Ecological Risk Assessment for Contaminated Sites - Chapter 10 pot

10 Post-Remedial Assessments The man who is certain he is right is almost sure to be wrong; and he has the additional misfortune of inevitably remaining so. —Michael Faraday, quoted in Johnson (1991) The assessment process does not necessarily end with the completion of a remedial action. If the contaminants are not all removed or destroyed or if the environment is not restored to an acceptable state by the remedial action, it may be necessary to perform additional sampling, analysis, and assessment. These monitoring and assess- ment activities have two purposes: (1) to ascertain whether additional remedial actions are needed and (2) to estimate the residual contaminant-induced injuries to the environment to determine what restoration activities are needed. Although these purposes are closely related, they are procedurally distinct under U.S. regulations and therefore are discussed separately below. A potential collateral purpose of these post-remedial assessments is to provide a basis for testing the predictions made in the prior assessments. Like any other activity based on scientific principles, the predictions of risk assessments must be tested to prevent the complacency of which Faraday warned. 10.1 REMEDIAL EFFICACY Remedial actions seldom remove or destroy all contaminants, and some simply control uses of the site. It is advisable to monitor incompletely cleaned environments to establish that the remedial actions were adequate and that unacceptable risks do not persist. The EPA notes that the efficacy of ecological risk reduction is sometimes easier to observe than that of human health risk reduction. The agency recommends that the effectiveness of remedial actions be assessed, following their implementation (Sprenger and Charters, 1997). In many cases, biological monitoring will provide a better indication of risk reduction than simple contaminant monitoring (Fox, 1993). In addition, the extent and rate of ecological recovery following the selection of the no-action alternative should be determined (Sprenger and Charters, 1997). In the United States under CERCLA, such monitoring must be carried out for 5 years, at which time an assessment must be performed. Requirements for post-remedial monitoring and assessment should be included in the record of decision along with the requirements for remediation. The extent and intensity of this activity should be inversely related to the extent and intensity of the sampling, analysis, and assessment performed prior to remediation. For example, if it was clearly established that the only significant risks prior to remedi- ation were due to consumption of contaminated fish by piscivorous wildlife, © 2000 by CRC Press LLC monitoring could be limited to the analysis of the contaminants of concern in fish from the areas where risks had been significant. The assessment would document the trends in contaminant levels and associated risks over time and would identify when risks dropped below prescribed threshold levels. In contrast, if an action is taken without an ecological risk assessment or with an inadequate assessment, and if all of the contaminant has not been removed or destroyed, the site should be treated like a newly discovered contaminated site. The process of problem formu- lation, analysis of exposure and effects, risk characterization, and risk management should be carried out as described in prior sections. Kondolf and Micheli (1995) provide recommendations for the post-project evaluation of stream restoration projects. Some of this guidance, including the relationship of initial restoration objectives and specific evaluation criteria, may be useful for remedial efficacy assessment. The importance of considering new toxic exposures and effects on new assess- ment endpoints from remedial measures is illustrated in the tests of the efficacy of remediation in Table 10.1. For many endpoints and actions, soil was more toxic following remediation than preceding it. The adverse effects resulted from (1) increased bioavailability following washing with solvents or (2) alteration of the wetting properties of soil. In most of these studies, the technologies are still in the development stage. The toxicity of environmental media following the use of reme- dial technologies has been studied only for the past few years. Nonetheless, these cases illustrate the importance of monitoring population parameters and ecological processes following remediation. In addition to the plant and earthworm endpoints described in Table 10.1, toxicity of remedial actions to aquatic and sediment organ- isms and terrestrial wildlife should be tested in the laboratory, observed in field surveys, or estimated using chemical concentration data. 10.2 RESIDUAL RISKS Even when remedial actions are rapid and thorough, some injury to the environment is likely to occur when sites have been contaminated by spills or improper waste disposal. Under CERCLA, the responsible party must pay “damages for injury to, destruction of, or loss of natural resources” (CERCLA, §107(a)(4)(c)). The liability is to natural resource trustees, which include agencies of the United States, individual states, and native American tribes. The trustees are required to