Sediment Operable Unit
ST. LOUIS RIVER/INTERLAKE/DULUTH TAR
NATIONAL PRIORITY LIST (SUPERFUND SITE)
ST. LOUIS, ST. LOUIS COUNTY, MINNESOTA
The St. Louis River Interlake-Duluth Tar (SLRIDT) Site (the site) is a National PrioritiesList (Superfund) site on the St. Louis River in Duluth, Minnesota. Beginning in about1904, pig iron, coke, and gas were produced at this site from iron ore and coal.
Over the years, different areas of the site have been open water. Currently the sitecontains three embayments (see Attachment #1): Stryker Bay (approximately 35 acres),Hallett Dock Company Slip #6 (Slip #6; approximately 23 acres), and Hallett DockCompany Slip #7 / Keene Creek (Slip #7; approximately 27 acres); as well as about 130acres of land between the inlets. The site lies about four miles upstream from the St.Louis River's outlet into Lake Superior. It is bounded by residential and light industrialareas to the north; residential areas to the west; by the St. Louis River to the south; andby a shipping dock to the east.
Approximately 960 and 4200 people live within 0.5 mile and 1 mile of the site,respectively (2000 census).
Stryker Bay, the westernmost boundary of the site, extends about one-half mile north-south and about 1000 feet in the east-west direction. Typically, within 100 feet of theshoreline, the depth of the bay drops to approximately 4 feet, and the depth across mostof the bay remains a fairly constant 4 - 6 feet. Stryker Bay is shallower at the northernend and forms a small wetland at the northern tip. Storm sewers and a small stream feedthis wetland. Stryker Bay is not in a shipping area. Furthermore, there are no commercialdocks for large lake vessels upstream of Stryker bay on the St. Louis River. There are anumber of residences along the western shore of the bay. Some of these residences haveprivate, recreational docks in the bay. The bay is intermittently used for recreation,including swimming, water-skiing, and fishing.
Slip #6 is located in the middle of the site and extends approximately one-half mile northfrom the St. Louis River , with a relatively uniform width of 250 feet. The Hallett DockCompany currently uses Slip #6 as a loading facility for ships traversing the Great Lakes,as well as ocean-going ships.
The western portion of the Slip #7 / Keene Creek (Slip #7) embayment is part of the site.Slip #7 has been used in the past as a loading facility. The eastern side of Slip #7 iscurrently used for docking barges. Keene Creek has been redirected so that it no longerflows into Slip #7. Area residents have been observed fishing in this bay.
Seiches (a sudden rise in water level), caused by wind forces shifting water in LakeSuperior, can cause 7.9 hour periodic water level oscillations of 3 centimeters (cm) to 25cm in the St. Louis Harbor area. This oscillation is considered the driving force for masstransport in the lower St. Louis Harbor (Stortz and Sydor, 1980). Near the site, seichescan cause water level fluctuations of up to 15 cm. As the water level changes, relativelystrong currents can be expected in narrow areas like the mouth of Stryker Bay (MDNR,2001). These currents will increase mixing and turbidity and affect the sedimentationprocess. Furthermore, large seiches can result in a reversing of St. Louis River flow. Forexample, upstream flow (26.5 cm/sec) at the site of the old Arrowhead Bridge (1/2 miledownstream from the site) is calculated to be similar to flow downstream (28.2 cm/sec)when a 15 cm seiche is modeled (Stortz and Sydor, 1980). Therefore, contamination fromthis site can be assumed to be dispersed upstream, as well as downstream from the site.
Concerns presently being addressed at this site are related to contaminated sedimentsassociated with historic industrial activities. The forum for addressing these concerns isa Peer Review process that was created by the Minnesota Pollution Control (MPCA)Citizens' Board in November 1999. The Peer Review Team (PRT) includes experts in groundwater, dredging, capping and pricing of remedial alternatives. Data is presentedby the MPCA and the potentially responsible parties (PRPs) to the PRT for review. Atthe request of the MPCA, MDH is also reviewing documents presented to the PRT andattending PRT meetings. An extensive amount of study has been conducted by the PRPsto provide information about the effectiveness and cost of proposed remedies. Thishealth consultation reviews the product of those studies: the Data Gaps Report (Service,2002).
MDH is specifically concerned about potential human exposure to contaminants at thepresent time, during any cleanup activities, and in the future. Therefore, our recentquestions have been on data gaps that may significantly affect estimates of humanexposure to contaminants or the longterm effectiveness of remedial actions. MDH haspreviously raised concerns about the cleanup of these sediments: in testimony to theMinnesota Pollution Control (MPCA) Citizens' Board (Attachment #2, (MDH, 1999));and in a health consultation (MDH, 2001b), written in cooperation with the U.S. Agencyfor Toxic Substances and Disease Registry (ATSDR).
One of the focuses of the Data Gaps Report (Service, 2002) is the potential publicexposure to contaminants that would be emitted during dredging of contaminatedsediments. This health consultation is a review of experimental data, the proposedemissions model and conclusions of the Data Gaps Report. In addition, this HealthConsultation contains a brief review of chemicals of concern; identifies routes ofpotential exposure; reviews concerns that are dependent on the choice of remedial option, and; provides MDH recommendations.
Human health may be adversely impacted when people are exposed to significantamounts of toxic chemicals. Because individual chemicals behave differently in theenvironment, potential routes of exposure are chemical dependent. Therefore,quantification of human health risks associated with contaminated sediments may requireanalysis of several routes of exposure. These include ingestion of chemicals partitionedinto sediments or surface water; dermal exposure to chemicals in sediment or surfacewater; inhalation of volatile chemicals partitioned first into surface water and then intoair, and/or; ingestion of fish or other foods contaminated by chemicals. If exposures arelimited by removal of contaminants from areas where exposures may occur or bycreating barriers to exposure, the probability of adverse health effects is also limited.
Chemicals of concern
Previous MDH documents have identified polycyclic aromatic hydrocarbons (PAHs) andmercury as the primary chemicals of concern for human health at the SLRIDT site. Polychlorinated biphenyls (PCBs), polychlorinated dibenzodioxins (PCDDs),polychlorinated dibenzofurans (PCDFs), octachlorostyrene (OCS), andhexachlorobenzene (HCB) are also potential chemicals of concern at this site and mostcontaminated aquatic sites in the St. Louis River area. A survey for these 5 chemicalshas not been conducted at SLRIDT. All 7 of the above listed chemicals and chemicalgroups are addressed in greater detail in the 2001 MDH Health Consultation (MDH,2001b).
The chemical listed above are of concern not only because of potential risks associatedwith exposure to them at this site, but also because they may be transported offsite intothe St. Louis River and then to Lake Superior. Canada and the United States havedeveloped a Binational Toxics Strategy (BNS) "...toward the goal of virtual eliminationof persistent toxic substances resulting from human activity, particularly those whichbioaccumulate, from the Great Lakes Basin, so as to protect and ensure the health andintegrity of the Great Lakes ecosystem" (http://www.epa.gov/glnpo/p2/bns.html). Thechemicals of concern are all Level I substances targeted by the BNS.
Naphthalene and similar PAHs
MDH has previously reviewed polycyclic aromatic hydrocarbon contamination data atthis site and discussed toxicity issues (MDH, 2001b). PAHs are a class of generallyhydrophobic, non-polar organic compounds that result from the burning of organicmaterials. They exist in mixtures in the environment. These mixtures have been shownto have both carcinogenic and non-carcinogenic effects on humans and animals. Historically, chemical analysis has been restricted to a handful of individual carcinogenicPAHs (cPAHs) and non-carcinogenic PAHs (nPAHs) out of the hundreds of PAHs foundin the environment. However, in early 2000, MDH began recommending PAH analysisinclude an extended list of 25 cPAHs with available California Office of EnvironmentalHealth Hazard Assessment (OEHHA) potency slopes or potency equivalency factors(PEFs) (CA OEHHA, 2002). MDH finalized recommendations in a 2001 memo toMPCA (MDH, 2001a) and cited this memo in the 2001 Health Consultation on SLRIDTsediments (MDH, 2001b). Recent samples from SLRIDT have not been analyzed forthese additional cPAHs (see below).
Naphthalene is the PAH of most concern during cleanup activities because it is veryvolatile and prevalent in sediments. Additional PAHs, primarily methylatednaphthalenes, will have similar volatilities, but they do not appear to be as prevalent insediments at this site.
