crab traps and horizontal gill nets. The net mesh size used was 2.0-inch bar or 4-inch stretch mesh (SM). Anglers who target mullet usually use 3 ½ to 3 ¾ SM nets, so failure to catch the target species, mullet, was virtually assured by their ability to pass through the net. This size net is good for sheepshead, which they did catch but discarded. In addition, the final report indicates the nets were never set at the sampling station near the plant outfall and crab sampling was abandoned after consultation with Hercules' technical advisors while two days of sampling effort took place at the sampling station farthest from the plant for the one spotted sea trout composite sample. During the four days of sampling effort, four samples were collected (3 crab, 1 fish) from three of the four sampling stations around the plant, and one crab sample from the Belle Point area.
Analysis of the fish and crab samples was done by Hazleton Environmental Services, Inc. (HES). Interestingly, Hercules personnel were placed in the chain of custody line, so samples went from the LAW samplers, to Hercules, to HES. The one fish sample underwent a 5X dilution by HES "due to matrix interference" so results were reported at five times the detection level of crabs.
Response: The PCC concentrations reported in the fish samples of 1990s were inconsistent among investigators with different experimental designs and protocols. The concentrations in fish were affected by many factors, such as feeding habits (e.g., bottom feeding mullet), the duration of residence time at contaminated areas when the fish sample was taken, and the location of the sampling station (e.g., the outfall ditch of the plant). In addition, reported concentrations could be affected by many factors in chemical analysis, such as the expertise of the analytical chemists, the completeness of solvent extraction of PCC from fish tissues (e.g., Soxhlet extraction), clean-up procedures, and the ways to calculate PCC concentration from the GC chromatograms (e.g., total area method or last four peak protocol.)
The 1995 Hercules report indicated that no detectable PCC was found in a sample of spotted seatrout, the only finfish sample. This result was contrary to USEPA's report on the February/March 1997 samples. USEPA calculated the PCC concentration with the total area method and found PCC in all four consumer fish samples (spotted sea trout as target species), with up to 3.9 ppm in fish from Dupree Creek. Up to 27 ppm of PCC was also reported in killifish at the Outfall Ditch of the Hercules plant25.
In April, 1997, the GA EPD collected 38 fin fish samples and prepared a draft release of laboratory reports in 1998. The report stated that Toxaphenes were not detected under various detection limits up to 1.0 ppm. This "no detection" statement was similar to the 1995 "no detection" result of Hercules report. GA EPD's 1998 draft release gave no detailed description on the chemical analysis methodology such as the sample extraction, clean up, and calibration curve for calculating the concentration.
On April 14, 2000, ATSDR formally received an analytical protocol from USEPA, Region IV describing the "Procedures for the Determination of Toxaphene," a three-page protocol dated August 14, 1997. This protocol, which was intended to be used by USEPA-Region IV and Hercules, employed "the last four to seven peaks in the 'back half' of the toxaphene chromatogram for calibration and quantification of toxaphene." The "four peak in the back half" methods dates to the packed column days, when there were only several usable peaks shown on the back half of toxaphene chromatogram (USEPA 1986 Method 8080). This "four-peaks-in-back- half" method was precise at that time, because most of the peaks (e.g., peaks u, v, w, x, y) were used in the calculation. This method, however, has lost its precision now because the powerful capillary column in modern gas chromatography instruments generates dozens of peaks in the back half of the chromatogram of toxaphene standard. Analytical chemists could then pick the four-peak set from any of the numerous possible sets. Although the Method 8081A of January 1995 kept this "four-peaks-in-back- half" method, the method was purged from the official December 1996 version of Method 8081 A, as well as the new Method 8081 B of January 1998.
Recently, GA EPD repeated the analysis of 56 samples from the old April 1997 samples with the specific methodology of both GC-ECD and GC-MS at Skidaway Institute of Oceanography. On June 19, 2000, quantitative data for the 56 samples became available and the PCC concentrations up to 26 ppm was found in fin fish. This work was published in peer reviewed, open literature 31 in September 2001. Skidaway lab is currently analyzing the toxaphene contents of another batch of fish samples collected in August 2001. These new data will be released soon. In a later health consultation, ATSDR will assess these new data, together with other data and additional reports, such as the RI/FS report, if available.
I-14. Comment: Page 17, Recent Data.