recover monetary damages for residual injuries, which are injuries that are not eliminated by the reme- diation. The assessment process for determining those damages is termed Natural Resource Damage Assessment (NRDA). The primary federal trustees are the U.S. Fish and Wildlife Service (FWS) and the National Oceanic and Atmospheric Administration (NOAA), but any landholding agency is also a trustee for the resources on their lands. The compensation may be for either the costs of restoring the injured natural resource services or the diminution of value of the natural resource services, plus the costs of performing the assessment. NRDAs follow a distinct set of regulations, procedures, and terminology (Box 10.1). The following discussion describes the procedure defined in Department of Interior’s (DOI) regulations (40 CFR 11). This procedure need not be followed, but NRDA results produced by trustees who follow the procedure have a © 2000 by CRC Press LLC TABLE 10.1 Studies of the Efficacy of Remedial Technologies in Protecting Ecological Endpoints Problem Remedial Technology Test for Efficacy Principal Conclusion Ref Soil contaminated with PCBs, other organics, and heavy metals from electrical transformers Washing with isopropanol Toxicity to earthworms ( Lumbricus terrestris and Eisenia fetida andrei ) and common onion ( Allium cepa ) Extraction removed 99% of PCBs, with 2 mg/kg remaining; reproductive toxicity to earthworms was not altered; phytotoxicity increased following washing, but when isopropanol was washed away with water, phytotoxicity was reduced Meier et al., 1997 Soil contaminated with lead to 700–800 mg/kg (also cadmium, zinc, chromium, nickel, silver) Soil washing/soil leaching process Toxicity to earthworms ( L. terrestris and E. fetida ), lettuce ( Latuca sativa ), and local grasses blue grama and sideoat grama Following washing, earthworm mortality increased in a mixture of contaminated soil and artificial soil; root elongation of the grasses was reduced; germination and root elongation were reduced in lettuce; the adverse effect on plants was eliminated after soils were rinsed Chang et al., 1997 Soil contaminated with PAHs Bioremediation using white-rot fungus, Phanerochaete chrysosporium Toxicity to lettuce, millet, oat, and Tradescantia — micronucleus test Plant tests showed that soil was significantly detoxified following treatment; the micronucleus test showed that soil was less genotoxic than untreated soil Baud-Grasset et al., 1993 Explosives- contaminated sediment from a lagoon at an army depot in Oregon Sediment composted with a 70% organic amendment comprising cow and chicken manure, sawdust, alfalfa, and potato waste Toxicity to lettuce ( L. sativa ), radish ( Raphanus sativus ), and soybean ( Glycine max ), earthworms ( E. foetida ), and isopods ( Armedillidium vulgare ) in mesocosms; toxicity to cabbage ( Brassica oleracea ), clover ( Trifolium repens ), and Arabidopsis thaliana . Germination of lettuce, clover, and Arabidopsis thaliana was reduced in the contaminated compost, relative to reference compost; soybean leaves were chlorotic, and plant biomass was reduced compared with the reference compost; few adverse effects to soil invertebrates were observed Gunderson et al., 1997 continued © 2000 by CRC Press LLC © 2000 by CRC Press LLC Agricultural topsoil contaminated with oil Bioremediation of soil to total extractable hydrocarbon level of 2% (from 4%) Toxicity to barley ( Hordeum vulgare ) The barley yield was not improved following bioremediation; shoot dry mass decreased slightly following bioremediation; the lack of improvement was attributed to poor soil water sorption because of the remaining hydrocarbons Li et al., 1997 Soil denuded of vegetation by contaminants from zinc smelter in Belgium Metals immobilized with beringite and compost Toxicity to Phaseolus vulgaris Growth inhibition to seedlings in treated soil was not observed Vangronsveld et al., 1995 Soil contaminated with uranium Treatment with carbonate and bicarbonate or citric acid Toxicity to oats, radish Treatment with sodium carbonate or ammonium carbonate reduced the seed germination fraction for all plants (unless soil was washed with CaCl 2 ); germination of radish and oats in soils treated with citric acid and CaCl 2 were reduced; treated soils were resistant to wetting and had a concrete-like surface Edwards, 1994 TABLE 10.1 (continued) Studies of the Efficacy of Remedial Technologies in Protecting Ecological Endpoints Problem Remedial Technology Test for Efficacy Principal Conclusion Ref © 2000 by CRC Press LLC © 2000 by CRC Press LLC “rebuttable presumption.” That is, the results are legally assumed to be correct unless the responsible party can prove otherwise. For a guide to current NRDA regulatory documents and guidance, visit http://www.doi.gov/oepc/frlist.html. The currently ongo- ing NRDA for the Fox River and Green Bay, WI can be followed at http://www.fws.gov/r3pao/nrda/. Because of the high profile of this site, including effects of PCBs on birds, it should be an influential case. The NRDA procedures have been developed to encompass assessments per- formed under the Clean Water Act (1977 Amendments), the Outer Continental Shelf Act (1978 Amendments), and the Oil Pollution Act of 1990 as well as CERCLA. The