Naphthalene is currently classified as an nPAH by EPA, but the International Agency for Research on Cancer (IARC) classifies naphthalene as possibly carcinogenic to humans (Group 2B) due to the evidence of carcinogenicity in laboratory animals exposed for long periods of time, at high concentrations (IARC, 2002). Current MDH criteria have been developed from data showing naphthalene to cause nasal and respiratory effects in rats exposed to low concentrations for 2 years, or exposed to much higher amounts for 4 hours. Using standard risk assessment methods, MDH has developed a shortterm (acute) health criterion of 200 micrograms per cubic meter (µg/m3), and a longterm (chronic) health criterion of 9 µg/m3 for naphthalene (Attachment #3; MDH, 2003). Exposures to air concentrations below these values are expected to be safe. A few people may be affected by naphthalene at concentrations around the odor threshold (between 200 and 440 µg/m3). It is likely that most people will not be affected by acute naphthalene exposures well above the acute criterion of 200 µg/m3.
Mercury and methyl mercury are neurotoxins and exposures are of particular concernduring fetal and early postnatal development. The methylated form of mercuryaccumulates in biota because it is readily absorbed and only slowly excreted. As a result,concentrations of methyl mercury are the highest in animals at the top of the food chain. Ingestion of fish tissue containing high levels of methyl mercury is the dominantexposure to mercury for most people. The MDH Fish Consumption Advisory has morerestrictive advice for women of child-bearing age and children than for the generalpopulation. In general, women of child-bearing age and young children are advised notto eat walleyes greater than 25 inches in length, and to limit eating walleyes 15 to 20inches from the lower St. Louis River to one meal a month or less. In addition, theadvice for the general population advises limiting consumption of walleye larger than 20inches to once a month. This advice is somewhat less stringent than fish consumptionadvice for the St. Louis River prior to 2001, and is based on changes in methods ofcalculating fish consumption advice as well as data from a limited number of fish.
Limited sediment mercury data show that mercury in sediments in Stryker Bay iselevated above current background levels (e.g. REMAP, 1995). Furthermore, asdiscussed in the previous Health Consultation (MDH, 2001b), industrial activities at thissite were, potentially, a large source of mercury to the local area and the region.
Mercury is of concern if it is converted to methyl mercury and enters the food chain. Methyl mercury accumulates in the food chain, with high trophic level fish having muchhigher concentrations of methyl mercury than sediments (USGS, 2001). Methyl mercuryconcentrations in sediments can change over the course of hours, days, months and years,because it is produced from inorganic mercury by bacteria (Gilmour et al., 1992) and it isreconverted to inorganic mercury biotically and abiotically (Matilainen and Verta, 1995;Sellers et al., 1996; Regnell et al., 1998). Therefore, changes in methylation ordemethylation rates may have a greater impact on fish tissue concentrations than theamount of inorganic mercury in sediments.
The National Oceanic and Atmospheric Administration (NOAA, 1996) has stated:
In order to select an effective remedial alternative, it may be necessary to characterize themajor pathways to receptors of concern and aspects of the aquatic system that enhancemethylation and influence mercury availability.
The role of [mercury] speciation in determining concentrations in, and toxicity to, biotamay need to be understood prior to attempts to control the geochemical cycling ofmercury within a waterbody. Remediation attempts have been unsuccessful at siteswhere these factors have been ignored.
MDH recommended in the 2001 Health Consultation (MDH, 2001b) that regularsediment analyses include an extended list of 25 carcinogenic PAHs (including 6 alreadyanalyzed and 19 additional cPAHs), and that sediments at the site be screened for anumber of chlorinated organic chemicals (i.e. polychlorinated biphenyls (PCBs),polychlorinated dibenzodioxins (PCDDs), polychlorinated dibenzofurans (PCDFs),octachlorostyrene (OCS), and hexachlorobenzene (HCB)). These additional chemicals ofconcern listed in Table #1, were not addressed in the Data Gaps Report.
|Polycyclic Aromatic Hydrocarbons||Chlorinated organics|
Potential routes of exposure
Six different potential routes of exposure to chemicals from contaminated sediments inthe St. Louis River and at the SLRIDT site have been identified. Based on MDHsediment screening criteria, a limited number of important (driving) pathways wereidentified for each chemical of potential concern. Table #2 lists the potentiallysignificant routes of concern for those chemicals.
|Chemical / |
|Ingestion of chemicals in sediment||Dermal exposure to chemicals in sediment||Ingestion of chemicals in fish tissue||Ingestion of chemicals in water||Dermal exposure to chemicals in water||Inhalation of chemical vapors|
|Polycyclic aromatic hydrocarbons (PAHs)||*||*||*||*|
|Polychlorinated dibenzo-dioxins (PCDDs)||*|
|Polychlorinated dibenzo-furans (PCDFs)||*|
|Polychlorinated biphenyls (PCBs)||*|
|Mercury / methyl mercury||*|
The Data Gaps Report (Service, 2002) contains reviews and reports of experimental data. Currently, two general remedial actions are being discussed for the contaminatedSLRIDT sediments: dredging and removal of contaminated sediment to a containeddisposal facility (CDF) - either onsite or offsite; or capping contaminated sediments inplace. Dredging and capping have the potential to cause future exposures in verydifferent ways.
Dredging will move most of the contaminated materials to a structure (e.g., tomb)designed to isolate them from the environment. The Peer Review Team had lengthydiscussion about the redeposition, or fallback, that may be expected during dredging(PRG, 2002). While there is evidence that dredging will leave contamination behind,MDH has been unable to locate any studies on the extent of "fallback" at dredging sites. The Data Gaps Report does not address this issue, and discussion of the magnitude of theproblem and possible solutions is needed.
On the other hand, the movement of contaminants away from the dredging area can besomewhat controlled during dredging (e.g. by use of hydraulic dredging), and ifnecessary, controls can be added so that there is very little movement of these suspendedmaterials offsite (National Research Council, 1997; pg 113-116). The biggest publichealth concern during dredging operations is that volatile contaminants (e.g. naphthalene)may escape from the water/sediment slurry and be emitted into the air (National Research Council, 1997; pg 112); and people can be exposed by breathing the contaminated air.
Capping of contaminated sediments will leave 100% of the contaminants on site, under ablanket of possibly 3-4 feet of clean materials. Long-term capping of contaminants in anaquatic environment, particularly a shallow environment, can be more difficult thanentombment in an upland area. Water flow around or through contaminated sedimentsand a cap can move contaminants or erode a cap. In addition, application of a cap andcompaction of underlying layers will displace porewater, and possibly non-aqueousphase liquid (NAPL). Such displacement could potentially cause a short-term releaseand introduce contaminants into the new overlying materials. If, as suggested in the DataGaps Report, emissions outside of the dredged cell are greater than emissions in theactive cell, then small disturbances of the sediment may have a disproportional effect onair emissions. In any event, it is expected that there will be a pulse of contaminants intothe aquatic environment when a cap is applied.
The site is in an area with a vertical groundwater gradient toward the surface, andgroundwater activity below, within and above contaminated sediments is of potentialconcern. Groundwater can potentially transport contaminants and it can also erode thecap. Of additional concern are frost or ice penetration into the cap and the effect of plantgrowth, including woody plant growth, in the cap. These activities can create channelsthrough parts of a cap, and also break up the matrix of a cap. Such activities couldincrease the mixing and erosion potential. Proposed capping materials include silt, whichwould increase the ability of the cap to grow vegetation and the effects of ice intrusion,and sand, that would decrease plant growth and matrix-disrupting effects of frost on thecap. If the contaminants remain on site, monitoring and repair of the cap will benecessary for many generations, or essentially forever.
If sediments are moved into a CDF, the footprint of contamination can be significantlyreduced and future exposures can be limited or eliminated by controlling contaminantaccess and interactions with groundwater, surfacewater, frost, air, plants and animalswith engineering controls. If contaminants are capped in place, the contaminationfootprint remains at its current size, and controls are generally limited to the top of thecontaminated sediments.
If contaminants are left on site, monitoring releases will be necessary while thecontaminants remain. Institutional controls have often proven to be ineffective(Environmental Law Institute, 1999). MDH is concerned that a cap or CDF may not bemaintained, monitored and repaired, and contaminants may become a health hazard inthe future. Therefore, MDH favors a remedial action that permanently restrictsexposures to levels below health concern at this site.
Hybrid remedial designs have been discussed and MDH recognizes that combinations ofdredging and capping may be considered. Unless the issues inherent to both remedies areaddressed in a hybrid plan, hybrid remedial actions may fail to adequately limit risks.
Air emissions experiments and proposed model
Concerns about emissions during dredging were discussed in a 2001 health consultation(MDH, 2001b). The health consultation recommended: "...monitoring of emissionsshould be conducted during the entire [dredging] operation and provisional plans should be prepared for further limiting emissions should emissions exceed levels of concern."
PAHs that volatilize from a slurry of sediment and water will be largely the light,sparingly soluble PAHs that have high Henry's Law Constants (Hk). These willprimarily be naphthalene and methylated naphthalenes. Not only does the Henry's Lawconstant (an equilibrium constant) decrease for larger PAHs, but molecular diffusion ofthese larger molecules is slower, thereby decreasing the speed at which they traverse theboundary layers.