What is the ATSDR trying to say in the following: "In addition to the complex composition of technical grade toxaphene, the toxaphene compositions in fish differ from that of the reference standards. The methods used to calculate total toxaphene from chromatograms also affects the estimated concentrations." If the analysis being performed is useless for public health purposes, say so! This would be the proper place to discuss the use of the 20 isolated PCCs in the toxaphene mixture for identification of known toxicants in seafood.
The ATSDR is obviously becoming embroiled in the convoluted argument being perpetuated by Hercules regarding what is or is not toxaphene. This argument about what is or is not "toxaphene" is irrelevant to the job of the ATSDR, which is the protection of public health. The ATSDR needs to come forward and clearly state that the testing and quantification being done for toxaphene that has been aged by several half-lives is irrelevant to determination of toxicity of remaining PCCs and other chemicals that comprise the toxaphene mixture and the hazards to public health.
Furthermore, please consider the following. Toxaphene is 67% to 69% chlorine by the patented description, PCB 1268 is 68% chlorine and PCB 1254 is 54% chlorine and considered more toxic than PCB 1268 because toxicity is lost with the increase in chlorine content. The obvious question is whether "toxaphene" becomes more toxic as the PCCs become less chlorinated, as is the case with PCBs. There are serious questions that this raises. First, toxaphene is quantified by the latter peaks of the chromatogram, which are the first to shift to the early eluted peaks when the toxaphene mixture enters the environment. As soon as toxaphene enters the environment, the chromatogram reading is biased toward underestimation.
The ATSDR's discussion continues in this section by discussing vague estimations of the quantification of toxaphene with no demonstrated connection or relevance to the protection of public health. Compounded by the uncertainties associated with several half-lives, the pertinence to the protection of public health is lost.
A rewrite of this section is desperately needed.
Response: The PCC residues in killifish and sediment at the Terry Creek site had simple compositions, as indicated by a recent study carried out at Skidaway26. There were only about 5 major PCC peaks in killifish collected from the open water of the creek. At the Outfall Ditch and a lagoon there were several secondary peaks in addition to the five or six major ones. GA EPD recently analyzed the old 1997 samples with an advanced method with both GC-ECD and GC-MS instrumentation. ATSDR lists several GC-MS analytical methods in the toxaphene toxicological profile, one of which was added in response to the public comments from industrial concerns. ATSDR will assess the recent data in a later health consultation.
Toxicity of technical Toxaphene is distributed among all fractions of the mixture in a fairly even fashion. The middle fractions are relatively (less than 3 times) more toxic to fish and mice than the technical Toxaphene, while the fractions at both the non-polar and polar ends are less toxic than the technical Toxaphene (Issensee, AR et al., J. Agric Food Chem 1979; 27:1041-46; Oshawa T et al., 1975 ibidem 23:98-106.) The major toxic components, toxicant A and toxicant B, are in the middle fractions. In the GC chromatogram that covers the range from early eluted Parlar 11 to the late eluted Parlar 69, toxicant A (i.e., A1 and A2) is located at Parlar 42 while toxicant B located at Parlar 32. Both of the major toxicants are located at the middle section in the range of the GC chromatogram.
On 4/20/2000, ATSDR formally received from USEPA, Region IV, a copy of the 8/14/97 version of the analytical protocol for toxaphene determination. This protocol, to be used by USEPA Region IV and Hercules, employed "the last four peak" method in determining toxaphene concentration in soil and water. The earlier GA EPD draft release on initial 1997 fish analysis did not mention whether this protocol was used for fish analysis.
I-15. Comment: Page 18, Off-Site Contamination.
This section needs to be deleted for the following reason. Brunswick is a not a cotton field nor has it been a cotton field. Toxaphene was used on cotton fields. The widespread use of toxaphene as a pesticide is not relevant unless the ATSDR intends to demonstrate that Brunswick was a cotton production area. Personally, I have never seen the first cotton plant in Brunswick. The ATSDR will most certainly "enter the storm" if they wish to contend that the contamination is from any of the sources in the opening paragraph of this section.
Response: The opening sentence has been rewritten to reflect that pesticidal use of PCC was insignificant in Brunswick compared with the amount of PCC discharged from the manufacturing plant. This section serves to define the extent of the PCC contamination beyond the boundary of the proposed NPL site. Historical data from 1967-1972 demonstrated that the distance (or dilution) from the PCC source reduced the frequency and concentration of PCC in oysters24. The oysters were 100 % contaminated in Terry Creek with maximum concentration of 54 ppm, the frequency of contamination reduced to 98% with maximum concentration of 7.5 ppm in St. Simons Sound, 60% with 3.5 ppm maximum concentration at Jekyll Island, and 13% and 1.0 ppm maximum concentration at Satilla River sampling station.
