LCP Chemicals Site Baseline Risk Assessment
January 2002

Overview

 Any site treatment, including doing nothing, must meet certain requirements under Superfund law. Legally, a site must pose no risk to people or animals that use the site. Under EPA rules, this requirement is met, in part, through evaluation by a risk assessment. At the LCP site, there have been a series of risk assessments examining the effects of site chemicals on plants, bugs, animals, birds, and people. These assessments have modeled risk for the groundwater, the upland soil, and the marsh sediment.

The purpose of the risk assessment is two-fold: both to get a “snapshot” of the present day risk, and to guide future cleanup. This site is expected to remain as an industrial site that is off-limits to the public for recreation or permanent homes. Therefore, the risk assessment focused on a “worker” model, rather than a “residential” model; and looked closely at the long-term environmental impact on animals that feed in the marsh.

Overall, this site is still a toxic waste hazard, both to future workers and to animals. Although much effort was expended to stabilize the site, more needs to be done to halt contamination in the marsh areas, more needs to be done to halt contamination of groundwater, and more needs to be done before the site is safe for human workers.

  Background

The LCP Chemicals Superfund Site is a 550-acre site consisting of 480-acres of tidal marsh plus “upland” dry areas. The site has been the location of several chemical plants over many decades. Oil refining, paint production, and bleach production are past industries. Each left behind buried debris and waste, and each used the marsh as a dumping zone for effluent. Generally, the site contains toxic levels of Mercury, Lead, cancer-causing “PAH” (polycyclic aromatic hydrocarbons- multi-ringed chemicals from oil), “PCBs” (Poly-Chlorinated Bi-phenyls), and traces of dioxins.

This site was the target of a major cleanup during the mid-1990’s that removed hundreds of tons of toxic waste. Many more tons of wastes were stabilized on-site. The site is no longer the major source of new contamination to the marsh and river; however, pollution migration still occurs from the marsh to the river, from the groundwater to the marsh, and from the upland soils to the groundwater. These sources are the focus of continued monitoring and cleanup efforts under the Remedial Investigation.

The initial cleanup occurred under the EPA’s Removal Action program. The present studies are conducted under the Remedial Investigation/Feasibility Study rules. This process surveys for chemical contamination and performs risk analysis to gain an understanding of effects pollution may cause. The site is divided into three areas for study: the uplands area consists of soils in the mainly dry inshore areas (about 70 acres); the marsh lands, which include the shores adjacent Purvis Creek and the Turtle River, plus wetlands that are inundated twice a day by the tides (about 480 acres); and the groundwater under both the uplands and the marshlands. Each area is complex. There are several strata of groundwater, for example, and the marshlands near shore are different from marshlands near the river.

The Glynn Environmental Coalition received several risk assessment documents, including the Baseline Ecological Risk Summary for the Estuary at the LCP Chemical Site  (September 2001), the Human Health Baseline Risk Assessment Groundwater (July 2000) and the Human Health Baseline Risk Assessment Marsh Sediment and Upland Soil (July 1999). In conjunction with earlier documents, these new documents gave the community the first comprehensive glimpse of risk factors for the LCP site.

Several different authors, in at least two different groups, using different approaches, constructed the studies. As a result, it was not always straightforward to interpret the information. To gain an overview of the Risk Evaluation we extracted portions of the various assessments into spreadsheets (attached as Tables 1a and 1b). To further simplify the Human Health Risk Assessment we primarily looked at a “child-resident” model, since children are most vulnerable to pollutants; and we examined a long-term “industrial worker” model, as the most likely future user of the site. On the environmental side, we looked for data representing the major trophic (feeding) levels at the highest concentrations of chemicals. We sorted chemicals based on danger and occurrence. Our main goals in this process were to look for any gaps in the analysis, and to spot any major trends.