The Data Gaps Report Air Analysis reports on emissions experiments conducted in thelaboratory and applies the resulting data to two models. The Data Gaps Report presents amodel for air emissions during dredging, then uses emissions estimates from that modelas inputs for the EPA ISCST3 air dispersion model. MDH has not reviewed theapplication of the air dispersion model to emissions data from this site. MPCA technicalstaff are better qualified to review air dispersion models. This review focuses onexperiments and theory used in the Data Gaps Report to develop emissions estimates fordredging.
The greatest errors in modeling ambient air concentrations are usually the inputs into thedispersion model. When the source is contaminated sediments, errors can be in theestimates of sediment concentrations from which emission estimates are extrapolated, orerrors in the model used to estimate emissions during dredging. Available data on PAHconcentrations as model inputs for Data Gaps Report are very limited. For the purpose ofmodeling air emissions, Stryker Bay was divided into 15 dredging cells, eachapproximately 3 acres. There are no sediment data for 2 of the 15 proposed dredgingcells in Stryker Bay. In addition, analytical data from a total of 38 point locations wasused to characterize some 35 acres of the bay. Furthermore, data are not available onwhether and where the contamination is in a non-aqueous phase liquid (NAPL).Therefore confidence in the emission estimates, even from a proven emissions model,would be limited.
Experimental data review
The Data Gaps Report contains experimental data for PAH and volatile organiccompound (VOC) emissions from sediment slurries. Emission data are reported for fourdifferent slurries with targeted concentrations of 45% (0.45 kgsed/kgslurry), 8%, 8% and 1%solids by weight (i.e. different volumes). One of the 8% slurries was stirred throughouttesting, while the remainder were stirred immediately prior to testing and allowed tosettle out during the experiments. Slurries were made from a single homogenizedsediment sample and diluted with St. Louis River water. The sediment used as the PAHand naphthalene source was taken from one of the most contaminated areas in StrykerBay. The Data Gaps Report does not indicate whether the sediment used in theexperiment contained NAPL.
The sediment slurry concentrations used to estimate emissions at different locationsduring dredging seem reasonable representations of sediment solids concentrations indredged materials. However, the accuracy of the experimental slurry concentrations andfactors which result in slurry concentration variation during dredging were not discussedin the Data Gaps Report, and MDH does not have experience with dredging operations. The 1% slurry, in particular, may be a very high estimate for suspended solidconcentration in the vicinity of a dredging operation, and emissions from that slurrylikely overestimate emissions in areas not being dredged. The National ResearchCouncil Report on contaminated sediments states that in most studies, suspended solidsare less than 100 milligrams per liter (mg/L) except in immediate proximity to thedredging head, and, in most field studies, less than 10 mg/L has been measured 100meters from the dredging head (National Research Council, 1997; pg 110). Therefore,while the model indicates that contribution of emissions distal to dredging is smallrelative to total emissions in the model, it actually exceeds emissions in the active cell(see Table #4). Basing estimates on emission rates from a 10 grams per liter (g/L) (1%)slurry in the inactive cells may overestimate suspended solids contributions from this partof the model by 2 to 3 orders of magnitude.
In addition because a very small amount of particulates are transported outside of thedredging area, naphthalene in water outside the active cell will be dissolved. Dissolvednaphthalene volatilizes rapidly, so it is unlikely that significant amounts of dissolvednaphthalene will remain in water far from the active cell. Therefore, the use ofexperimental data from a 1% slurry to describe volatilization over about 42 acres outsideof a 3 acre dredge area, appears to be a large overestimation. Use of 1% slurry emissionsdata for cells bordering the active cell should provide a very conservative estimate ofnon-active emissions from those cells. Current emissions from the rest of Stryker wouldprobably provide a reasonable emission rate estimate for the non-adjacent cells in StrykerBay.
Experiments described in the Data Gaps Report Air Model appendix were intended tohelp quantify naphthalene emissions from dredging operations and a CDF. Emissionrates from sediments were estimated using an experimental apparatus that allowed asteady air stream to flow over sediment / water slurries and subsequent air samplecollection for analysis. In this manner it is possible to measure potential air emissionsunder controlled conditions.
It is critical to an experiment that the initial quantities of the chemicals of interest areknown. In addition in complicated experiments where errors can occur, it is useful tomeasure the amount of chemical in different compartments (e.g. sorbed and dissolved) atthe beginning and end of the experiment. Minimally, the amount of chemical in thesediment slurries at the beginning of each experiment and the amount remaining at aftereach experiment should have been measured. In this way the mass balance of anexperiment could have been checked. Monitoring mass balance is extremely importantin experiments with multiple physical compartments and a potential different preparationof samples or volatile losses.
In the Data Gaps Report air emissions experiments, not only was no mass balanceperformed, the actual quantity of chemicals available in the slurries prepared for theexperiments is uncertain. While Data Gaps Report Table (DGR Table) A1-1 shows areasonable match between the targeted and measured solids composition of the slurries,DGR Table A2-3 shows a very poor agreement between their targeted chemicalcomposition of the slurries and the measured chemical composition of the slurries. Because water used to make up the slurries contained very little PAHs, the sedimentfraction should initially contain all the PAHs in the slurries. Furthermore, the PAHconcentration in the sediment (solid) fraction should be the same for all slurries. A totalof 2.1 kg of material, sediment, and water, was used in each slurry mixture tested (DGRSection A1-1.3). Dried sediment has about 1.5 times the density of water; therefore, thetotal amount of sediment in each experiment was approximately 980, 172, 189 and 32grams (g) for the 45%, 8% mixed, 8% quiescent, and 1% slurries (calculated frommeasured solids composition (Table A1-1) and total mass of 2.1 kg). The volume ofeach experimental slurry was not reported.
The targeted amount of each chemical (including naphthalene) in each experiment can be calculated from the assumed concentration in source sediments and the amount of source sediments in each slurry. DGR Table A2-3 shows the results of chemistry analyses of source sediments. This table shows analytical data for two 'solids' samples (bulk and 45% sediment; reported in milligram chemical / kilogram sediment (mg/kg)) and three 'liquid' samples (micrograms per liter (µg/L)), for which no solids or moisture data are available. The Data Gaps Report implies that DGR Table A2-3 includes all of the chemical data for each experimental slurry. If this is the case, there is a large discrepancy between the measured amount of naphthalene and the targeted amount of naphthalene in the experimental slurries. Table #3 below compares the total amount of naphthalene in each slurry (calculated from DGR Table A2-3 and total experimental mass of 2.1 kg for slurries) with its targeted amount.
|Solids in Slurry||Bulk sediment||46.80% (of 2.1 kg)||8.20% (of 2.1 kg)||9.00% (of 2.1 kg)||1.50% (of 2.1 kg)|
|Measured naphthalene||(1300 mg/kg)||10811 mg/test||32 mg/test||173 mg/test||15 mg/test|
|Target naphthalene||(11000 mg/kg)||10811 mg/test||1894 mg/test||2079 mg/test||347 mg/test|
If there are no more analytical data on the chemical mixtures in the slurries than the datadescribed in DGR Table A2-3, comparison of emission data between experiments isproblematic. In addition, the use of data from these experiments to describe emissionsduring dredging becomes problematic as well. Note in Table #3 (and DGR Table A2-3and DGR Section A2-3.2) that bulk sediment was analyzed. However, chemical analysisof the bulk sample showed about 9 times less naphthalene than the 45% sedimentsamples. While the Data Gaps Report explained that the bulk sample was not mixedproperly, there is no way to tell if the sediment added to the other slurries was mixed. The Data Gaps Report says "for purposes of air modeling, it is assumed the sourcesediment is as characterized by sample '45-SED'." Without offering data orcorroboration, the Data Gaps Report suggests that the chemical concentrations measuredin the 45% sediment sample are "more consistent with the other analytical results." Massbalance is very important in experiments that characterize contaminant movement, andfailure to accurately measure chemicals in source media makes the usefulness of theseexperiments to describe air emissions during dredging questionable.
There is another possible explanation for the confusing 8% and 1% slurry data in DGRTable A2-3. The laboratory report on these slurry samples says that the media wereliquid, and suggests by the absence of solid and moisture data that the analysis may havecharacterized dissolved chemicals including naphthalene. This would suggest that thesolubility of naphthalene in the slurries was up to 86 mg/L (from DGR Table A2-3). Clearly this is not possible, as the solubility of naphthalene in water is 31 mg/L. Any co-solvent effect that could elevate the amount of naphthalene in solution significantlywould require considerable (maybe 1% or more) organic solvent dissolved in the aqueousmixture. The data discrepancies outlined in Table #3 suggest: non-homogeneoussediment; different pretest preparation of samples; different methods for samplingslurries for analysis, or; analytical problems. Therefore, it is difficult to draw any conclusions from the chemical data.