In the off-site samples of 1997 by Skidaway staff, PCC level in finfish livers and ova are relatively low in comparison with the on-site samples of March 1997 by USEPA. The off-site PCC concentration of 0.01 ppm in whole grass shrimps was also very low compared to the 1.2 ppm found in the on-site grass shrimps in the1998 phase I studies.
I-16. Comment: Page 19, Soil. Residential yards near the Hercules
The data shows "toxaphene" levels up to 14 ppm in residential yards. I strongly suggest that the ATSDR review all available data and make their own determination on any pattern of distribution from the plant, rather than quoting a company hired by Hercules to defend their position in a lawsuit. If ATSDR desires to use arguments presented in lawsuits, the positions of the plaintiff and defendant should be presented with the supporting documentation and studies from both sides. The quality of research by the ATSDR is seriously lacking in this document.
Response: ATSDR will assess any new data when they become available. There are uncertainties in the reported PCC concentrations from different analytical laboratories with different methods. For example, soil concentration in a general area was reported at about 20 ppm by one camp and 330 ppm by other camp. Without detailed, quality reports, ATSDR can only cite the available data with reservations.
Another example of the inconsistency in early data of soil-borne toxaphene occurred in the school yards of Glynn county. No toxaphene was detected in the sampling effort using the gas chromatography with electron capture detector (GC-ECD) when using the characteristic peaks on the latter half of the chromatogram to quantify the amount of toxaphene present. But an average of 1.8 (0.37-11.2) ppm of toxaphene was found in 20 soil samples from the same sample pool when total area method was used in the GC-ECD analysis. Quality data were hard to find in the data base at that time.
I-17. Comment: Page 19, Schools.
Again, it appears that the ATSDR only has the data that Hercules wants you to have. There have been samples and results that have shown "toxaphene like" compounds on school property. The ATSDR is advised to look for PCC on school properties. Again, it is the ATSDR's job to make an objective assessment of the potential for adverse human health effects from PCC, degradation compounds, other toxic components of the toxaphene mixture, and not define what is or is not "toxaphene." Over 670 chemicals were distributed over the community. All the community wants is an objective assessment of the risk from the remaining chemicals and the breakdown products of PCC.
Response: ATSDR is aware of the great uncertainty on PCC content in school yard soils. The reported concentrations of the general area can be as low as 1.8 ppm from one source and 330 ppm from the other source. ATSDR received this information through either a simple e-mail message or a copy of a legal notice to the State of Georgia. ATSDR has no detailed information on the chemical analysis or method for calibration and quantification. A new test with the combined, advanced methodology of GC-ECD and GC-MS for school yard soils is in progress at the Skidaway lab. The new testing is sponsored by the Glynn Environmental Coalition and the Glynn County School Board, Brunswick, GA.
The toxicity of technical grade toxaphene is distributed in all fractions. The fractions in the middle portion are a little (up to three times) more toxic than the technical toxaphene in acute testings with aquatic organisms. Those fractions on both ends (the least polar fractions and the most polar fractions) are a little less toxic than technical toxaphene ( Nelson J and Matsumura F. Separation and comparative toxicity of PCC components. J Agr Food Chem 1975; 23:984-90.) In this health assessment, ATSDR reviews data of various PCC mixtures or components that present a potential hazard to public health.
I-18. Comment: Page 19, Air.
For "toxaphene" to be detected, all of the peak ratios need to be correct. Unless the "toxaphene" is attached to a dust particle (as it was when detected at the Hercules 009 Landfill Superfund Site, it will not be detected. The ATSDR should require testing for individual PCCs (congeners). The individual chemical components of "toxaphene" will volatilize at different temperatures and under different environmental conditions. Interestingly, in the next section of the PHA, the ATSDR discusses calculating the quantity of toxaphene on just a few selected peaks of the over 670 peaks on the chromatogram.
The ATSDR should require all air samples be analyzed for individual PCCs, total PCC, degradation compounds and other components of the toxaphene mixture.
Response: ATSDR has only preliminary information on the absence of detectable airborne toxaphene at this site. No actual quantitative number was available to ATSDR. ATSDR learned in August 2001 that the detection limit was 2.5-6.6 ug/cubic meter (USEPA Region IV letter to Lake Superior Binational Forum) without information on the analytical methodology. This Health assessment is based on available data reported from USEPA, State of Georgia, and other sources. Only limited information was available to ATSDR on PCC in air. ATSDR considers individual PCCs, total PCC, degradation compounds relevant to public health. When these data become available, ATSDR will evaluate their public health implications.
