Sediment Operable Unit
ST. LOUIS RIVER/INTERLAKE/DULUTH TAR
NATIONAL PRIORITY LIST (SUPERFUND SITE)
ST. LOUIS, ST. LOUIS COUNTY, MINNESOTA
October 25, 1999
Carl Herbrandson, Ph.D.
Minnesota Department of Health
At MDH, our concern is human health--and we focus our analysis of effects during the choice of remedy only on its implications as related to human health.
- I work in collaboration with MPCA toxicologists--and since MPCA toxicologists are intimately involved in this site my involvement has been limited primarily to consultation and review.
- A couple of years ago I did work with Helen Goeden on development of health criteria for assessing potential risk to human health at this site
- I have also briefly reviewed the ROD
- Directly involved in USX Site (a site which has had less involvement by MPCA toxicologists) for the last 2 years--spent some time looking at dredging and its effect on distribution/redistribution of persistent chemicals including mercury
While MDH has some concern about PAH contamination at this site, I believe that the MPCA has covered this concern. Mercury in the sediments of Stryker Bay has potentially broader Public Health implications, and is of concern to MDH because it may be contributing to the methyl-mercury pool in the lower St. Louis River and to levels of methyl-mercury in fish which are then caught and eaten.
- This is the only river in the state with a mercury-based DoNotEat fish consumption advisory for predator fish (25-30" walleye) in the state. While mercury in sediment and mercury in fish do not always correlate well due to issues like: variable rates of mercury methylation and fish movement throughout a watershed, the EPA has suggested that if sediment concentrations of mercury are less than 1/10th of those found in layers 101 and 102 at this site, fish consumption is probably a human health issue.
- There is a lot we do not know about mercury geo and hydro-chemistry. We know that mercury has an affinity for materials which make up sediment, so generally, it is not very mobile in sediments. We also know that for mercury to be accumulated in fish it needs to be converted to methylmercury. And we know that conversion to methylmercury is increased in anoxic sediments. Therefore, we expect the production of methylmercury to be elevated in a wetland.
- So how can mercury move in the aquatic environment??
- We know that mercury can be transported with suspended sediments. Mercury, and (mono)methyl-mercury, movement in groundwater is very slow, but it is unknown if mercury or methyl-mercury could be co-transported with other compounds in a wetland or what other chemical species of mercury may be manufactured in a wetland. It is also not known if or how vegetation in wetlands may affect the mobility of buried sediments. There is some information demonstrating that significant amounts of methyl-mercury can be found in macrophytes (such as cat-tails - Miles CJ; Fink, LE), presumably taken up from sediments. Some portion of the methyl-mercury appears to be volatilized into the air. This may be just one way in which mercury can move in a wetland.
- The presence of mercury contamination in Stryker Bay, at concentrations up to 2.7 ppm is of concern to MDH. If it could be covered in place and we could be assured that it would stay there, we would agree that the capping alternative is protective of human health. Given our understanding of mercury in the environment at this time, MDH believes that there may be significant uncertainty in any such assurances.
- The sediments in Stryker Bay are spread over a large area; they have a very large surface area relative to their volume--both on top (currently in contact with water--and potentially in contact with wetland in the future)--and they are also exposed, on the bottom, to groundwater which may discharge into the bay/river.
- MDH believes that there are significant uncertainties which may be associated with the maintenance of a barrier in Stryker Bay which is impermeable to mercury--these include
- the effect of groundwater movement and hydraulic pressure
- maintaining the integrity of a cap in a wetland/river shoreline environment.
- The greatest advantage to dredging alternatives are that they provide the opportunity to control the infiltration of groundwater into sediments (or dredged wastes) through the minimization of the surface area to volume ratio, and possibly by construction of barriers, or through stabilization of the spoils. With a decreased profile to groundwater, surface water, and wetland, long-term uncertainties and the long-term potential availability of mercury from sediments can be minimized.
- While dredging may provide greater long-term security - and less long-term uncertainty, dredging is not a clean operation. Sediments can be resuspended and contaminants can be redistributed and made available to aquatic organisms, or they can volatilize into the air we breathe.