Emissions model description
The Data Gaps Report proposes a model for use in quantifying naphthalene emissionsfrom sediments during dredging. This model is based on a proposed linear relationshipbetween sediment concentration of naphthalene and the amount of emissions up to amaximum emission rate, above which any increase in naphthalene concentration insediment will not increase emissions. The Data Gaps Report describes the sedimentnaphthalene concentration at which the maximum emission rate is first achieved as the"breakpoint." Further, the Data Gaps Report says that the emissions experimentsdescribed above were conducted at naphthalene sediment concentrations above the"breakpoint." The Data Gaps Report asserts that as the concentration of naphthalene insediment increases, the movement of naphthalene from sediment to water becomesrestricted as the naphthalene concentration in water approaches the solubility ofnaphthalene in water. Therefore, the "breakpoint" where maximum emissions occur canbe calculated from the maximum solubility of naphthalene in water.
The following rearranged equilibrium partitioning equation was the basic equation usedin the Data Gaps Report to calculate the naphthalene sediment concentration"breakpoint."
Solubility x Koc x foc = [naphthalenesed] (@ "breakpoint")(1) -equation #1
The Data Gaps Report modified the above equilibrium partitioning equation to include anadditional factor (a solubility correction factor) that represents solubility interactionsbetween different PAHs (% naphthalene in the non-aqueous phase liquids (NAPL)). Justification for this additional factor of 0.052 is not clear. Furthermore, if a"breakpoint" is used in this model without sound justification, it should have beensupported experimentally. The resulting equation used to calculate "breakpoints" for theData Gaps Report (Section A2-3.3.3 Source Strength Adjustment) is:
Solubility x % naphthalene in NAPL x Koc x foc = [naphthalenesed] (@ "breakpoint") -equation #2
The 'solubility' proposed in the Data Gaps Report (7 mg/L) is a naphthaleneconcentration measured in a sample from the site. It is stated that this measuredconcentration is "the highest measured naphthalene solubility (7 mg/L) in any of thesolubility related tests (DRET, SBLT)" (DGR Section A2-3.3.3). Actually, 22 mg/Lnaphthalene was recorded in a 0.7 m filtered sample from the SBLT (DGR SBLT Table4-4). See Appendix 2 of this health consultation for further discussion.
The dissolved naphthalene concentration in the air emission experimental slurries wasnot measured, and the assumption of a dissolved concentration less than the solubility ofnaphthalene is problematic. Measuring the solubility of naphthalene is difficult since thechemical is volatile. The solubility for naphthalene in water is about 31 mg/L (31.7mg/L; ATSDR, 1995). While this solubility cannot be decreased, it can be increasedabove 31 mg/L if there is substantial co-solvent present. However, when a NAPL ispresent, the equilibrium concentration in water may be considerably less than thesolubility (i.e. 5-11 mg/L: more discussion in a later section of this document).
In addition, the Data Gaps Report proposes using a site-specific Koc. Determination ofKoc from laboratory experiments is not a simple task. MDH reviewed the Kocexperiments conducted for the Data Gaps Report, and concluded that they werecalculated using incorrect data as well as data from experiments that contained errors inmethods. Appendix 2 contains a discussion of the derivation of the Data Gaps Report"site-specific Koc," including discussion about why Koc cannot be determined fromsamples containing NAPL.
The Data Gaps Report emissions model assumes that the naphthalene concentration insediment in the experimental slurries was the target or "nominal" concentration, andcalculated the hypothetical emissions total during dredging for 2 different "breakpoints"(238 mg/kgsed and 1000 mg/kgsed naphthalene). Figure #1 shows the proposedrelationship between naphthalene sediment concentrations and total naphthalene airemissions (first 2 hours) proposed by the Data Gaps Report for the 238 mg/kgsed"breakpoint." Note that the actual emissions may be different under differentexperimental conditions, but that the "breakpoints" in the Data Gaps Report modelremain the same regardless of experimental conditions (more discussion of this below).
Also, note that for a single experimental data point (triangles in Figure #1), the modelproposes 2 lines: the maximum emission rate between 238 and 11,000 mg/kgnaphthalene in sediment and; a line from the maximum emission rate to zero forsediments with 0 - 238 mg/kg naphthalene. While there are no experimental data thatsupport this model, an analysis conducted in Appendix 1 of this Health Consultationdemonstrates that experimental data demonstrates that the model is incorrect, especiallyin predicting emissions from slurries with low suspended solids concentrations.
For comparison, Figure #2 graphically compares 8% mixed slurry emissions modeledwith "breakpoints" at 238 mg/kgsed, 1000 mg/kgsed, 5,000 mg/kgsed and 15,000 mg/kgsednaphthalene. Note that use of the Data Gaps Report model with a 15,000 mg/kgsed"breakpoint" results in the same predicted emission rates as a model without a"breakpoint" for naphthalene concentrations at the site (< 15,000 mg/kg naphthalene). Note in Figure #2 that there is one experimental datapoint (triangle) from which the Data Gaps Report proposed an emissions model.
Table #4 (below) shows the effect of using different breakpoints to calculate airemissions during dredging.
| DGR model "breakpoint"|
- minimum sediment concentration at which maximum emissions occur -
|15,000 mg/kgnaphthalene (same as no"breakpoint")||5000 mg/kgnaphthalene||1000 mg/kgnaphthalene||238 mg/kgnaphthalene|
|Active emissions during dredging||947||12%||1,403||7%||1,582||5%||2,302||7%|
|Emissions from inactive cells||1,195||15%||1,769||9%||1,996||6%||2,904||9%|
|CDF active emissions||5,161||66%||15,482||77%||25,888||81%||25,888||77%|
|Quiescent CDF pool emissions||469||6%||1408||7%||2,354||7%||2,354||7%|
|Total naphthalene air emissions||7,752||100%||20,062||100%||31,821||100%||33,448||100%|
|Total naphthalene dredged|
(calc from cell avgs)
|Percent total naphthalene dredged that is emitted||3.5%||9.0%||14.3%||15.0%|
It is clear that the Data Gaps Report model is very sensitive to the choice of a"breakpoint". In other words, a low "breakpoint" assumes a very high rate of increase inair emissions as sediment concentrations of naphthalene increase. Appendix 1 of thishealth consultation demonstrates that the proposed model breakpoint of 238mgnaphthalene/kgsediment predicts values that do not agree with the experimental data from theData Gaps Report. If the Data Gaps Report model is correct, better experimentalconfirmation than presently exists is needed. In order to accept this model right now, onehas to accept very implausible (and likely incorrect) assumptions that are not supportedby the poor quality experimental data.
Partitioning: equilibrium versus dynamic systems
In a closed system at equilibrium, concentration ratios between all compartments areconstant and all net fluxes between compartments are zero. When a closed system isdisturbed and no longer exists at equilibrium, there is movement of chemicals betweencompartments, and ratios of chemical concentrations between compartments will change. Concentrations under non-equilibrium conditions are not determined by equilibriumratios, but are determined largely by the chemical flux between compartments. Atsteady-state, as at equilibrium, concentration ratios between compartments are alsoconstant; however there is a constant rate of movement of contaminants from one phaseto another. Chemical concentration ratios between compartments at steady-state may bevery different than ratios at equilibrium. And certainly, chemical ratios betweencompartments in a dynamic open system are not at equilibrium.
Movement of naphthalene from sediment to air requires the chemical to move betweendifferent compartments: sediment, water, and air. The speed at which a chemical canmove from one compartment to the next is determined by the amount of mixing of thebulk medium (e.g. sediment, water or air), the molecular diffusivity of a chemicalthrough the mediums, and the distance that the chemical must diffuse. Any movementthrough a compartment that is not aided by mixing must be accomplished by diffusion. Therefore, diffusion is usually confined to boundary layers that border the bulk media. Diffusion is also important in stagnant sediments.
Models of mass transfer across these boundary layers and into the different mediadescribe slow diffusion across the boundary layers and rapid transfer from one layer toanother to maintain equilibrium conditions across extremely small areas of adjacentboundary layer surfaces. Therefore, as boundary layers become thinner (a result ofmixing) or molecular diffusivity increases (a function of temperature, typically a minorfactor), flux from one compartment to another increases. This was demonstratedexperimentally by the data from the 8% mixed slurry (assuming targeted naphthaleneconcentrations): mixing increased the emission of chemicals from 8% sediment slurries. These experimental emission rates are marked with triangles in Figure #1. (However,note in Figure #1 that the measured emissions from the 45% and 1% slurries are similar,and the emissions from the 8% quiescent slurry is about 40% less. This suggests eitherinconsistent preparation of samples, non-homogenous source sediments, problems withsample collection, analytical errors, or the presence of a NAPL.)