Risks are divided into three categories. Hazard Indexes (HI, the sum of the Hazard Quotients, HQ) measure the relative toxic effects of certain chemicals. A Hazard Index greater than “1” indicates a threat. Generally, the higher the number the greater the threat, but  levels greater than 1 are “actionable” signifying remediation is required under the National Contingency Plan (the “NCP”). A probability scale that looks at the increase in chance of new cancers measures cancer risk. Both the type and amount of chemical is important in cancer probability models. If a chemical concentration has a one-in-one-million chance of causing a new case of cancer it is considered to be of no risk. As the concentration of the chemical increases so does the risk for new cases of cancer. Generally, EPA considers as “actionable” any cancer risk greater than “one-in-ten-thousand” probability. A finding of  “two-in-ten-thousand” excess cancer risk would mean that action in the form of a cleanup is required under the NCP rules. Risk considerations also include a “miscellaneous” category. For instance, Lead in children under the age of six years old does not use either HI or cancer risk probability. Instead, a computer model is used to estimate the potential blood Lead level, since this is a better predictor of injury. Likewise, for environmental risk there are reproductive-effects tests and community-structure tests that fall outside of the standard Hazard Indexes.

For all risk analysis, the chemicals, chemical concentrations, health-effects, and completed pathways must be documented. The scientific literature is relied upon for these assessments. The kinds of chemicals, the kinds of effects, which target organs (liver, kidney, etc.) are affected, and how chemicals get from the environment (breathing a gas, breathing an aerosol, drinking liquid, eating, or through the skin) to the person or animal all must be found and measured. In many cases, there is little or no prior information, other than laboratory tests, to use in making a decision.

Findings

Upland Soils:

The uplands area was industrial property for most of the last century and is expected to remain industrial in the future. For risk analyses, site areas are designated “residential,” “commercial,” or “industrial” use. A residential use means that people of all ages and sexes would be exposed 24 hours a day for 50 weeks per year. An industrial usage provides only for adult workers for a few hours per week. However, the LCP site sits between residential areas and the marsh, so trespass is likely. Both a long-term industrial worker and several trespasser models were the used in assessing risk for the uplands. Additionally, the risk assessment examined a future residential model, although the authors argue residential development is unlikely. This was needed since Glynn County has grown appreciably in the last two decades and waterfront property is increasingly valuable as recreational and living space. Under the residential scenario, we focused much of our attention to child-resident models, since children are the most sensitive human group. However, the most likely near-term users at LCP are trespassers and industrial workers.

The risk assessment models for uplands did not extensively model the subsurface (the HBRA indicates in Table 6-4 very short exposure durations, with minimal exposure). Subsurface soils remain contaminated at this site, and the risks are high for contact with soil. Much of the on-site pollution remains buried. Any construction, including digging, trenching, or grading, would likely expose workers to toxins. These effects should be modeled more extensively. Surficial groundwater moves from the uplands to the marsh. The contaminated subsurface soils need to be modeled for continued contribution to groundwater.

Children under the age of six years are especially vulnerable to Lead exposure. The exposure, both acute and chronic, does not cause immediate or obvious injury, but shows up in later years as nerve impairment affecting intellectual and emotional development. Although the study’s authors argue that child Lead exposure is unlikely, it was modeled at the request of EPA and Georgia EPD. The model uses conventional exposure assumptions and found no danger (to the 98% confidence level) to child future residents. One of the main goals of the Removal Action was to reduce heavy metals in surface soils, and the Lead levels found now are consistent with meeting those goals.

A future adult resident was found to have a Hazard Index of 3 for surface soils, which is “actionable” under the National Contingency Plan. “Actionable” under NCP rules means that the site must be treated in some way to reduce the threat of harm. Any HI above 1 is considered actionable.

A future industrial worker was found to have a Hazard Index of 1 for surface soils. Both residential and industrial models show cancer risks less than the actionable levels. Looking over the calculations it appears the study’s authors maneuvered the models somewhat to get data showing the future industrial workers with HI’s at exactly 1. The model uses factors such as hours per day of exposure, days per year, skin covered or otherwise protected, averaging of chemical concentrations instead of using the highest values, and other techniques to arrive at an estimate of risk. The factors used in these studies were not the most conservative values. If the authors had used more conservative values, industrial workers would have had Hazard Indexes above 1. It seems probable that future workers could violate the stringent conditions needed for safety. Overtime employment, failure to wear complete skin coverings, or spending too much time in areas that have toxins above the “average” levels would result in higher HI’s. It is fair to say that the site requires surface soil treatment for the safety of future residents, and that conditions are “borderline” or marginal for future industrial workers exposed to site soils.

