C:\Documents and Settings\Mike\My_Documents ... - City of Biddeford

4 downloads 25 Views 3MB Size Report
over time asymptotically approaching a steady state value at which the deposition and loss terms would be equal. ...... United States Department of Agriculture.
Risk Assessment for the Evaluation of Direct and Multi-pathway Impacts of Emissions from the Maine Energy Recovery Company Facility, Biddeford, Maine Prepared for: Maine Energy Recovery Corporation by: Michael R. Ames, Sc.D., Stephen G. Zemba, Ph.D., P.E., Kyle Satterstrom, and Laura C. Green, Ph.D., D.A.B.T.

June 2006

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

Contents Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ES–1 1

Introduction and Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–1 1.1 Basic facility information and site description . . . . . . . . . . . . . . . . . . . . . . . . . 1–1 1.2 Risk assessment methods and study area characteristics . . . . . . . . . . . . . . . . . . 1–5 1.3 The concept of a ‘most exposed individual’ . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–8 1.4 Uncertainty and conservatism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–9 1.5 The meaning of risk estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–10 1.6 Risk assessment basis and organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1–11

2

Facility emissions characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–1 2.1 Facility process information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–1 2.2 Compounds of Potential Concern (COPCs) . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–2 2.3 COPC emission rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–4 2.3.1 Procedures for estimating stack COPC emission . . . . . . . . . . . . . . . . . 2–5 2.3.2 Procedures for estimating odor scrubber system COPC emission rates 2–6 2.3.3 Data and procedures for estimating stack COPC emission rates under process upset conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–8 2.3.3.1 Combustion control upsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–10 2.3.3.2 Spray dryer absorber upsets . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–10 2.3.3.3 Baghouse/fabric filter upsets . . . . . . . . . . . . . . . . . . . . . . . . . . 2–11 2.3.3.4 Combustion startup/shutdown conditions . . . . . . . . . . . . . . . . 2–11 2.3.3.5 Effects of upset emissions on long-term average . . . . . . . . . . 2–12 2.3.4 Procedures for non-detected compounds . . . . . . . . . . . . . . . . . . . . . . . 2–13 2.3.5 Chromium speciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–14 2.3.6 Mercury speciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–15 2.3.7 Summary of COPC emission rates used in Maine Energy Risk Assessment modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2–18 2.3.8 Recent variations and long-term trends in COPC emissions . . . . . . . . 2–22

3

Air dispersion and deposition modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–1 3.1 Background and general air modeling description . . . . . . . . . . . . . . . . . . . . . . 3–4 3.2 General modeling options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–5 3.3 Receptor locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–6 3.4 Meteorological data processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–11 3.5 COPC deposition estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–19 3.5.1 Plume depletion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–20 3.5.2 Particulate-phase COPC deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–20 3.5.3 Vapor-phase COPC wet deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–21

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ii

3.6 3.7

3.5.4 Vapor-phase COPC dry deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 3–28 Modeling of startup and shutdown emissions . . . . . . . . . . . . . . . . . . . . . . . . . 3–30 Summary of atmospheric dispersion and deposition modeling results . . . . . . 3–30

4

Exposure scenario selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4–1

5

Estimation of media concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–1 5.1 COPC concentrations in soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–4 5.2 COPC concentrations in produce, grain, and vegetation . . . . . . . . . . . . . . . . . . 5–8 5.3 COPC concentrations in livestock and related farm products . . . . . . . . . . . . . 5–10 5.4 COPC concentrations in surface water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–11 5.4.1 COPC loading to nearby ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–12 5.4.2 COPC dissipation in nearby ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–19 5.4.3 COPC partitioning in nearby ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–20 5.4.4 Bounding estimates of COPC impacts on Saco River water . . . . . . . . 5–23 5.5 COPC concentrations in fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5–24 5.5.1 The use of a site-specific value for the BAF fish for mercury . . . . . . 5–25