Experimental data from the Data Gaps Report suggest that emissions were not restrictedby a solubility limit. Solubility remains constant over time. If, as suggested in the DataGaps Report, solubility restricts movement of contaminants from sediment to air,experimental data should show steady emissions of naphthalene. As shown in Figures#3-#6 (chemical emission rates corrected for chemical loss and normalized to emissionrates over the first 2 hours), all experiments had decreasing volatilization of naphthaleneover time. Changes in emission rates must be a result of changes in the system. Thereare two apparent potential causes for this rate decrease: 1) Mixing during the test wasinsufficient to maintain the initial rate of transport of naphthalene through the watercolumn, and 2) there are two (or more) compartments of sediment-adsorbed naphthalenethat release naphthalene at different rates.
If the lack of mixing decreases the naphthalene renewal at the water-air surface orsolubility limits desorption from sediments, the emission rate reduction over time shouldbe less for the 8% mixed slurry than from the 8% quiescent slurry. As shown in Figures#4 and #5, the reduction in emission rate for the 8% quiescent slurry may be less than thereduction for the 8% mixed slurry; however the difference is small and likelyinsignificant (56% and 91% for hours 2-6 and 6-24, respectively, versus 47% and 85%). Therefore, it is unlikely that the lack of mixing or solubility limitations are responsiblefor most of the decrease in emission rates through time.
If naphthalene is sorbed to sediment particles differently, with different affinity fordifferent types of particles (e.g. different types of organic carbon) or with longerdiffusion paths as more desorbs, then experimental data may show less of a decrease inemission rate in experiments with the 45% slurry than in other experiments. However, ifthe rate of transfer through the water column is sufficiently fast, and there is no solubilitylimitation to the desorption from sediment, the difference may be insignificant. Whiledecreases in the naphthalene emission rate may have been the least in the 45% slurryexperiment, variation between all experiments was not large (see Figures #3-#6). Because water mixing does not appear to limit emission rate decreases, by default,slowed diffusion of naphthalene from sediments over time is the suggested explanationfor the emission rate decreases.
There is considerable information in the environmental chemistry literature that showsPAHs, in general, have different affinities for organic carbon of different origins. PAHpartitioning into water has been shown to be decreased in areas with significant tar orpyrogenic materials (Maruya et al., 1996). In addition, research by Gustafsson et al.(1997) and others has shown that soot or black carbon in sediments does not allow PAHsto desorb as readily as natural or other organic carbon.
In summary, the Data Gaps Report experimental data suggest that there is probably not asolubility limit to naphthalene emissions under conditions similar to those encountered inthese experiments. Further, if there is a solubility limit under some conditions, thesolubility limit will change as experimental conditions change. The known mechanismswhereby chemicals partition between different phases or media, as well as very limitedexperimental data, suggest that even if there is a sediment concentration above whichthere will be no increase in emissions (i.e. a "breakpoint"), this sediment concentrationcould be very different under different experimental conditions. Therefore, the single"breakpoint" concept is of little utility when attempting to model emissions duringdredging.
While experiments in the Data Gaps Report were not designed to find a possible cause ofdecreasing naphthalene emission rates over time, the data do suggest that emission ratesmay be limited by diffusion and desorption of naphthalene from the sediments.
Naphthalene partitioning model
Naphthalene partitioning from sediments can be very different if contamination exists asa NAPL, or if it is sorbed to organic particulates in the sediment. Equilibrium betweennaphthalene in a NAPL and water, and equilibrium between sediment organic carbon andwater, are two distinctly different conditions. If there is a NAPL in the sediments, thenpartitioning should be calculated between the NAPL and water, as well as directlybetween the NAPL and air. If there is no NAPL, then the important partitioning isbetween the organic carbon in the sediment and water. These are important distinctions,because the presence of a NAPL can lead to very different emission rates than will beencountered when dredging sediment-sorbed naphthalene. The concentration ofnaphthalene in water at equilibrium (i.e. in a non-dynamic system) with naphthalenesorbed to sediment is:
dissolved naphthalene (mg/L) = [naphthalene] (mg/kg) / Koc (L/kg) / foc
for dissolved < 31 mg/L -equation #3
The concentration of naphthalene in water at equilibrium with naphthalene in a NAPLcan be approximated by (Schwarzenbach et al., 1993):
dissolved naphthalene (mg/L) » Cwsat * Exp((+)DSm/R * ((Tm / T) - 1) * fNAPL * gorg mix
for dissolved < 31 mg/L -equation #4
|Cwsat (mg/L)||= water solubility (31 mg/L)|
|DSm (J/mol/K)||= entropy of melting (~48 - 59 J/mol/K)|
|R (J/mol/K)||= gas constant (8.3145 J/mol/K)|
|Tm (K)||= melting temperature (353.75 K)|
|T (K)||= temperature (293.15 K)|
|fNAPL||= fraction of naphthalene in NAPL|
|gorg mix||= fugacity of naphthalene in the organic mixture (~1 - 1.8)|
Note that the naphthalene concentration in water (e.g. 5 - 11 mg/L @ fNAPL = 5.2 % naphthalene) is almost totally dependent on the fraction of naphthalene in the NAPL and therefore, the dissolved concentration (at equilibrium) will remain constant as long as a similar composition NAPL is present. Experimental data from the Data Gaps Report Sequential Batch Leaching Report shows up to 22 mg/L naphthalene in water (DGR SBLT Table 4-4; see Appendix 2 for discussion). Dissolved naphthalene was not measured in the air emissions experiments, and it is possible that samples contained NAPL. As concentrations of naphthalene decrease in sediments, there is probably a point where there is no longer a NAPL. At those locations, the concentration in water is determined (again, at equilibrium) by the equilibrium partitioning equation (equation #3) above. Figure #7 shows potential dissolved naphthalene concentrations at equilibrium with sediment, with and without NAPL.
NAPL in a slurry can also emit volatile chemicals directly to the air. The potential forthis type of air release was not studied in the Data Gaps Report, or by MDH during ourreview.
The Data Gaps Report emissions model is structured to model emissions fromcontaminants that are sorbed to sediments, or solid phase contaminants. The presence ofa NAPL in the sediments or slurries is not discussed in the air emissions sections of theData Gaps Report, but NAPL could have large effect on emissions calculations. Whenattempting to apply the Data Gaps Report experimental data to the emissions model,estimates of dissolved naphthalene could be off by a factor of 6 depending on thepresence or absence of a NAPL. If there is a NAPL, there is likely a maximumnaphthalene concentration in water that is below the solubility of naphthalene. Thismaximum is not related to the naphthalene sediment concentration, but is reached whenthere is a NAPL present, and is absent when there is no NAPL. Sediment-sorbednaphthalene will be associated with generally higher dissolved concentrations ofnaphthalene than will naphthalene from a NAPL, because naphthalene in a PAH NAPLhas less fugacity than naphthalene sorbed to natural organic carbon. If there is no NAPL,the dissolved naphthalene concentration adjacent to sediment containing 11,000 mg/kgnaphthalene may be about 31 mg/L.(2) The presence or absence of a NAPL is critical inevaluating experimental data, and it may also be needed to calculate conservative andrealistic emission estimates from available physical and chemical data on naphthalene.
Additional air emission issues
- Appendix I, attached, ('Using the Data Gaps Report air emission model to predictnaphthalene emissions') compares experimental data with model calculations, and showsthat the Data Gaps Report air emission model does not accurately predict emissions from sediment slurries.
- DGR Section A2-3.3.2 (Air-side Resistance) Air speed in the thin boundary layer at the surface is considerably less than the air speed higher up. Therefore, it is incorrect to compare pseudo-laminar flow at the water surface of the experimental apparatus directly with wind speed.
- DGR Section A2-3.3.1 (Temperature Variability) Henry's Law Constant describesthe equilibrium partitioning of a chemical between air and water. Henry's Law forindividual chemicals varies as a function of temperature. In addition, diffusivity for aspecific chemical varies as a function of temperature to the 1.75 power (T (as Kº)1.75). The rate of volatilization from water is a function of Henry's Law and a function oftransfer velocity, which is dependent on molecular diffusivity. Therefore at equilibrium,changes in concentrations of chemical in air and in water due to temperature changes arereflected in the change in Henry's Law. But changes in the rate of transfer indisequilibrium conditions are also related to T11.75 - T21.75.
- DGR Section A3-1.2.2 (Use of monitoring) Other sites have monitored PAHsemitted by dredging. The section quotes "highest average" PAH concentrations, but doesnot say what the averaging time of the samples was. However, the interview informationreviewed in the Data Gaps Report suggest that air emissions at PAH dredging sites havenot been problematic.
- DGR Section A3-1.2.4 (Chemical specific Emissions) Naphthalene data citedsuggest that not only are emissions during dredging low (maximum ratio 137 mgnaph/kgsed: 1488 ngnaph/mair3), but that post-dredging levels in the dewatering basin were about ½ ofpre-dredging levels. Current naphthalene concentrations in water and air near StrykerBay should also be measured.