I-19. Comment: Page 20, Quality Assurance and Quality Control
The statement, "Although the chemistry of toxaphene is now well-developed as twenty isomers have been isolated and identified," would be the funniest statement in the PHA, except I believe the authors are serious! Since when is 20 of over 670 chemicals "well-developed chemistry"? The problem here is ATSDR is discussing "toxaphene" and not PCC. ATSDR must identify what are the toxic components of PCC and the toxic degradation products and stop this infatuation with identifying "toxaphene." Again, after several half-lives, the amount of "toxaphene" is irrelevant to the job at hand, which is determining the threat to public health from PCC and the degradation products.
I would be interested to see ATSDR's data supporting the increase in chlorine under anaerobic conditions. In addition, I have never seen the first study that found any atmospheric transformation of toxaphene (even though you cite a five-day half-life for the vapor phase to the contrary), they all say there is no transformation during atmospheric transport. Please, rewrite this section (and find a new source).
Response: The status of the PCC chemistry and toxicology is about the status for PCBs 20 years ago. Of the 177 or so significant PCC components, about 20 individual compounds are commercially available for the researchers. About 20 of the available standard congeners were used by the Skidaway lab in the chemical analysis of the PCC congeners in the fish samples collected by the State of Georgia in April 1997.
About five or six toxic PCC compounds have been identified since mid-1970s. They include toxic fraction A (Toxicant A1 and Toxicant A2), Toxicant B, Toxicant Ac, toxicant 7 and Toxicant 8. Acute toxicity of toxaphene seems to distributed in all fractions of technical grade toxaphene. The most lipophilic (most soluble in oil) and least lipophilic (least soluble in oil) fractions are less toxic than the fractions in between. Nevertheless, these fractions have toxicity comparable to the technical toxaphene, usually within the range of 30% to 300%.
Under anaerobic conditions, the highly chlorinated fraction of the technical toxaphene readily degraded and lost some of chlorine atoms in their molecules. This sentence was revised in the text to improve clarity. It is possible in some cases that bioaccumulation in fish can favor the lipophilic PCC fraction (with higher chlorine content) already in water than the less lipophilc fraction from the PCC mixture in contaminated water.
The text was revised and the short half-life (five days) in vapor phase was modified. Two sentences are added to the text to qualify the half-life statement. They indicated that the airborne toxaphene in particulate form is resistant to photo degradation and that the airborne mixture of PCC had an atmosphere residence time of 46-70 days.
I-20. Comment: Page 21, Community-Based Environmental Protection
Add 4th Street Landfill, and the T Street Landfill to your list of where toxaphene is concentrated in Brunswick. The ATSDR can confirm the presence of toxaphene at these sites in the CBEP documents.
Response: Text was revised accordingly.
I-21. Comment: Page 22, Air.
The ATSDR should discuss the potential for exposure from air during the drying of sediments in open drying beds during remedial activities at the site. In addition, the ATSDR Toxicological Profile calls for a 1-foot clay covering of toxic waste sites containing toxaphene to stop volatilization from sandy soils. What is the potential for air exposure from the other identified source areas and the site?
Response: ATSDR received short preliminary reports and photographs taken during the removal operation. USEPA sent some one-paragraph reports informing ATSDR that toxaphene was not detected in the air. During the removal operation a salt crust covered the sediment piles. This salt covering might depress the dust formation and reduce the particulate PCC in the air. ATSDR learned in August 2001 that the detection limits for the airborne PCC as 2.5-6.6 ug/cubic meter. The analytical method is unclear and will not be known until ATSDR receives the requested detailed report from USEPA.
I-22. Comment: Page 24, Toxaphene.
Rewrite the section about the hypothetical fisherman. 30 grams of fish is about one ounce (210 grams is one meal or an average of one ounce per day). Otherwise, the discussion is meaningless unless ATSDR discusses PCC and the degradation compounds that are not currently detected. (Remember the later peaks are used to quantify and the peaks shift to the left when TOXAPHENE enters the environment. Also, two half-lives.) Furthermore, levels up to 33,000 were found near the Hercules dock.
On page 25, statements number 1, 2, 3 and 4 are very good! These statements embody much of my concern. It is so important that people know that there is a great unknown at the site about what are the possible exposures and potential health threats.
Response: Text was revised and the fish meal was specified as 210 grams.