Groundwater:

Groundwater contamination at the LCP site is very complex. A contaminated upper or “surficial” aquifer drains from the upland portion down gradient towards the River. The contamination comes from remaining buried chemicals including old refinery chemicals (PAH) and heavy metals from LCP operations. At several points, this upper aquifer seeps into the marsh. During low tides after heavy rains the seepage is likely more severe. In places, there is only a thin layer of gravel-and-sediment separating the contaminated surficial aquifer and less contaminated marsh surface water. This thin layer can easily be broached; a heavy walker/wader could sink through the thin separation layer becoming exposed to surficial aquifer contaminants. There is a lower aquifer called the “rock” aquifer underlying the site, and an even deeper aquifer used for drinking water called the Floridan aquifer. These lower aquifers presently appear not contaminated; but should be considered “threatened” since LCP groundwater contamination may still be spreading.

In addition to the contamination in the surficial aquifer, there is a “caustic brine pool”  (or “CBP”) of dense alkaline chemicals and heavy metals. The CBP begins on the upland portion around the location of the former bleach facility and flows underground to a form a large pool beneath the marsh. Skin contact with the CBP would be similar to the injury caused by some household “lye” type drain cleaners. The CBP essentially flows through and under the upper aquifer. The risk study’s authors argue that little mixing of the CBP and aquifer occurs due to differences in liquid density (the same principal that allows drain cleaners to sink through standing water). Under the ground and the marsh, the CBP has “eaten” away the rock by dissolving the silica rock of the limestone, to form silica gel. The authors argue that the gel, which is less “soluble” (cannot easily dissolve) than the lime rock, has formed a synthetic barrier thus decreasing the rate of further rock break down. While the plant was operating portions of the old cell building began to collapse from the action of the CBP dissolving the ground beneath the site. Obviously, as long as the CBP is present, there is a limit on the types of structures that could be developed over the CBP. Hazard Indexes for material such as the caustic brine pool are not really required. As already mentioned, drinking and skin contact with the CBP results in acute injury. Chemical burns could actually occur from exposure, with skin damage, loss of limbs, or death possible.

The upper surficial aquifer is highly contaminated with industrial chemicals, including heavy metals and cancer-causing PAH. The authors argue that no resident would likely drink surficial groundwater in quantity, or use the water for bathing (PAH can enter through the breathing exposure pathway from showering in contaminated water). The Hazard Index cited for a child resident was 54, which is very high above the action level of 1, the “trigger” for cleanup. The child cancer index was two-in-ten-thousand probability, which is also above the trigger of one-in-ten-thousand probability. For a long-term industrial worker the Hazard Index was exactly 1, and cancer-causing models were acceptable. It did not appear that the study examined such things as irrigation of garden vegetables, which is a common use of water from surficial aquifers.

The “wader” model used in this study was very poor in our opinion. We think wading trespassers are likely, and might occur at this time since the site is situated between residential areas and the marsh, and is poorly enclosed. Exposure of waders to the known contaminated groundwater seeps can occur and likely has and will. In this study, the wader was considered wading one hour per day, for one day per year, for only ten years. That adds up to only ten hours in ten years, a number we believe seriously underestimates wading trespassers at this site.

Marshlands:

The LCP marshes are complex areas of near-shore tidal flats, a network of tidal creeks and deeper drainages toward the Turtle River. Some areas are inundated during seasonal high tides but are otherwise dry soils. Some structures in the marsh are man-made. Dredging may have changed the natural flow patterns in some areas. Other areas are covered with water year-round. Some areas are more saline (salty), other areas receive surface runoff from the uplands, and still others receive seepage from contaminated upper aquifer water. Different areas of the marsh received different types of pollution, and portions of the banks near-shore were filled or bermed with waste, including heavy metals.

To conduct the environmental risk analysis a roughly rectangular grid was used in the marsh, with sampling stations at intervals throughout the grid. Obviously, with such a complex marsh system, the individual points on the grid system can vary widely in the types of pollutants and the nature of marsh plant and animal communities. Describing the total effect of pollutants on the marsh requires very accurate data over a long period or time, or considerable interpretation. This risk analysis relies on extensive interpretation, rather than extensive data.