- DGR Section A4-1.4 (Objectives of dispersion modeling) MDH agrees that "Acutestandards and odors (should be) modeled as one-hour concentrations. Chronic standards(should be) modeled as the average of the seven-month construction season."
Increasing windspeed will increase emissions from dredging areas and containmentfacilities. But increased windspeed also increases the dilution of contaminants,thereby decreasing potential exposures. Therefore, inputs to the dispersion model arepotentially complex, and the results are not easily predictable. The highest acuteexposures typically occur when there is an atmospheric inversion and the windspeedis low. Because emissions from water are somewhat dependent on windspeed (windcauses a thinning of the water and air-side boundary layers), emissions at lowwindspeed may be below those predicted by a simple emission model. Conversely,emissions at high windspeed may exceed those in the emissions model.
Analysis of the relative relationship between increasing dilution and increasingemissions as functions of windspeed should be conducted.
Henry's Law constants at different temperatures as well as ratios of Henry's Law andratios of molecular diffusivities at other temperatures with their values at 20º arelisted in the Table #5 below (EPA, 2001).
|Temperature||Hk - naphthalene||Hk Ratio (to 20º C)||Molecular diffusivity ratio (to 20º C)|
|22.4º C||40.48 pa/m3/mol||120%||101%|
|20º C||33.85 pa/m3/mol||- -||- -|
|17.3º C||27.56 pa/m3/mol||81%||98%|
|8.5º C||13.63 pa/m3/mol||40%||93%|
These data suggest that temperature-dependant changes in the rate of volatilization will be related to changes in Henry's Law, and will not be appreciably affected by changes in molecular diffusivity. However, given that Henry's Law is an equilibrium constant, the temperature dependence of the rate of volatilization is very dependent on naphthalene concentrations in the water / air boundary layers. These data suggest that emissions may be significantly decreased if the cells containing the highest quantities of naphthalene are dredged in the early spring or the late fall.
Review of the Data Gaps Report suggests that numerous steps can be taken to decreasethe potential exposure of the public to emissions during dredging. These include:
- Cells containing high naphthalene concentrations can be dredged in the earlyspring or late fall.
- Cells dredged during a single year can be chosen so that both highly contaminatedand relatively clean cells can be dredged each year. All yearly calculations inTable #4 (above) are based on dredging cells in a single year that have maximumnaphthalene concentrations that average 2,990 mg/kg. The average of all of thecell maximums for Stryker Bay is 1780 mg/kg naphthalene. Therefore, there isroom to balance emissions between dredging years.
- CDF emissions are the largest source of naphthalene. Offsite handling of wastesand CDF design alterations, including the use of simple floating covers orgeotextile tubes, could significantly reduce emissions.
- Ambient air should be monitored at all times during dredging. Real-timemonitoring should identify hazardous conditions and trigger responses such as;use of shortterm emission control measures, temporarily shifting dredging to aless contaminated area, stopping activities until meteorological conditionschange, or notification and/or shortterm evacuation of residents.
Summary of additional MDH reviews of the Data Gaps Report
Dredging Elutriate Test Report
The purpose of the Dredging Elutriate Testing (DRET) was to "measure the effects of St.Louis River/Interlake/Duluth Tar Site (SLRIDT) sediments on St. Louis River waterquality when mixed together in the immediate area of the disturbance"(Service, 2002;Appendix DRET, Executive Summary). MDH did not review the DRET Report entirely,but confined review to sections and tables that characterized sediment samples. MDHhas the following comments:
- DGR Table DRET3-2 states that the moisture content of samples collected fromStryker Bay, Slip #6 and Slip #7 were 157%, 55% and 88%, respectively. This is anapparent error that is not identified or explained in the text.
- DGR Table DRET4-3 shows analytical results of three samples (one from each of thebays). Numerous PAHs in samples from Slip #6 and Slip #7 were detected at higherconcentrations in (operationally defined) dissolved phase, than in the total samples. These discrepancies were not explained, and without duplicate or additional samples, theconflicting data cannot be reconciled. In addition, the fraction total organic carbon (ftoc)was only determined for 1 sediment sample. Not only is it important to determine spatialvariability of organic carbon measurements, but analytical variability is also important. (In addition, ftoc was improperly reported as 230,000 mg/L OC in DGR Table DRET4-3).
Minimally, the Data Gaps Report should have identified and addressed the uses andlimitations of these data.
MDH reviewed the original Ice Report in December 2001. Other than the addition ofobservations during the winter of 2001-2002, there are no notable changes in the newreport. The two winters of observation now include an "atypical winter," 2000-2001, and"an uncharacteristically warm (winter) with minimal snow cover," 2001-2002.
The Report includes mention that "where the ice thickness was equal to the water depththe ice was typically frozen to the bottom, eliminating the potential for ice movement." However, there is no discussion of frozen sediments. In addition, while the Ice Reportstates that thermal expansion under minimal snow cover "can create significant lateralforces that produce ice movement, bottom gouge, ridge formation, and ride-up," thereport also states that "Duluth always has a snow cover." This is not entirely true, asevidenced by conditions this winter (Attachment #4; MDH and MPCA, 2003).
The Ice Report does not address concerns raised by MDH with regard to ice intrusioninto sediments or a cap, the potential formation of ice feet, or lenses, and the effect of iceon the sediment matrix. These issues were discussed in the 2001 MDH HealthConsultation (MDH, 2001b).
Additional Data Gaps Report Review
MDH has reviewed most sections of the Data Gaps Report that directly apply to humanexposure to contaminants during dredging and the longterm protectiveness of a remedy atthe SLRIDT site. Currently we are in the process of reviewing some related sections ofthe Data Gaps Report, and if we have additional contributions to make to the ongoingdiscussions we will write an additional review. In lieu of a full review, MDH offers thefollowing comments.
MDH expects to complete a review of the groundwater sections of the Data Gaps Report soon. Initial areas of concern are:
- Groundwater flow can move contaminants through a cap or a CDF.
- Groundwater can disturb the matrix of a cap or a CDF, leading to localized failure.
- Modeling contaminant flow through a cap with a one-dimensional model can notaccount for heterogeneous and non-contiguous layers at the site.
- There appear to be numerous springs in the bays and slips. These springs wereshown in a temperature survey of the bottom water at the site conducted in 2000, andtheir existence is supported by observations in Stryker Bay during an MDH site visit(Attachment #3; MDH and MPCA, 2003).
Extent of sediment contamination
The Data Gaps Report (GH 7-3.1, GH 7-3.2) states that there is a lack of associationbetween metals data in layers and total PAH (tPAH) data in the same layers andconcludes that there is no association between releases of tPAH and metals onsite. Thelack of co-location of contaminants does not exclude the probability that both metals andorganics were discharged by facilities that operated at this site. Co-location ofcontaminants assumes that transport of these contaminants to sediments is the same, andthat movement in sediments is the same. While movement of non-polar organics can beeasily modeled, the partitioning and movement of metals and metalliccompounds/species can be more complicated. Some metals may remain relativelyinsoluble as long as conditions remain reduced and there is sufficient sulfide. Groundwater and seiche flow through contaminated sediments can make the predictionof metallic species and movement in sediment even more difficult.
Information not contained in the Data Gaps Report
There are limited data on the mercury contamination in sediments at the SLRIDT site. Any sediment sampling and analysis performed at this site should include analysis ofmercury.
In the MDH 2001 Health Consultation (MDH, 2001b) on sediments at the SLRIDT site,MDH discussed concerns about building a wetland over sediments contaminated withmercury. The most serious concern related to mercury at the SLRIDT site is that creatinga wetland cap may increase the methylation rates of mercury from the site. Wetlandshave been shown to be the major source of methyl mercury in watersheds (Rudd, 1995;Saint Louis et al., 1996). In addition, field studies have suggested that groundwater flowmay increase the flow of mercury into a system (Krabbenhoft et al., 1998). Mercury insediments can form many different sulfide complexes, including some that are soluble(Paquette and Helz, 1997; Benoit et al., 1999). While MDH does expect that mercuryconcentrations in surficial sediments will be lowered with remedial action, we areconcerned that methylation rates may be increased. If Stryker Bay is converted into awetland because of the construction of a cap, methylation rates are expected to rise. Ifmethylation rates increase, it will be important to limit the amount of mercury availablefor methylation by maintaining low concentrations of total mercury in surficialsediments. The Data Gaps Report only addresses mercury by suggesting, with the one-dimensional groundwater model, that the movement of mercury through a cap will beminimal.
Dredging also offers some challenges for containment of mercury. Some mercury willbe resuspended during dredging. However, using a conservative estimate, the totalamount of mercury discharged in decanted dredged water will be at most about 35 grams(MDH, 2001b). A CDF should be designed to isolated contaminants, including mercury,from groundwater and surfacewater. Following dredging, maintenance of current waterdepths should not result in a long-term increase in methylation.