To advance the understanding of the toxicity of PCC residues in local fish, ATSDR plans to study the acute toxicity of the suites of PCC-congener mixtures in fish samples. The PCC mixture in contaminated finfish will be measured with the chemical analysis and the bioassay that is based on acute toxicity. This study could provide the relative acute toxicity of PCC residues in fish to the toxicity of the technical grade toxaphene.
I-23. Comment: Page 26. C. Children's Exposure and Health
Very well stated.
Response: No response needed.
I-24. Comment: Appendix 3
The discussion in this section is needed concerning the potential health threat and quantifying that threat. The work being done by Skidaway Institute should be of value for future evaluations of potential health threats from individual PCC. In addition, the actual toxic components will be identified and quantified.
Response: The Skidaway data for individual PCC and total PCC for the April 1997 fish samples was submitted to GA EPD on June 16, 2000. PCC concentrations were quantified using both GC-ECD and GC-MS. ATSDR will evaluate the new data later in a health consultation.
I-25. Comment: General Comment.
There needs to be an inclusion of a discussion of the endocrine disrupting properties of PCC in the PHA.
Response: ATSDR toxicological profile of 1996 did not cover this topic, on which some reports exist. When enough relevant studies accumulate in the database and adequate review papers become available, ATSDR will address this issue.
I-26. Comment: General Comment.
Dioxin has been found in all samples containing toxaphene at the Hercules 009 Landfill Superfund Site. The lack of dioxin data for the site should be identified as a priority need for future analysis in the PHA by the ATSDR. The PHA should be deemed incomplete until the needed data is provided for sediment and fish tissues. Furthermore, testing has been positive for dioxin and PCBs on the Hercules Plant Site. In addition, the possibility of dioxin/PCB's levels in seafood being additive to the known dioxin discharges from Georgia Pacific Pulp and Paper, dioxin/PCBs from the LCP Chemicals Superfund Site, dioxin from the Brunswick Wood Preserving Superfund Site, and PCBs from Plant McManus Site should be discussed in the PHA.
Response: There are many potential sources of dioxins in the Brunswick area, but only one sample taken from this site was tested for dioxin. In 1997, USEPA found some dioxins in two sediment samples, one from the Terry Creek/Dupree Creek junction and another off-site control sample from Jove Creek. The concentrations in the on-site station did not exceed the concentrations in the off-site control station25. The most abundant components of sedimentary dioxins, octachlorodibenzo-p-dioxin (OCDD), was found to be 77 ppt at an on-site station and 80 ppt at an off-site station. Total heptachlorodibenzo-p-dioxins (HpCDD) in on-site and off-site samples were 9.3 and 19 ppt, respectively. The most potent dioxin component, 2,3,7,8-TCDD was not detected in either sample at the detection limit of 10 ppt. In the estuaries of urban U.S. cities, the range of HpCDD and OCDD concentrations in sediments, on a dry weight basis, were 500-3,000 ppt and 2,000-37,000 ppt, respectively1. ATSDR will assess dioxin contamination at this site if new data become available.
Public Comments and Responses; Set # II:
II-1. Comment: The exposure scenarios are potentially biased. For exposure modeling, the report defines certain pathways, such as the Terry Creek Drive area for soils, of the Main Dredge Spoil area. The report correctly notes, however, that Brunswick has numerous areas where toxaphene exposure is likely. The model should reflect the aggregate opportunity for toxaphene contact rather than just the proximity to one of two courses, such as immediately adjacent gardens. Pathway analyses were never designed, or intended to be used, for threats such as toxaphene in Brunswick.
Response: Among the many potentially contaminated areas, the health assessment used the most contaminated areas to estimate the maximum potential exposure. The worst case scenario at the Terry Creek site usually included the highest PCC concentration for lifetime exposure. This approach will cover the other areas with lesser degree of contamination.
II-2. Comment: The toxaphene analyses used in these studies historically under-reports toxaphene. Both the USEPA and ATSDR are to be commended for efforts in finally obtaining scientific counsel other than Hercules in matters regarding the analysis of toxaphene. However, after noting the shortcoming of the "back-half" GC-ECD methodology, the report draws sweeping conclusions using that same methodology.