- We are not engineers and we do not expect to be involved in the development of a detailed remedial action plan for this site, but from the studies we have reviewed involving PCB and toxaphene contaminated sediments, we believe that dredging can be done with minimal resuspension and redistribution of contaminants. It is our understanding from our reading of the literature that choice of dredging methods, appropriate use of sediment screens, and controlled dewatering facilities when properly employed by experienced contractors can control contamination and make overall dredging effects on the environment very minimal. (for example: One study conducted during a PCB cleanup showed no increase in PCBs in water (and suspended particulates) 15-70 m downstream from the dredging activity - Hafferty, Pavlou, Hom)--hydraulic dredge--no sediment screen.
- There are sites/conditions where we would definitely not recommend dredging (such as in a rapidly flowing river--or in the deep limnetic sediments). This is not one.
Volatilization of contaminants from dredged spoils does occur. However, we expect volatilization from dredging at this location to be limited because the sediment will not be dryed. During dredging at this site, we would expect that very little mercury would be volatilized, and that the only compounds which may be released into the air in significant amounts would be the light - non-carcinogenic PAHs, such as naphthalene.
- Studies have shown that chemicals similar to the light - non-carcinogenic PAHs at this site volatilize at a rate related to the evaporation of water and the drying of the sediments. Therefore if drying is kept to a minimum, volatilization can be kept to a minimum.
- Most studies/concern about volatilization I have seen are specifically related to dredging operations where there is land disposal of the dredged materials. It is my understanding that at this location, the sediments will remain wet, thereby limiting the volatilization of light PAHs. It is also my understanding that previous dredging or soil excavation at this site was accompanied by drying and baking of the sediments prior to burning--clearly activities which maximize volatilization.
- Dredging is always associated with smells--typically swamp gases such as methane and H2S are released. Both of these gases volatilize from water and sediment much more readily than PAHs. PAHs are ubiquitous - and it is not known how much dredging could contribute to already existing levels of PAHs in air---studies have shown that levels of PCBs in air may be increased by dredging PCB containing sediments, but that study also raised questions about the possiblity that some of the significant levels of PCBs attributed to the dredging may have been volatilizing from open water in the bay itself. Again--at this site, these dredged spoils will not be dried out, therefore, volatilization/odors will be minimized.
In conclusion, MDH supports this ROD and hopes that the issues we have raised and discussed will be considered during the development of the detailed Remedial Action Plan.
|Date:||February 19, 2003|
|To:||Shelley Burman, Supervisor|
|Risk Evaluation/Air Modeling Unit|
|From:||Hillary M. Carpenter, Ph.D.|
|Health Risk Assessment Unit|
|Subject:||Air Health-based Value for Naphthalene|
This memo is in response to your agency's December 5 request for an air health-based value (HBV) for naphthalene to be used in a clean up of the St. Louis River Interlake Duluth Tar site. During consultation with staff familiar with the project, it was decided that two numbers were needed; consequently, the Minnesota Department of Health (MDH) has derived both an acute HBV (200 µg/m3) and a chronic HBV (9 µg/m3) for inhalation exposure to naphthalene. A description of the techniques, assumptions and caveats used in developing these numbers follows.
Acute. There are limited data addressing the impacts of acute exposures of experimental animals to naphthalene. However, because there are a number of anecdotal reports of naphthalene toxicity in humans (nausea, vomiting, abdominal pain, and hemolytic anemia) at concentrations above those that trigger an odor (200440 µg/m3) MDH recommends the use of an acute HBV (one hour exposure) of 200 µg/m3 as a reasonable maximum exposure level.
This use of this number is supported by results from a study on rats that reported respiratory changes (cell swelling and sloughing) following four hours of exposure to 380 mg/m3 of naphthalene (Buckpitt, 1982). In this study 204 mg/m3 was a No Observed Adverse Effect Level (NOAEL). Applying an uncertainty factor of 1000 (10 for intraspecies variability, 10 for interspecies variability, and 10 for database deficiencies) gives an acute value of 200 µg/m3 for a four-hour exposure. As an additional precaution MDH recommends that this number be applied using a 1 hour averaging time.