MDH General Comment on the Data Gaps Report
Review of the experimental studies in the Data Gaps Report generally shows inadequatelaboratory procedures and incomplete explanation of methods and results in sectionsreviewed by MDH. A significant amount of information in the reviewed studies isimported from other sections or reports that have not been reviewed by MDH. MDH isconcerned that there is a cascading of effects that is not clearly presented in the DataGaps Report - where errors in methods, analysis, or interpretation are carried over andimpact other calculations and interpretations. This segregation of errors, and lack of fulldisclosure of the sources and limitations of data, can lead to improper and unjustifiedconclusions.
For instance, computation of the naphthalene site-specific Koc used in the Data GapsReport Air Analysis Appendix contains significant errors. Appendix 2 of this healthconsultation contains a review of the site-specific Koc derivation. This review concludesthat the data presented are not sufficiently reliable to replace a more general, published,carefully derived Koc of 933 L/kg (ATSDR, 1995) with the Koc (3200 L/kg) used in theData Gaps Report.
As mentioned elsewhere in this health consultation, any partitioning study should includea mass balance component or study. Not only did the air emissions experiments notinclude a mass balance study, only the emissions appear to have been appropriatelycharacterized. Furthermore, the presence or absence of a NAPL will significantly affectexperimental results and air emission model predictions. The Data Gaps Reportcontained no discussion of the presence or absence of NAPL in sediments that may bedredged. Chemical composition data for the slurries used in the experimental studieswere not reliable. This makes drawing quantitative conclusions impossible. Someinferences can be made by comparing emission rates within individual studies, as done inFigures #3-#6, and by then comparing normalized results, but any conclusions can not bestated with confidence.
As part of an MDH Health Consultation, issues related to children's health are addressedexplicitly. Children are exposed to chemicals differently than adults. Children willsplash and play in water, sometimes drinking it and any suspended particles. They alsoeat, drink and breathe more than adults on a body weight basis. Therefore, they mayhave significantly greater exposures than adults. In addition children are not fullydeveloped, and developing organs may be more sensitive to chemical hazards.
Scenarios for exposure to chemicals in sediments or other media are evaluated by MDHusing a child receptor. Activities of a child and child intake rates are used in determiningpotential exposures by all routes of exposure. In addition, criteria used to evaluate thetoxicity of chemicals are calculated to be protective of children and other sensitive sub-populations.
The Data Gaps Report contains a large amount of information characterizing a number ofissues related to cleanup of the SLRIDT site. MDH reviewed sections of the Data GapsReport that discussed topics previously raised by MDH. These included sections onpotential air emissions during dredging; ice impacts on a cap over contaminatedsediments and; the effect of groundwater on remedial options.
MDH review of the air emissions experiments identified several instances for whichsampling methods were poorly described or not available, targeted naphthaleneconcentrations in experimental samples were not achieved, and there was incomplete ormissing data. Problems encountered in the experiments or in data analysis were notidentified or discussed in the text of the report. In addition, the proposed Data GapsReport air emission model contains significant errors and is inconsistent with theexperimental data in the Report.
No chemical mass balance was performed in the experiments conducted for the DataGaps Report. This is a significant weakness: there is no way to verify that the design andmethods allowed analyses of intended compartments. It is not possible to resolveapparent errors in chemical analyses of the 1% slurry and both 8% slurries withoutadditional data. The Data Gaps Report suggests that target concentrations betterrepresent the actual concentrations of contaminants in sediment than do the analyticalmeasurements, so it is assumed that there was an error in methods in sampling theslurries. These errors do not affect the conceptual or the comparative analysis ofnormalized data within samples, but they may affect between sample comparisons andthe quantification of emissions by any model that uses the experimental data. If analyseswere bad, the experimental data are useless for describing a functional relationshipbetween: naphthalene in sediments, naphthalene in water, and naphthalene in air.
The Data Gaps Report proposes that there is a sediment concentration of naphthaleneabove which no additional naphthalene will be emitted from a slurry. A consequence ofthis proposal is that the modeled relationship of air emissions to sediment naphthalenehas a very rapid rise at very low naphthalene concentrations. In addition, Data GapsReport proposes that this "breakpoint" is the same under different experimentalconditions. This analysis is not substantiated by experimental data, and is likelyincorrect. Further, the model uses one experimental point to determine two lines (seeFigure #1); therefore the experimental data neither support or refute the model, as onepoint is "consistent" with any model (see Figure #2). While solubility may have aneffect on naphthalene emissions, this effect is very difficult to approximate in a dynamicsystem. Changing experimental conditions will change any solubility limit. Therefore,while the output of the model is very sensitive to a single parameter (bulk chemicalsolubility), the model contains no adjustment to account for: dispersion of naphthalene inthe water due to mixing of the slurry, or; additional dilution by added water (i.e. themodel uses a single sediment concentration, solubility limit, regardless of the amount ofsuspended solids in the slurry).
Furthermore, the Data Gaps Report errs by including a correction factor of 0.052 whencalculating a theoretical, solubility-limited "breakpoint." In addition, it is likely thatthere was NAPL present in some experiments. This is suggested by data that shows highconcentrations of naphthalene in sediments, yet water concentration of naphthalene inthose experiments well below the solubility of naphthalene.
The general conclusion of MDH analysis of the air emissions data and model is that thepublic can be protected from exposure to chemicals at levels of concern during dredging. However, the effort necessary to achieve protection is not well understood.
Review of the air emissions model in the Data Gaps Report suggests that:
- Even a good model of air emissions may err by a factor of 10. Given theuncertainties of air emission models; real-time monitoring of PAHs in ambient air, andcontingency plans for reacting to exceedances of acceptable concentrations, should be inplace during dredging.
- The Data Gaps Report proposes the use of experimental emission rates that are notwell supported by data.
- The Data Gaps Report emission model is not accurate, does not reflect environmentalprocesses. Furthermore, the model appears to overestimate emissions, especially at lowsolids concentrations, and provides results that do not agree with experimental data orthat are irrelevant.
- Combining uncertain experimental emission rates with an incorrect model thatappears to overestimate emission rates yields unusable inputs into a dispersion model.
- Emissions during dredging may approach levels of concern, but the majority of thoseemissions will come from a CDF that can be covered. Enclosed processing of dredgedspoils, offsite transport of slurries and utilization of geotextile sediment containmenttubes may also significantly decrease PAH emissions.
- The most significant data gap, that could impact emissions calculations, may be thelack of data on the magnitude of sediment contamination and the extent of NAPL in thebays.
- Mass balance should be performed when conducting experiments with volatilechemicals in different phases or media.
- 2-methyl naphthalene is emitted from sediment slurries at similar rates asnaphthalene. The toxicity of 2-methyl naphthalene and other methylated naphthalenesmay be similar to naphthalene. Therefore, methylated naphthalene concentrations insediment should be added to naphthalene concentrations when determining air emissionsfrom sediments.
- Temperature could have a significant effect on the naphthalene emissions duringdredging. Therefore, cells containing the highest concentrations of naphthalene insediments should be dredged in the early spring or late fall.
The Ice Report in the Data Gaps Report was essentially unchanged from the Ice Reportpreviously submitted by the PRPs. It did contain new field observations from last winter. Analyses of ice intrusion into sediment, the possible formation of ice feet or lenses, andthe disruption of the sediment matrix, which have been identified by MDH as processesthat may impact the longterm effectiveness of a cap, are not included in the Data GapsReport.
MDH has also begun reviewing selected areas of the Groundwater Report and expects tohave full comments shortly. Concerns are generally related to the groundwater-aidedmovement of contaminants through a cap or from a CDF, and the potential forgroundwater to contribute to the erosion and ultimate failure of a cap or a CDF.
Discharges of contaminants from the sediments during placement of a cap have not beenaddressed. Furthermore, a review of fallback or the redeposition of residualcontamination on sediments after dredging needs to be conducted. This would help usdevelop a common understanding of these problems.
Mercury issues related to this site are not discussed in the Data Gaps Report. MDH ismainly concerned about the increase in methylation that can occur in wetland areas thatwould be created by capping. This may be offset because mercury movement from areasunder a cap will be below those currently experienced, but the mobility of mercury is notwell understood and therefore cannot be accurately assessed.
Sediment samples taken for the Data Gaps Report were not analyzed for all chemicalsthat have been identified as potential chemicals of concern by MDH. These include anextended list of carcinogenic PAHs and chlorinated organics.
- Dredging at other PAH-contaminated sites should be observed, or a pilot studyshould be conducted on-site to gather better air emission data.