Response: The limited fish data (3.9 ppm) used in the dose estimation of this health assessment were measured with the total area method25. It is not clear whether the "back-half" method was used in the first analysis of the April 1997 fish samples by GA EPD28. In the draft release28 from GA EPD, analytical procedures were described in a very general fashion. On April 20, 2000, ATSDR formally received from USEPA a copy of the August 14, 1997 protocol using the back- half method for PCC residue estimation and learned of the possible use of this method by the GA EPD. Recently, GA EPD and the Skidaway Institute of Oceanography made another chemical analysis on all of the old fish samples collected in April 1997. The specific method with advanced GC-ECD /GC-MS procedures produced quality data. Total PCC concentrations, the sum of all individual PCC components, was found up to 26 ppm in mullet. The final report was released on June 16, 2000. It was also published in a peer-reviewed, open literature in September 200131. ATSDR will use this report and other data such as future RI/FS reports later in a future health consultation.
II-3. Comment: This report is biased since it draws mainly on conclusions previously drawn by ATSDR and USEPA, not an independent evaluation of this particular site. Further, a plurality, perhaps a majority, of the cited references are merely prior opinions, not scientific data or facts. For example, reference 38 is not a scientific or statistical validation of any method; however, it is cited in the text (page 19, 4th paragraph) as a critical review of the total toxaphene peak analysis methodology. Where there is known ambiguity from sample environmental degradation or the sample material is from a manufacturing process, but not a product, the total area method may be as valid as any other technique. It is recommended that the toxaphene analysis and health assessment be revisited by a panel of outside experts, both analytical chemists and toxicologists, rather than using old opinions as factual models.
Response: The reference #38 about the 1996 debate on "back-half vs total area" was deleted from the text. There is new information on the PCC residues at this site since 1996. The work at Skidaway lab26 clearly showed that PCC residues eluted in the front end of the GC-chromatogram of technical toxaphene. And a major component (Hx-SD) in the PCC residues could be a dechlorinated product of Toxicant B. ATSDR reviews any available information on PCC components or degradation products that are relevant to public health. Under the sponsorship of GA EPD, Skidaway analyzed the old fish samples of April 1997 and produced quantitative data for PCC residues (the sum of all PCC components). PCC concentrations up to 26 ppm were measured in the samples that yielded no detectable toxaphene in the initial 1998 analysis with old method28.
The suggestion on the use of outside panel is a useful idea. ATSDR will continue to seek help from outside experts as needed. In fact, the total area method vs. back-half method was debated in 1996. It is an ongoing issue at this site. On February 6, 2001, administrators and scientists from USEPA, GA EPD, and Skidaway lab came to ATSDR for a discussion on how to count the PCC residues: the whole body or just the tail end. About 20 members attended the meeting. USEPA proposed three ways for the determination of PCC residues, but no conclusion was reached at the meeting.
II-4. Comment : The potential threat from seafood may be underestimated. The report relies heavily on recent fish studies in reaching conclusions regarding toxaphene exposure. There is too little recent data to allow such generalizations. Current and previous studies are not necessarily matched in seasonal, species and tissue fluctuations, and bias is possible, perhaps likely. Further, there are too few studies over time to permit a stated Conclusion (page 28, first paragraph) that a trend exists in the decline in biological toxaphene. Some studies used whole animals, others used samples prepared as if for consumption. The "Conclusion" section of this assessment might better be entitled "Generalizations."
Response: A qualification sentence is added at the end of this section. It states that ATSDR will re-evaluate the health hazard of this proposed NPL site when new data become available. The new data (6/16/2000) that reported maximum concentration of 26 ppm by Skidaway scientists on the PCC level of the old (April 1997) fish samples will be assessed later in a health consultation. The average PCC concentration in the whole body of a forage fish (killifish) declined from 62 ppm in 1970 to 7 ppm (1.9-27 ppm) in 1997. GA DEP and the Skidaway lab are analyzing a recent batch of fish samples harvested in the early fall of 2001.
II-5. Comment: Risks from chemicals other than toxaphene are ignored. The report correctly notes that many chemicals co-chromatography with toxaphene and can be hidden when performing the toxaphene analysis. The material deposited within the creek was manufacturing effluent, not commercial toxaphene. The analysis and the conclusions are shortsighted in focusing on just the toxaphene standards. In particular, dioxins are underanalyzed in this review.
Response: The total concentration of PCBs and non-toxaphene organochlorine pesticides (OCPs) in most contaminated finfish is relatively low according to the new Skidway report of June 16, 2000. In the mullet with 26 ppm of toxaphene, the sum of all the other PCBs/OCPS contaminants is just a fraction (8%) of the toxaphene. The other organochlorine pesticides have low concentrations when the toxaphene concentration was low--in the ppb range.