Chronic. Two studies chronic rodent bioassays, one in mice (NTP, 1992) and one in rats (NTP, 2000) are the basis for MDH's chronic HBV of 9 µg/m3 naphthalene. Both of these studies involved the administration of naphthalene for 6 hours per day, five days per week for two years. Both studies produced a Lowest Observed Adverse Effect Level (LOAEL) of 10 ppm naphthalene with fairly marked respiratory and nasal impacts as adverse endpoints.
Manipulating this exposure to allow for a 24 hour/day and a seven day/week yields an adjusted LOAEL of 1.78 ppm which converts to a value of 9.3 mg naphthalene/m3. Applying an uncertainty factor of 1000 (10 for intraspecies variability, 10 for interspecies variability, and 10 for the use of a LOAEL rather than a NOAEL) results in a final chronic HBV for naphthalene of 9 µg/m3. Although, by definition MDH considers a chronic exposure to be one that occurs on a daily over a 70-year lifetime, MDH recommends that exposures that take place for more than 10 percent of an individual's lifetime be assessed using chronic values. MDH anticipates that chronic HBVs will be applied using annual emission estimates. MDH understands that the exposures associated with the remediation of this site will be occurring for a maximum of seven months each year, but feels that the potential for public health impacts be assessed using the chronic number.
Please be advised that although MDH has a reasonable level of confidence in the chronic naphthalene number, and in fact intends to propose this value as a HRV during the next rule revision, available data do not address two additional toxic endpoints reported in humans, cataracts and the blood disorder hemolytic anemia. MDH is therefore less certain about the conservative nature of the naphthalene number for these endpoints.
MDH has less confidence in the acute value and suggests that it be considered a site-specific screening number to be used to trigger some remedial action.
If you have any questions regarding the development of these numbers please call me at (651) 215-0928.
Buckpitt, A.R. (1982). Comparative biochemistry and metabolism. Part 2: Naphthalene lung toxicity. AFAMRL-TR-82-52, pg 25-30. Air Force Aerospace Medical Research Laboratory, Wright-Patterson Air Force Base, Ohio.
NTP (1992). National Toxicology Program. Toxicology and carcinogenesis studies of naphthalene (CAS No. 91-20-3) in B6C3F1 mice (inhalation studies). NTP Technical Report Series No. 410. NIH publication no. 92-3141. U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, Research Triangle Park, N.C.
NTP (2000). National Toxicology Program. Toxicology and carcinogenesis studies of naphthalene (CAS No. 91-20-3) in F344/N rats (inhalation studies). NTP Technical Report Series No. 500. NIH publication no. 01-4434. U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, Research Triangle Park, N.C.
|cc:||Barb Jackson, MPCA|
|Doug Beckwith, MPCA|
|Bruce Brott, MPCA|
|Doug Wegstein, MPCA|
|DEPARTMENT:||Health/Pollution Control Agency||State of Minnesota |
|DATE:||February 7, 2003|
|TO:||Jane Mosel and Mike Bares, Minnesota Pollution Control Agency|
|FROM:||Carl Herbrandson, Ph.D., Toxicologist, MDH; |
Steve Hennes, M.S., Toxicologist, MPCA
|PHONE:||651 / 251-0925|
|SUBJECT:||SLRIDT Site visit|
On February 6 2003, from about 2:30 to 4:30 PM, Carl Herbrandson and Steve Hennes walked the ice at the SLRIDT site. It was a cloudless sunny day with a temperature of about 0° F and a N-NE wind of about 8-10 MPH. The ice looked about 3 feet thick and in most areas it was covered by about 1 inch of snow. In some areas, mainly in the river and near the mouth of the bays, the ice was uncovered and clear, but crisscrossed by cracks.
We started at the tip of Slip #7 and walked to the river. We proceeded up river to Stryker Bay, walked around the bay clockwise, and then walked back to our start point in Slip #7. We did not walk into Slip #6, but observed it from the river. See attached figure for route and location of notable observations.
Ice in the middle of Slip #7 was about 1 1/2 foot below ice along parts of the western and northern sides of the bay. This phenomenon was also noted at places in Stryker bay and also along the river.
The artesian (?) well in the shed attached to the large warehouse on the eastern shoe of Slip #7 was flowing rapidly (maybe > 1 L/sec). The water heated the area enough to keep ice from building up around the well or the shed. The pool of water extended to the gravel roadway between the shed and the bay, and then disappeared into the ground.