- If the air emission model is refined:
- recognize that even a good model may only be accurate to within a factor of 10;
- review the experimental data from the air emissions experiments to determine ifthere are usable data from which to calculate emissions of naphthalene;
- gather additional sediment concentration data. These data may be 2-ringhomologue data, accompanied by an estimate of chemical ratios, or more typicalanalytical data;
- determine the extent of NAPL at the site and use this information in the model;
- research and discuss direct volatilization from NAPL to air; and,
- do not include a "breakpoint" that attempts to describe a solubility limitation ofemissions.
- A better understanding of the extent of residue contamination following dredging isneeded.
- Isolation of a CDF from impacts of groundwater and surfacewater needs to beaddressed in a dredging feasibility study.
- Plans for air monitoring should be developed with community input early in theremedial planning phase.
- Dissolved and ambient air concentrations of naphthalene should be determinedunder present conditions, as well as during dredging and capping of contaminatedsediments.
- Contaminant releases that accompany application of a cap need to be itemized andquantified.
- Longterm effectiveness of a cap (and a CDF) needs to be further addressed in theData Gaps Report
- Ice effects on the sediment and cap matrix are potentially serious and need studyand clarification.
- Groundwater effects on contaminated sediments and a cap in areas where itbreaks through, need study and clarification.
- Plant root penetration and potential woody growth on a cap needs to becharacterized.
- The effectiveness of longterm maintenance, monitoring and repair strategies for a capand a CDF need to be addressed, preferably prior to a Feasibility Study.
- Sediment PAH analyses should include analysis of the extended list of cPAHs.
- Chlorinated organic concentrations in sediments should be measured, especially ifsediments are left in place and capped.
- MDH will continue to be active in the community workgroup for the SLRIDTsite.
- MDH will continue to monitor progress at this site and be available forconsultation.
- MDH will assist with review of further documents, as requested.
Carl Herbrandson, Ph. D.
Site Assessment and Consultation Unit
Environmental Surveillance and Consultation Section
Minnesota Department of Health
Agency for Toxic Substances and Disease Registry (1995). Toxicological Profile forNaphthalene, 1-Methylnaphthalene and 2-Methylnaphthalene. Science International Inc.,U.S. Department of Health and Human Services, Public Health Service, Atlanta, GA.August 1995.
Benoit, J.M., C.C. Gilmour, R.P. Mason and A. Heyes (1999). Sulfide controls onmercury speciation and bioavailability to methylating bacteria in sediment pore waters.Environmental Science & Technology 33(6): 951-957.
California Office of Environmental Health Hazard Assessment (2002). Cancer PotencyValues. Sacramento, CA, 11/2002,http://www.oehha.org/air/hot_spots/pdf/TSDNov2002.pdf.
Environmental Law Institute, (1999). Protecting Public Health at Superfund Sites: Caninstitutional controls meet the challenge?, Washington, D.C. Research Report, 1999, 121pg.
Environmental Protection Agency (2001). Correcting the Henry's Law Constant for SoilTemperature Superfund, Washington, D.C. Factsheet, June, 2001.
Gilmour, C.C., E.A. Henry and R. Mitchell (1992). Sulfate stimulation of mercurymethylation in freshwater sediments. Environ Sci Technol 26(11): 2281-2287.
Gustafsson, O., F. Haghseta, C. Chan, J. Macfarlane and P.M. Gschwend (1997).Quantification of the dilute sedimentary soot phase: Implications for PAH speciation andbioavailability. Environmental Science & Technology 31(1): 203-209.
International Agency for Research on Cancer (2002). Monographs on the Evaluation ofCarcinogenic Risks to Humans: Volume 82: Some Traditional Herbal Medicines, SomeMycotoxins, Naphthalene and Styrene, Lyon, France. February, 2002, 367 pg.
Krabbenhoft, D.P., C.C. Gilmour and J.M. Benoit (1998). Methyl mercury dynamics inlittoral sediments of a temperate seepage lake. Canadian Journal of Fisheries and AquaticSciences 55(4): 835-44.
Maruya, K.A., R.W. Risebrough and A.J. Horne (1996). Partitioning of polynucleararomatic hydrocarbons between sediments from San Francisco Bay and their porewaters.Environmental Science & Technology 30(10): 2942-2947.
Matilainen, T. and M. Verta (1995). Mercury methylation and demethylation in aerobicsurface waters. Canadian Journal of Fisheries & Aquatic Sciences 52(8): 1597-1608.
Minnesota Department of Health (1999). Testimony to the MPCA Citizens' Board. C.Herbrandson, Environmental Health Division, St. Paul. October 1999.
Minnesota Department of Health (2001a). Methods for Estimating Health Risks fromPoly Aromatic Hydrocarbons (PAHs). C. Stroebel, St. Paul, MN. Memo To: H. Goeden,Minnesota Pollution Control Agency. February 22, 2001.
Minnesota Department of Health (2001b). St. Louis River/Interlake/Duluth Tar NationalPriority List (Superfund) Site: Sediment Operable Unit. C. Herbrandson, Duluth, MN.Health Consultation ATSDR, Atlanta, GA: April 2001.
Minnesota Department of Health (2003). Naphthalene Health-based Value. H. Carpenter,Health Risk Assessment Unit, St. Paul, MN. Memo To: S. Burman, Minnesota PollutionControl Agency. February 19,2003.
Minnesota Department of Health and Minnesota Pollution Control Agency, (2003).SLRIDT Site Visit. C. Herbrandson and S. Hennes, MDH and MPCA, St. Paul, MN. To:J. Mosel and M. Bares, MPCA. 2/7/2003.
Minnesota Department of Natural Resources (2001). Surface water flow at the InterlakeSite. J. Lindgren, MDNR, Duluth, MN. Telephone conversation with: C. Herbrandson,Minnesota Department of Health, March 1, 2001.
National Research Council (1997). Contaminated Sediments in Ports and Waterways.Washington, D.C., National Academy Press.
National Oceanic and Atmospheric Administration (1996). Contaminants in aquatichabitats at hazardous waste sites: Mercury, Department of Commerce, Seattle, WA.Technical Memorandum NOS ORCA 100, December 1996.
Paquette, K.E. and G.R. Helz (1997). Inorganic Speciation of Mercury in SulfidicWaters: Importance of Zero-Valent Sulfur. Environ. Sci. Technol. 31(7): 2148 -2153.
Peer Review Group (2002). Meeting of Peer Review Group, SLRIDT Site, St. Paul, MN.December 10, 2002.
Regnell, O., G. Ewald and E. Lord (1998). Factors controlling temporal variation inmethyl mercury levels in sediment and water in a seasonally stratified lake. Limnologyand Oceanography 42(8): 1784-1795.
REMAP, E. (1995). Regional Environmental Monitoring and Assessment Program. U.S.Environmental Protection Association.
Rudd, J.W. (1995). Sources of methyl mercury to freshwater ecosystems: A review.Water Air And Soil Pollution 80(1-4): 697-713.
Saint Louis, V.L., J.W. Rudd, C.A. Kelly, K.G. Beaty, R.J. Flett and N.T. Roulet (1996).Production and loss of methylmercury and loss of total mercury from boreal forestcatchments containing different types of wetlands. Environmental Science & Technology30(9): 2719-2729.
Schwarzenbach, R.P., P.M. Gschwend and D.M. Imboden (1993). EnvironmentalOrganic Chemistry. New York, John Wiley & Sons, Inc.
Sellers, P., C.A. Kelly, J.W. Rudd and A.R. Machutchon (1996). Photodegradation ofmethylmercury in lakes. Nature 380(6576): 694-697.
Service Engineering Group (2002). Data Gap Report: St. Louis River/Interlake/DuluthTar Site, Duluth, Minnesota, St. Paul, MN. November 28, 2002.
Stortz, K.R. and M. Sydor (1980). Transports in the Duluth-Superior Harbor. Journal ofGreat Lakes Research 6(3): 223-231.
U.S. Geological Survey (2001). A National Pilot Study of Mercury Contamination ofAquatic Ecosystems Along Multiple Gradients: Bioaccumulation in Fish. W. G.Brumbaugh, D. P. Krabbenhoft, D. R. Helsel, J. G. Wiener and K. R. Echols, U.S.Department of Interior, Washington, D.C. Biological Science ReportUSGS/BRD/BSR-2001-0009, September 2001, 32 pg.
This St. Louis/Interlake/Duluth Tar Site Health Consultation was prepared by theMinnesota Department of Health under a cooperative agreement with the Agency forToxic Substances and Disease Registry (ATSDR). It is in accordance with approvedmethodology and procedures existing at the time the health consultation was begun.
Alan W. Yarbrough
Technical Project Officer, SPS, SSAB, DHAC
The Division of Health Assessment and Consultation, ATSDR, has reviewed this publichealth consultation and concurs with the findings.
Chief, State Program Section, SSAB, DHAC, ATSDR
1 DGR Section A2-3.3.3
2 Calculated with foc = 23%; Koc = 933