6

Quantifying exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6–1

7

Risk and hazard characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7–1

8

Uncertainty evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8–1 8.1 Facility characterization—emission uncertainties . . . . . . . . . . . . . . . . . . . . . . . 8–2 8.1.1 Estimation of long-term emission rates from maximum rather than average measured concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8–2 8.1.2 Extrapolation of risks to account for un-analyzed compounds . . . . . . . 8–3 8.1.3 Treatment of COPCs below detection limits in stack tests . . . . . . . . . . 8–5 8.1.4 Use of a DRE to estimate some COPC emission rates . . . . . . . . . . . . . 8–7 8.1.5 Chromium speciation in emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8–8 8.1.6 Mercury speciation and distribution in emissions . . . . . . . . . . . . . . . . . 8–9 8.2 Air dispersion and deposition modeling uncertainties . . . . . . . . . . . . . . . . . . . 8–10 8.2.1 Superposition of maximum concentration and deposition values . . . . 8–11 8.2.2 Bounding estimate for COPC concentrations in the Saco River . . . . . 8–13 8.2.3 Receptor grid spacing at far-field maximum impact locations . . . . . . 8–15 8.3 Estimation of media concentrations uncertainties . . . . . . . . . . . . . . . . . . . . . . 8–17 8.3.1 Use of non-zero kse in watershed soil concentration calculations . . . 8–17 8.3.2 Bounding estimates of COPC levels in fish in the Saco River . . . . . . 8–17 8.3.3 Site-specific, BAF fish values for mercury . . . . . . . . . . . . . . . . . . . . . 8–20 8.4 Uncertainties in quantifying exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8–21 8.5 Inherent uncertainties in toxicologic data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8–22 8.5.1 Toxicity of coplanar PCB congeners . . . . . . . . . . . . . . . . . . . . . . . . . . 8–24

9

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9–1

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

iii

10

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10–1

Appendix I

Cambridge Environmental, Risk Assessment Protocol (RAP), Review Comments on the Protocol by TechLaw, and Cambridge Environmental’s Response of Comments

Appendix II

COPC-specific properties

Appendix III Data used to calculate COPC emission rates Appendix IV Air Dispersion Modeling and Data Files Appendix V

Air Dispersion Modeling Results Figures

Appendix VI Calculated Concentrations of Compounds of Potential Concern (COPC) in Environmental Media Appendix VII Effects of the Proposed Scrubber Stack Height Increase (TRC)

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

iv

Executive Summary The Maine Energy Recovery Company processes municipal solid waste at its facility in Biddeford, Maine (referred to in this report as the Maine Energy facility). The facility produces refuse-derived fuel (RDF) from the municipal solid waste and combusts this fuel in boilers to generate steam and produce electricity. Incidental to its operations, the facility emits air pollutants from its boiler stack and odor control system. To determine whether these emissions present significant risks to human health, Cambridge Environmental Inc. has performed a risk assessment which evaluates the potential direct and indirect exposures of individuals living near the facility, as well as the likelihood that these exposures might lead to adverse health effects. Cambridge Environmental Inc. conducted a similar risk assessment of the facility’s emissions in 1996. The 1996 risk assessment found that these emissions would not lead to significant risks to human health. However, since the time of the 1996 risk assessment, several conditions have changed that warrant reexamination of the health risk assessment to determine if the 1996 conclusions remain valid. Significant changes that have transpired include: • • •

the addition of the odor control system at the facility; the development of new regulatory guidance and models for conducting multi-pathway risk assessments; and the City of Biddeford’s enactment of its Air Toxics Control Ordinance designed to evaluate emissions of facilities that release potentially hazardous air pollutants and to ensure that such emissions do not cause impacts that are above the City’s health-based ambient air limits.

The risk assessment methods and results described in this report update the 1996 Maine Energy risk assessment to contemporary standards and expand upon the previous work to include an evaluation of emissions from the odor control system. The pollutants emitted from the stack of the Maine Energy facility at the greatest rate are those referred to as ‘criteria pollutants.’ Criteria pollutants include very fine particles (particulate matter, classified as PM10 and PM2.5), sulfur dioxide (SO2), nitrogen oxides (NO and NO2, or NOx), lead (Pb), and carbon monoxide (CO). The U.S. Environmental Protection Agency (U.S. EPA) has established National Ambient Air Quality Standards (NAAQS) and the Maine Legislature has established Maine Ambient Air Quality Standards for each criteria pollutant designed to protect public health with an adequate margin of safety. The Maine Energy facility’s emission of most criteria pollutants is monitored at the facility’s stack. Compliance with the NAAQS is verified through a combination of modeling and direct measurement of ambient levels at selected locations.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–1