There were several “lines of evidence” used in modeling the marsh. “Line of evidence” as applied to this site is a “qualitative” term (not based entirely on numbers). The “evidence” ranged from true quantitative numerical data to obviously raw opinion. In some cases, animal species known to inhabit the marsh were used in studies that calculated Hazard Indexes and cancer-effect probability models. The efficacy (usefulness of the outcome) of these models is dependent on accurate knowledge of the effects of known chemicals on the lifecycle of the animal. In most cases, such scientific information is lacking, so values from the scientific literature on related species were used. As one example, the effect of Mercury on clapper rails (marsh birds) was not known, but Mercury studies had been done on different birds, such as grackles and eagles. Although it is not known if the data is comparable between the species, the literature data was used for this modeling study. Two types of hazard models were used, both the direct effects of exposure to the chemicals and the effects of eating contaminated prey species. To do the “uptake” effects (amount of chemical in the animal from the environment) the authors could use the chemical concentrations in the water and make computer predictions on uptake, or they could sacrifice some of the animals and perform a target-tissue chemical analysis. Both types of studies were done, however, not consistently across species lines. For instance, marsh turtles were only computer modeled in this new study, but were measured directly in an earlier risk estimation. Other sets of experiments measured the community structure as a whole. Observations were taken of the numbers and types of species found in an area, and compared to control sties that do not receive LCP pollution. The control sites were at areas such as nearby Troup Creek. Major differences between LCP marsh and the reference marsh are considered an indicator of pollution damage (these types of changes are well known). Some studies collected contaminated marsh sediment and water, and exposed laboratory raised fish or bugs to see if the LCP mud or water was toxic. In other experiments, animals were sacrificed and experts performed a pathological test for the effects of toxins on livers, kidneys, or other target organs.

Obviously, there is a great deal of uncertainty present in interpreting any results of these exposure models. Animals may not be consistently exposed, the exposed animals may have moved on the day collections were done, or the sublethal effects for LCP marsh species may be very different than for species that were described in the literature.

Despite the uncertainty, there is a “consensus” of evidence showing the marsh is impacted by toxins and will continue impacted, unless intervention in the form of a cleanup occurs. The rationale for this is the evidence of reduced survivability to amphipods (small shrimp-like animal mostly known for its chemical sensitivity) when exposed to marsh sediment and the Hazard Indexes of a number of animals ranging from Clapper Rail birds (HI of 12) to Raccoons (HI of 47) to Finfish (HI of 6.4). With evidence of toxicity ranging from the sediment all the way up the food chain, there can be little doubt that the marsh is a hazardous place for local wildlife.

There are several problems with some of the conclusions reached by the study’s authors. With respect to marsh turtles, a 1995 preliminary study found turtles with wasting disease and reproductive problems, although the lesions were not directly comparable with other reports of effects of Mercury and PCB on reptiles. The authors of the recent study point out the earlier study had a small sample size (very few turtles could be found) and the authors used a pure computer model, based on no turtles, to show damage could not occur. This is poor science.

Some animals expected to use this marsh are not found. Mink, a common predator, was noted as absent from the LCP marshes. The risk assessors ignore this line of evidence in their summation, but the absence of top predators is highly predictive of, at the very least, territory avoidance, if not outright damage to a species. Marsh turtles, a species of very limited range, should derive the great majority of their diet from LCP local sources; and should be constantly exposed to sediment. Consequently, the limited sample size for marsh turtles is itself a point for suspecting ecological damage, not a rationale for discontinuing direct studies on turtles.

In the absence of top trophic species, other prey species should be found in higher abundance, another point ignored by the study’s authors. Consequently, much of the marsh eco-community models are probably suspect. One entire section of the report detailing complex “back-calculation” models was of no scientific merit, and should not be considered in filling statutory requirements under the NCP.