There were numberous pressure ridges of ice in the river. While most were about 2-3 feet high, some reached about 4 feet. The largest ridge extended from the eastern tip of Slip #7 to Indian point. There were several branches from this ridge. Most went out into the river, but one went from this ridge north to the land between Slip#6 and Stryker Bay.
There was an ice fishing house in the mouth of Slip #6 (see attached figure) that appeared to have been brought form shore at Hallett Dock Co.
A small pressure ridge was apparent at the narrowest point in the mouth of Stryker Bay (see attached figure). Residents' docks, about 7-8 total, were all pulled to shore except for the Simonson's small dock, a wooden dock that sat on the edge of a wetland on wooden blocks, and a dock that was suspended above the ice from a single piling (see pictures). Snowmobiles and 4-wheelers appeared to regularly use the bay.
Four apparent groundwater springs were noted in shallow areas of Stryker Bay (see attached figure and pictures): one in the middle of the wetland at the head of the bay; a second a little south of the first, in cattails along the eastern shore; the third in the area of the former discharge pipe, along the shoreline; and the fourth about 30 feet from the eastern shore, and about 100 feet north of the mouth to the bay.
Along the western shore of the mouth of Slip #6 we observed ice (4-6 inches thick) with about 3-4 inches of sediment attached to the bottom, pushed up on shore (see attached figure and pictures).
The contaminated sediments sign at the mouth of Slip #7 had apparently been unbolted from the sign posts and turned to face the shore. In addition, the area around the sign posts had either been eroded away, or the post had been lifted or heaved about 1 foot. About 1 foot dia. of sediment/concrete remained attached to the posts, about 1 foot above the terrain (shore? sediment?)(see attached figure and pictures).
In the wetland area of Slip #7, on the vegetated sandbar, there was a large wooden box that contained 2-10x12x7 inch data logger-type metal boxes. One was identified as a multiplexer. Several cables extended out into the water (ice) of the bay. There were several sets of footprints around the box.
The air emission model in the Data Gaps Report proposes that as the naphthalene concentration in sediment increases, air emissions increase proportionally until naphthalene emissions reach a maximum. The naphthalene concentration in sediment at this point is considered to be the 'breakpoint' and naphthalene emissions at all sediment concentrations above this 'breakpoint' will be the same as emissions at this 'breakpoint.' According to the model, this 'breakpoint' is a result of flux from sediment to air being limited by the maximum solubility of naphthalene in water. The model also proposes that this chemical concentration limit is the same under all experimental conditions: high and low ratios of sediment:water, and with and without mixing. Theoretical problems with the model are discussed in the main body of the Health Consultation.
The Data Gaps Report proposes that the solubility limit to naphthalene emissions are achieved with sediment containing 238 mg/kg naphthalene. Application of this proposed 'breakpoint' to experimental data from the Data Gaps Report suggests that there are significant errors with the emission model proposed in the Data Gaps Report.
The Health Consultation contains an extensive discussion on the inadequacies of the analytical characterization of the experimental slurries. The lack of reliable data may impact the actual calculations in this appendix, but it does not affect a qualitative analysis and it is unlikely to affect the conclusions of this analysis. For the sake of the analyses below, it is assumed that all naphthalene in each experimental slurry came from additions of solid sediment containing 11,000 mg naphthalene per kg solids.
Data from the Data Gaps Report (Table A2-2) shows that naphthalene emissions from the experimental slurries over the first 2 hours were 9.7, 14.6, 6.4, and 9.5 mg for the 45% (46.8% measured solids - Table A1-1), 8% mixed (8.2%), 8% quiescent (9.0%), and 1% (1.5%) slurries respectively. The Data Gaps Report model says that emissions would be the same if the naphthalene concentration in sediment used to make up the slurries was 238 mg/kg. Under these conditions, the total amount of naphthalene in the experimental slurries would have been 234, 41, 45, and 7.5 mg naphthalene for the 45%, 8% mixed, 8% quiescent, and 1% slurries respectively. Comparison of the predicted emissions from the model and the total naphthalene in the 238 mg/kg sediment additions shows that the predicted emissions exceed the total naphthalene in the 1% slurry by about 27%. In addition, if the emissions are estimated from the 24-hour Data Gap Report emissions, the expected emissions will be about 2.5 times the total naphthalene in the 1% slurry. Similarly, predicted 24 hour emissions from the 8% mixed slurry of 238 mg/kg naphthalene and water will be about 1.1 times the total amount of naphthalene in an experimental slurry.