Although these criteria pollutants are generally released in the greatest quantities, there are many other compounds that the Maine Energy facility releases in much smaller quantities. These numerous non-‘criteria’ pollutants are the focus of this study. Many of these compounds are regulated as Hazardous Air Pollutants under section 112 of the Clean Air Act, and also under the City of Biddeford’s Air Toxics Ordinance. These compounds are evaluated according to risk assessment methodologies that have evolved over the past few decades. Various metals and products of incomplete combustion account for most of the compounds of potential concern (COPCs) emitted from the Maine Energy facility’s main boiler stack. Any metals present in the waste received at the facility are not destroyed in the combustion process, and, although most are removed as bottom ash or by pollution control equipment, very small levels are emitted to the environment. Although most organic compounds present in the municipal solid waste and RDF are destroyed in the combustion process, some organic compounds are not fully destroyed and other organic compounds are formed in the combustion zone. Together, these compounds are often referred to as products of incomplete combustion (PICs). In addition, the odor control and treatment system releases various volatile organic COPCs in the effluent of its three exhaust stacks. The fact that the Maine Energy facility releases small quantities of COPCs is not unusual. For example, automobiles emit a myriad of compounds that, if breathed at concentrations present in the tailpipe, could be hazardous to health. From experience, however, people are exposed to undiluted auto exhaust only for limited amounts of time (if at all), and tailpipe emissions disperse rapidly once introduced to the atmosphere. Thus, like any other source of air pollution, a relevant issue regarding the compounds released from the Maine Energy facility is the degree to which they become dispersed and diluted in the atmosphere. Additionally, to fully assess potential exposures to these emitted compounds one must consider other plausible ways humans may be exposed to them. Some compounds are capable of entering and concentrating within soil, water, and plants, thereby becoming available to people through means other than inhalation. This report focuses on the evaluation of the small levels of COPCs that are released by the Maine Energy facility and evaluates the various ways that individuals could be exposed to the compounds, starting with the direct inhalation of the compounds while they are present in air, followed by indirect pathways whereby compounds deposit to the ground, become incorporated within soils and foodstuffs, and are then consumed either inadvertently (within soil) or purposely (within people’s diets). The consideration of both direct and indirect exposure pathways is termed multi-pathway exposure assessment, and represents the attempt to develop upper-end estimates of a person’s total potential exposure to compounds released from the Maine Energy facility. This risk assessment is based on measurements of the compounds that are released from the Maine Energy facility. Drawing from experience with similar facilities, the U.S. Environmental Protection Agency (EPA) has compiled a list of COPCs typically emitted from waste-to-energy facilities, and the Maine Department of Environmental Protection (DEP) requires the Maine

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–2

Energy facility to periodically test for these compounds in its boiler stack emissions. In addition, through cooperative arrangements with the City of Biddeford, the Maine Energy facility has installed an odor control system, and the three exhaust stacks of this system have been tested for the presence of a wide variety of compounds. These data are used to quantify the rates at which any of these compounds are released from the facility. Table ES-1 is a list of the compounds measured in emissions from either the boiler stack or the odor handling system, and which are hence considered in the risk assessment. As described further in this report, the emissions estimates are conservatively health protective for the purposes of this risk assessment. Given estimates of the rates at which these compounds are released from the Maine Energy facility, a series of computer-based mathematical models are used to predict the concentrations and distribution of the compounds that occur throughout the environment. These models are based on chemical and physical principles as well as empirical data, and they have been developed and updated over time by the U.S. EPA and others. To ensure that the models do not underestimate the degree to which compounds might accumulate in the environment and food chain, most uncertainties in the models have been resolved in a manner that over-predicts the concentrations likely to occur. The modeling is thus designed to make high-end estimates of the degree to which people may be exposed to compounds released by the Maine Energy facility. The philosophy of developing high-end exposure estimates also influences the scenarios examined within the risk assessment. These scenarios focus on a model of the people, animals, and plants living in the vicinity of the Maine Energy facility that have the highest potential to encounter compounds emitted by the facility. Table ES-2 lists the exposure scenarios considered within the risk assessment. The goal of estimating high-end exposure scenarios is met in three ways: • • •

scenarios are evaluated at locations where the highest concentrations are predicted to occur due to emissions from the Maine Energy facility; the types of personal exposure scenarios considered are those for people that consume large amounts of the foods that tend to accumulate compounds from the environment; and the rates at which compounds are encountered (e.g., through the amount of food consumed) are assumed to be at high-end or higher-than-average values.