provisions of NRDA differ under the other acts in that the injured site would not have been subject to a remedial investigation and feasibility study but may have been subject to prespill planning to predict and minimize injuries (Deis and French, 1998). Ideally, the needs of the NRDA would be anticipated in the planning of a remedial investigation so that the costs of sampling, analysis, and assessment could be minimized. In addition, it would be advisable to consider natural resource dam- ages when planning the remedial action so that the total cost of remediating the contaminated site and the natural resource damages could be minimized. However, this integration seldom occurs in practice, because different agencies are involved. The EPA is required to notify natural resource trustees of CERCLA sites that potentially involve natural resource injuries, and collaboration between the EPA and trustees could and should continue (Office of Emergency and Remedial Response, 1992d). The responsible parties should want to integrate the two assessments, because they bear the cost for both remediation and restoration. However, in the authors’ experience, responsible parties do not want to think about NRDA during the remedial investigation. BOX 10.1 Key Terms in Natural Resource Damage Assessment Natural resources : “Land, fish, wildlife, biota, air, water, groundwater, drinking water supplies and other such resources. . . .” The resources must be public property. Injury : “A measurable adverse change, either long- or short-term, in the chemical or physical quality or the viability of a natural resource resulting either directly or indirectly from exposure to a . . . release of a hazardous substance. . . .” Services : “Physical or biological functions performed by natural resources, including human uses of those functions.” Damages : Money paid to a trustee in compensation for the injuries to natural resources. Restoration: Return of the natural resource services to a baseline condition. Replacement/Acquisition of the Equivalent: Substitution of another resource that provides equivalent levels of services for the injured resource. Source : DOI (1986). © 2000 by CRC Press LLC There are two distinct NRDA procedures. Type A assessments use simulation models to characterize injuries and estimate damages for spills of oil and other hazardous materials in coastal marine waters and in the Great Lakes (French et al., 1994). The trustee’s assessors must paramaterize the appropriate model with char- acteristics of the receiving environment, conditions at the time of the spill, and characteristics of the spilled material. The Type A model then generates estimates of injuries and associated damages such as loss of commercial and recreational fisheries and loss of recreational use of beaches. Regulations for performing a Type A NRDA were published in DOI (1996). Type B assessments estimate injuries on a site-specific basis using methods similar to a remedial investigation and calculate damages using local values. The Type B method, which is far more generally useful, is discussed below. 10.2.1 P REASSESSMENT S CREEN P HASE The preassessment screen phase is equivalent to the scoping and screening steps in a remedial investigation. Its purpose is to determine whether to proceed with a full NRDA. The criteria for proceeding to full NRDA are as follows: • Release of a hazardous substance has occurred • Natural resources for which a federal or state agency or native American tribe may assert trusteeship under CERCLA may have been adversely affected by the release • The quantity or concentration of the released substance is sufficient for the potential injury of those natural resources • Data to pursue an assessment are readily available or are likely to be obtained at a reasonable cost • Response actions, if any, do not or will not sufficiently remedy the injury to natural resources without further action (i.e., residual injury) It is intended that the preassessment screen be carried out rapidly with existing information, and that any sampling follow the screening phase. However, sampling may be performed before completion of the preassessment screen if necessary data or materials would otherwise be lost. Such sampling is likely to be needed for spills or CERCLA emergency actions, but is unlikely to be needed for typical CERCLA NRDAs. Where an RI/FS has been performed, the ecological risk assessments in those documents should provide a more-than-adequate basis for a preassessment screen. Another part of the preassessment process is notification of trustees. In many cases, there are multiple natural resource trustees (e.g., a federal land manager, the U.S. Fish and Wildlife Service for migratory birds and threatened or endangered species, a state agency managing nonmigratory species, and a tribe with treaty rights to fisheries). In such cases, a lead trustee should be designated to conduct the NRDA, while cotrustees are consulted in planning the assessment and other important decisions. The lead trustee appoints an “authorized official” with authority to make decisions on behalf of the agency, including the decision whether to proceed with the NRDA. In a few major NRDAs (e.g., the Exxon Valdez case), agreement on a lead trustee is difficult, so a trustee