Table #1A shows the amount of naphthalene in experimental and proposed slurries, and the actual experimental results as well as the predicted results.
|Experimental Results:||Projected results from |
Data Gaps Report model:
|Slurry made with 11,000 mgnaphthalene/kgsed||Slurry made with 238 mgnaphthalene/kgsed|
|Initial naphthalene |
|Total naphthalene emitted |
|Initial naphthalene |
|Projected total naphthalene emitted |
|2 hr||24 hr||2 hr||24 hr|
|45% Slurry |
8% Mixed Slurry
8% Quiescent Slurry
|9.7 (0.09%) |
| 41.2 (0.38%) |
|9.7 (4%) |
|41.2 (18%) |
Since the Data Gaps Report predicts greater naphthalene emissions than contained in the sediments it cannot be expected to accurately predict emissions during dredging.
A sequential batch leaching test (SBLT) was performed for the Data Gaps Report on sediment samples from the 3 bays of the SLRIDT site, to derive site-specific partitioning coefficients. In order to derive coefficients that describe partitioning between sediment (normalized to total organic carbon (Toc)) and water, it is necessary for concentrations to reach equilibrium between the 2 mediums, and for the concentration in water to be below the solubility of the chemical (e.g., naphthalene). If the concentration in the sediment is too high, the dissolved concentration cannot increase to reflect the Koc of the chemical, but is limited by its solubility. If, on the other hand, the dissolved concentration does not approach the solubility, Koc can be determined from the relationship described in equation #1.
Koc (L/kg) = [naphthalenesediment] (mg/kg) / [naphthalenedissolved] (mg/L) / foc -equation #1
(for dissolved naphthalene < 31 mg/L)
Where foc = the fraction of organic carbon in the sediment
In an SBLT, determination of Kocs can be made from sediment and dissolved analytical data from multiple samples that are successively prepared by serial dilution of the same sediment sample. If the dissolved concentration of a specific chemical remains the same over successive iterations, then the experiment for that specific chemical is probably limited by the solubility of the chemical. Therefore, the Koc cannot be determined from these data. On the other hand, if the dissolved portion decreases, Kocs can be calculated from each iteration, or by combining the data from all iterations.
If calculated Kocs are similar for each iteration for the same chemical, then at least the portion desorbed was adsorbed similarly to sediments. If the Koc increases with successive leachings, the chemical may be adsorbed to different types of organic carbon in the sediment with different affinities. Differing affinity for different types of organic materials was briefly discussed in the Health Consultation, with references to scientific literature that describe an increased affinity (increased Koc) of PAHs for aromatic and anthropogenic organic particulates.
The presence of NAPL in a sediment sample makes determination of a Koc useless. Furthermore, Koc is not a useful metric in a system that contains a NAPL. Partitioning between a NAPL and water is not driven by the amount of NAPL in a system, but by the fraction of the NAPL that is the chemical of interest. Therefore, while sediment concentration of the chemical may vary broadly in the presence of a NAPL, the dissolved concentration can remain constant.
Problems encountered when calculating Koc
Measuring dissolved concentrations of non-polar organic chemicals, especially very hydrophobic (i.e. with log Kows > 4 or 5) and very volatile chemicals (e.g. benzene, naphthalene), can be difficult. Significant amounts of hydrophobic chemicals can be sorbed to organic carbon in small colloids that may pass through filters used to separate phases. These sorbed chemicals may be operationally-defined as 'dissolved' which can skew a calculation of Koc and bias it low. Care was taken in the SBLT report to limit this bias. On the other hand, sequential filtration of volatile chemicals can lead to increased loss of dissolved chemical, leading to Koc calculations for volatile chemicals being biased high. Therefore, if additional handling of laboratory samples increases a calculated Koc (as a result of decreased dissolved chemical) it may not be possible to determine if the increase is a result of removal of colloids or volatilization of chemical. To be assured that volatilization losses have not occurred in the laboratory during phase separation, it is important to conduct mass balance on chemicals, especially volatile chemicals.