People fishing in the marsh for fish and shellfish are exposed to marsh waters and sediment directly, as well as through eating contaminated food. Although the study’s authors argue the local ban on fishing in the marsh should prevent exposure, bans are not effective institutional controls, since they can only be effectively monitored in public areas, such as on bridges. There is some controversy over what constitutes a “subsistence” fisher. Subsistence fishing is different than sport fishing. Many sport fishers engage in game catch-and-release, so no exposure occurs through ingesting contaminated meat. Subsistence fishers do supplement their diets routinely by catching and consuming fin and shellfish. Some studies indicate subsistence fishing is a function of socioeconomic factors, with low-income fishers obtaining a far greater portion of their diet through fishing. Further, lower socioeconomic strata fishers are more likely to provide fin and shellfish to family members, including children. Consequently, the 30-grams of seafood per day used in calculating threat in this study may be low for the actual consumption in the LCP area. This study can underestimate risk from use of the marsh by the public.

  Conclusions

Throughout the marsh risk analyses, there was a strong tendency to oversimplify the studies and interpretations. This was regrettable, since without a solid baseline assessment it becomes much more difficult to design countermeasures to halt contaminant spread, or to accelerate natural cleanup. The contaminated marsh will likely remain polluted indefinitely.

A trespasser impinges on all three modeling areas since they must cross the uplands to get to the marsh, the likeliest reason for trespassing. Once in the marsh, they are exposed to contaminated sediments, contaminated surface water, contaminated groundwater from one of the seeps, and contaminated fin and shellfish. Trespassers can also enter the marsh by boat. One trespasser can impact others by providing contaminated seafood. These risk assessments greatly underrate the risk to a serial trespasser. These unrealistic trespasser models raise questions regarding the realism of other conclusions the authors have reached.

The study’s authors trivialized the dangers from upland subsurface soils. There are significant levels of toxins left in the subsurface and the model should be more realistic for site excavation. Likewise, while we agree with the finding that upland site trespassers likely have hazard indexes below 1, and low probability of cancer from site soils, we find the models used less than conservative factors for building the estimates.

The site is hazardous to wildlife. Nearly every aspect of the site shows potential toxicity, from the mud all the way up to the top predators. Without intervention by additional cleanup, the site will continue to place marsh species at risk.

Finfish and shellfish from the LCP marsh can be harmful to people. Controversy over what amount of food is considered a normal “dose” for subsistence fishing only serves to hide the fact that catching and eating seafood from this area is inherently “risky.”

The surficial aquifer flows through contaminated upland soils to the marsh, dissolving and carrying pollutants along the way. In the marsh, there is an interface where toxins can seep into the marsh or mix with subsurface waters for eventual dilution in the creek and river drainage. A direct route of exposure exists between the upland subsurface contamination and the waters of the marsh, via the surficial aquifer. This route needs to be modeled since there is a lack of data on the long-term impact. Using groundwater to flush soil toxins into the marsh for dilution is a poor, perhaps ineffectual, way of cleaning upland subsurface soils. 

Much money has been, and is being, spent to “prove” the caustic brine pool can be safely left in place. It seems very doubtful that safety can be achieved merely by monitoring. The “bottom line” is that the money may be better spent on removal, since the time span for monitoring is likely measured in eons, not decades.

 ·         It is obvious the site could not be used for any residential or recreational purposes.

·         The site is only marginally safe for many commercial purposes as well.

·         No large structures could be built over or near the caustic brine pool.

·         The marsh is harmful to marsh life.

·         A route for contaminant spread and recontamination exists between the uplands and marsh via the contaminated surficial aquifer and caustic brine pool.

·         No use could be made of groundwater at this site, for the next few hundreds of years, unless action is taken to clean up the aquifers.

·         Essentially, the 550 acres of the LCP site cannot be used for anything at this time; it is even unsafe as wetlands.

 Written by R. Kevin Pegg, Ph.D.; edited by Dr. Mary S. Saunders. Copies of the newsletter are available from the GEC, at the Glynn County library, or at www.NucleicAssays.com/eco  on the Internet.

"This project has been funded wholly or partly by the U.S. Environmental Protection Agency under Assistance Agreement Number 1-994850-01-0 to The Glynn Environmental Coalition, Inc. The contents of this document do not necessarily reflect the views and policies of the U.S. Environmental Protection agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use."