council manages the assessment. © 2000 by CRC Press LLC 10.2.2 A SSESSMENT P LAN P HASE The assessment plan for an NRDA is equivalent to the analysis plan for a remedial investigation. The discussion of analysis plans in Section 2.7 is applicable here. The DOI procedure specifies that the plan should ensure that the NRDA is performed in a well-organized and systematic manner, and that the methods selected can be conducted at reasonable cost. The emphasis on cost of assessment is based on a concern that trustees not use the NRDA process as a means of supporting staff or acquiring equipment at the expense of the responsible party. The cost of the assess- ment should not be a large fraction of the expected damages. However, in cases with large natural resource damages, such as the Exxon Valdez oil spill, the assessment may be quite extensive, intensive, and expensive. 10.2.3 A SSESSMENT P HASE The assessment phase is equivalent to the analyses of exposure and effects and the risk characterization in a conventional ecological risk assessment. It consists of three steps: injury determination, quantification of effects on services, and damage determination. 10.2.3.1 Injury Determination Phase The injury determination phase must demonstrate the occurrence of adverse effects on a property of a natural resource that constitutes a service and the causal relation- ship of those effects to the release. The identification of effects that constitute a loss of services is equivalent to the selection of assessment endpoints in a remedial investigation (Section 2.5), but is restricted by the need to assess damages. That is, NRDAs do not simply estimate effects on societally or ecologically significant properties. They must estimate effects on properties that constitute services to some component of society that can be associated with monetary values. In some cases, the estimation is straightforward. For example, if the contaminant levels exceed Safe Drinking Water Act standards, the inability to use that water for drinking is an injury that has clearly associated costs. However, the term services has been broadly interpreted to include nonuse values along with the more conventional market and nonmarket use values. Definitions of injuries contained in the NRDA regulations are presented in Box 10.2. Note that this is not a list of all compensable natural resource injuries. However, trustees who select injuries from this list maintain their rebuttable presumption. The process is complicated by the concept of services of one resource to another. For example, water provides a service to the fish community by providing habitat, and the fish provide recreation, food, and nonuse values. To prove that a resource has been injured, the NRDA must confirm the potential exposures and effects identified in the preassessment screen and must demonstrate that the injury is associated with the subject release. Exposure may be confirmed by characterizing the release and modeling the transport and fate processes. More commonly, exposure is determined by measuring contaminant concentrations in biota or abiotic media. Confirmation of effects is most commonly performed by directly measuring reductions in the quality of the resource, but less direct measures, including toxicity tests, may also be used. © 2000 by CRC Press LLC Selection of measures of effects for many injuries is straightforward. If the measured contaminant concentrations exceed criteria or standards, or if they change the properties of a sediment or soil so that it constitutes a hazardous waste under RCRA, then an injury has occurred (Box 10.2). However, appropriate measures of effects on biological resources are less obvious. The regulations provide the follow- ing acceptance criteria for those measures. BOX 10.2 Natural Resource Injuries Specified in 43 CFR 10.62 Surface water resources Contaminants in excess of Safe Drinking Water Act criteria or standards Contaminants in excess of Clean Water Act criteria or standards Contaminant concentrations sufficient to cause bed, bank, or shoreline sedi- ments to exhibit characteristics identified under the Resource Conservation and Recovery Act Contaminant concentrations sufficient to injure other resources Groundwater resources Contaminants in excess of Safe Drinking Water Act criteria or standards Contaminants in excess of Clean Water Act criteria or standards Contaminant concentrations sufficient to injure other resources Air resources Emissions in excess of National Environmental Standards for Hazardous Air Pollutants Emissions sufficient to injure other resources Geologic resources Contaminant concentrations sufficient to cause bed, bank, or shoreline sedi- ments to exhibit characteristics identified under the Resource Conservation and Recovery Act Soil pH below 4.0 or above 8.5 Sodium absorption ratio above 0.176 Decreased water-holding capacity Impedance of microbial respiration Inhibition of carbon mineralization Restricted mineral access, development, or use Physical/chemical changes in unsaturated zone groundwater Toxicity to soil invertebrates Phytotoxicity Contaminant concentrations sufficient to injure other resources Biological resources Adverse changes in viability: death, disease, behavioral abnormalities, cancer, mutations, physiological (including reproductive) malfunctions, and phys- ical deformations Exceedence of Food, Drug, and Cosmetics Act levels in edible portions Exceedence of state health agency directives in edible portions © 2000 by CRC Press LLC • The measured response often results from exposure to the contaminants (i.e., is well documented in the literature) • The response has been demonstrated in free-ranging organisms • The response can be produced by exposure to the contaminant in con- trolled experiments • The response measurement is practical and produces scientifically valid results The demonstration that the injury is associated with the release requires (1) the identification of a pathway from the source to the receptor and (2) a qualitative or quantitative analysis of the transport and fate processes sufficient to indicate that the exposure that is causing the injury could result from the release. A conceptual model is sufficient for the first step and either a simple transport and exposure model or measurements of contaminant concentrations in intermediate media are sufficient for the second. For example, if the confirmation of exposure is based on measurement of the contaminant in fish tissues, a pathway from the release to the fish must be identified (e.g., leaching of contaminants in soil to groundwater, input to a gaining reach of a stream, and uptake by fish), and its magnitude must be shown to be sufficient to result in the fish tissue levels (e.g., by measuring concentrations in groundwater and stream water and showing that they are sufficient to cause the concentrations in fish). 10.2.3.2 Quantification Phase In effect, the DOI procedure requires that, having completed a qualitative assessment of injuries in the injury determination phase, the assessors then perform a quantitative assessment. The procedure for quantification calls for identification of the services provided by the injured resource, estimation of the baseline service level (i.e., services provided in the absence of the release and contamination), and quantification of the reduction in services due to the release. Factors to be considered in the selection of resource services and measurement methods according to the DOI include • The degree to which the resource is affected • The degree to which the resource may represent a range of related resources and their services • The consistency of the method of measuring services with the require- ments of the economic methodology • Technical feasibility of the methods • Preliminary estimates of the services As in the injury determination, the DOI procedure provides considerations for selection of biological measures to be used in the quantification step. • Emphasize population and ecosystem levels • Choose representative species or habitats • Choose sensitive species, habitats, or ecosystems that provide especially significant services © 2000 by CRC Press LLC • Focus on injured resources • Provide data that can be interpreted in terms of services Quantification of reduction in services requires quantification of a baseline state. The baseline data may come from existing sources, including prior environmental assessments, scientific literature, databases, and other scientific studies. In the absence of adequate existing data, measurements should be taken from reference areas which are comparable to the affected area but are not exposed to the release. Shaw and Bader (1996) argue that the greatest problem with NRDA is its requirement for baseline data. For example, in Prince William Sound, baseline data prior to the Exxon Valdez oil spill were sparse, and some data reflected conditions resulting from a major earthquake in 1964 (Shaw and Bader, 1996). Disturbances that could have influenced apparent effects include winter storms, changes in salinity, changes in suspended sediment, changes in otter populations due to the sediment loads or due to the Marine Mammal Protection Act, and releases of salmon from hatcheries. Shaw and Bader (1996) suggest that the lack of baseline data and high natural variability may have contributed to the trustees’ willingness to settle the damage claims. In the CERCLA context, managers of large sites, especially those managed by government entities, may be expected to have more baseline data than managers of small, privately owned sites. Factors to be considered in the quantification of reduction in services include • The total area, volume, or numbers of the resource affected • The degree to which the resource is affected • The ability of the resource to recover • The proportion of the available resource affected • The services normally provided by the resource that have been reduced as a result of the release The choice of baseline can have a large effect on the estimate of injury. For example, the NRDA for the Blackbird Mine focused on the loss of spawning salmon in the creek due to metal contamination beginning in the early 1900s (Renner, 1998). The population prior to the mine of 2000 to 3000 salmon might have been used as a baseline, but that would not reflect the fact that salmon runs have been diminished by dam construction and commercial fishing since that time. Instead, a baseline of 200 salmon was established, based on current runs in similar streams in the region. 