Serial filtration of samples is appropriate for high Kow compounds such as benz[a]anthracene (log Kow = 5.91), but does not appreciably improve estimates of dissolved concentrations for chemicals with low Kows such as naphthalene (log Kow = 3.36). In addition, naphthalene may be lost from the dissolved phase during successive handlings which could decrease the dissolved concentration. Therefore, successive filtrations with 0.7, 0.5, and 0.22µm filters, such as was done in the SBLT is not advised for characterizing dissolved naphthalene concentrations.
Data from the colloid study in the SBLT Report support this analysis. Naphthalene concentrations in operationally-defined dissolved samples from Stryker Bay were about 22,000 µg/L (100%), 9400 µg/L (42%) and 4,500 µg/L (20%) after being filtered with 0.7µm, 0.5µm and 0.22 filters successively. Benz[a]anthracene concentrations in operationally-defined dissolved samples from Stryker Bay were about 4.1 µg/L (100%), 4.0 µg/L (96%) and 0.24 µg/L (6%) after being filtered with 0.7µm, 0.5µm and 0.22 filters successively. The Kow of benz[a]anthracene is 355 times greater than the Kow of naphthalene, and yet the ratios of the amount lost following sequential filtration differ by only a factor of 4. This suggests that sequential filtration of sediment containing naphthalene may not only remove colloids, but may also release significant amounts of volatile naphthalene. (It is uncertain, but doubtful, that this sample contained naphthalene in a NAPL, because naphthalene in a NAPL (~ 5% naphthalene) will equilibrate with about 5 - 11 mg/L dissolved naphthalene.)
The Stryker Bay SBLT data in Data Gaps Report SBLT Table 4-4 show that the dissolved naphthalene concentration does not change through successive leachings of sediment. Therefore, the dissolved concentration is likely limited by the presence of naphthalene in a NAPL or by sediment concentrations that can saturate water at equilibrium. Because successive leachings result in similar dissolved concentrations between 5 and 11 mg/L (0.7 µm filtered concentrations of 8,000, 9,500, 7,500 and 7,800 µg/L) it is likely that the sediment used in this test contained a NAPL. In addition, because the initial sediment concentration was 3,300 mk/kg naphthalene, and the final sediment concentration was 1600 mg/kg, and yet the dissolved concentrations of naphthalene were similar (8,000 and 7,800 µg/L), either: the sediment contained a NAPL; the system did not reach equilibrium, or; there were discrepancies between the 2 tests. It is clear from the data that, in the case where there was no NAPL, the sediment concentration used to calculate a Koc for naphthalene in Stryker Bay must be below the lowest measured sediment concentration for the Stryker Bay sample, i.e. less than 1600. Therefore, the resulting Koc must be below 892 L/kg (naphthalenesed / Toc / naphthalenewater; 1600 / 23% / 7.8). The Koc of these sediments is expected to be greater than the published value of 933 L/kg (ATSDR, 1995) because the sediments contain considerable anthropogenic carbon. Yet these results suggest that a Koc calculated from this sample is equal to or less than the published Koc. This may also suggest the presence of a NAPL in this sample.
The Data Gaps Report calculated a Stryker Bay site-specific Koc from the pre-leaching SBLT sediment concentration (3,300 mg/kg), and the assumed maximum dissolved naphthalene (4.5 mg/L; see above, colloid test). The resulting Koc that is cited in other sections of the Data Gaps Report is 3,200 L/kg (3,300 / 23% / 4.5 = 3,200). This Koc was calculated using improper data and should not be used.
The discussion in this appendix suggests the importance of discerning if PAHs in sediment are in a NAPL, or sorbed to the sediment. In addition MDH recommends that because there are no reliable data from which to calculated a site-specific Koc, published Kocs be used at this site.
Agency for Toxic Substances and Disease Registry (1995). Toxicological Profile for Naphthalene, 1-Methylnaphthalene and 2-Methylnaphthalene. Science International Inc., U.S. Department of Health and Human Services, Public Health Service, Atlanta, GA. August 1995.