10.2.3.3 Damage Determination The natural resource damages include the cost of restoring the resources, the diminution of the value of resource services prior to restoration of those services, and the cost of performing the assessment (Figure 10.1). Restoration costs and assessment costs may be controversial, if the responsible party believes that the trustee is performing a gold-plated assessment or restoration, but the quantification of these components of the damages is not controversial. The difficulties arise in the quantification of the diminution of the value of services. Those values may © 2000 by CRC Press LLC [...]... addition, ecological risk assessment has not focused on the monetary value of the endpoint entities and processes as has NRDA However, as cost–benefit analysis becomes an increasingly important adjunct to risk assessment, NRDA and ecological risk assessment practices may converge The NRDA process is particularly problematical for federal agencies that have contaminated sites for which they are trustees For. .. above However, the two are not incompatible, and three of the phases (preassessment screen, injury determination, and quantification) could be thought of as tiered assessments with increasingly rigorous analysis NRDAs might benefit from the logic of the ecological risk framework and from the ability to adopt risk assessments performed for the remedial investigation, but the legal advantage of the rebuttable... (Section 10. 2.5) suggests that the NRDA system may be evolving toward greater rationality and efficiency The relationship of NRDA to risk assessment in general is unclear The NRDA procedure was published in 1986, the same year as the first ecological risk assessment framework (Barnthouse and Suter, 1986; DOI, 1986) As a result, the structure of the NRDA procedure is quite different from the ecological risk. .. been legally accepted for determining natural resource damages (NOAA Panel on Contingent Valuation, 1993) However, the method is highly controversial Studies have shown that it is unreliable as an estimator of how much people will actually pay for preservation of a resource, and results are sensitive to the context and wording of the questions 10. 2.4 POST -ASSESSMENT PHASE The post -assessment phase consists... the services that the environment performs for human society As discussed in the preceding chapter, the remediation performed following the RI/FS may actually result in a net degradation of the environment However, the NRDA regulations do not allow for restoration of the loss of natural resource services due to the prior remediation Damages may be collected only for injuries due to contamination Because... is performed by the responsible party, even the payment of damages can be eliminated (Renner, 1998) NOAA (1996a) has included provisions for collaboratively moving toward restoration in its regulations for NRDA under the Oil Pollution Act, and has developed guidance for identifying equivalent habitats (NOAA, 1996b) Reportedly, the DOI is considering similar provisions in revised regulations for NRDA... under CERCLA (Renner, 1998) 10. 2.6 CONCLUSIONS The relationship of the CERCLA RI/FS risk assessments to the CERCLA NRDAs has been slight to nonexistent Logically, there should be a single process of determining the risks to the natural environment posed by contaminants and selection of a remedial action that not only removes or destroys contaminants that pose significant risks, but also restores the... travel to an uncontaminated beach), hedonic pricing (e.g., how much more do people pay for a house that does not adjoin a contaminated site), and contingent valuation (a survey technique that asks how much one would be willing to pay to prevent loss of a resource) Nonuse values are consumer surpluses that people may have for resources that they are not using They include willingness to pay for the option... Estimates of lost-use values may be based on behavior and therefore are reasonably reliable Even when contingent valuation is used to estimate use values, the reasonableness of the results can be checked against the actual amounts spent on vacations, etc Because there is no market for nonuse values and no behavior is involved in not using a resource, only survey techniques can be used for measuring nonuse... POST -ASSESSMENT PHASE The post -assessment phase consists of documenting the results of the assessment, creating an account for the damages awarded, developing a restoration plan, and performing the restoration Restoration may include additional removal or destruction of contaminants beyond the remediation performed under the CERCLA record © 2000 by CRC Press LLC of decision Alternatively, restoration . adjunct to risk assessment, NRDA and ecological risk assessment practices may converge. The NRDA process is particularly problematical for federal agencies that have contaminated sites for which. of performing the assessment (Figure 10. 1). Restoration costs and assessment costs may be controversial, if the responsible party believes that the trustee is performing a gold-plated assessment. likely to be needed for spills or CERCLA emergency actions, but is unlikely to be needed for typical CERCLA NRDAs. Where an RI/FS has been performed, the ecological risk assessments in those documents