The exposure scenarios listed in Table ES-2 reflect the first two criteria. The individuals considered in the human health risk assessment do not represent actual people, but rather serve as examples of the populations that are expected to encounter higher-than-average exposures to compounds released by the Maine Energy facility. The hypothetical people that are studied include a resident, a recreational farmer, and a recreational fisher; exposures to both children and adults are evaluated. The resident exposure scenario is intended to characterize individuals who engage in typical activities and who live in the vicinity of the location where emissions from the facility are expected to produce the highest concentrations in the environment and hence have the greatest potential to affect soil, homegrown vegetables, and drinking water that can serve as indirect avenues of exposure.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–3

The recreational farmer and fisher scenarios represent high-end exposures of individuals who derive a substantial portion of their food from home-grown or local sources. Like the resident, both the recreational farmer and fisher are assumed to live near the location predicted to be most affected by emissions from the Maine Energy facility, but their exposure profiles are supplemented with the consumption of locally-derived foods that tend to accumulate compounds to the greatest degree. Thus, the recreational farmer is assumed to raise a substantial portion of his or her meats, eggs, and dairy products near the location where the influence of the Maine Energy facility is predicted to be highest. Similarly, the recreational fisher, is assumed to consume a substantial quantity of fish which have been caught at locations near the Maine Energy facility. The amount of local foods raised and caught by the recreational farmer and fisher constitute a substantial portion of their diets. In fact, the EPA risk assessment guidance uses the term subsistence farmer and fisher to describe these scenarios.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–4

Table ES-1

Compounds evaluated in the Maine Energy facility risk assessment.

Metals/Inorganics Arsenic * Beryllium * Cadmium * Chromium (total) * Chromium (hexavalent) Copper Lead * Mercuric chloride Mercury * Methylmercury Nickel * Selenium Silver Tin Vanadium Zinc Hydrogen chloride

Semi-volatile Organic Compounds Benzoic Acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Diethyl phthalate Di-n-butylphthalate Methyl naphthalene, 2Methyl phenol, 3&4Methyl phenol, 2Naphthalene Phenol

Polychlorinated dibenzo(p)dioxins and furans (PCDD/PCDFs) and polychlorinated biphenyls (PCBs) 2,3,7,8-TCDD * 1,2,3,7,8-PCDD * 1,2,3,4,7,8-HxCDD * 1,2,3,6,7,8-HxCDD * 1,2,3,7,8,9-HxCDD * 1,2,3,4,6,7,8-HpCDD * OCDD * 2,3,7,8-TCDF * 1,2,3,7,8-PCDF * 2,3,4,7,8-PCDF * 1,2,3,4,7,8-HxCDF * 2,3,4,6,7,8-HxCDF * 1,2,3,7,8,9-HxCDF * 1,2,3,4,7,8,9-HpCDF * OCDF * PCBs

Organic Compounds Acetone Benzene Bromomethane 2-Butanone Carbon disulfide Chloromethane Chloroform Cyclohexane 1,4-Dichlorobenzene Ethanol Ethylbenzene Freon 11 Freon 12 Heptane Hexane Methylene chloride 2-Propanol Styrene Tetrachloroethylene Toluene 1,1,1-Trichloroethane 1,2,4-Trimethylbenzene Vinyl chloride Xylenes

* Compounds also evaluated in the 1996 health risk assessment

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–5

At each stage of the environmental transport and exposure analyses, mathematical algorithms are used to model a variety of physical, chemical, or biological processes. Properties of the facility, its surroundings, the COPCs, and the hypothetical human receptors are used as inputs to these algorithms to estimate the various derived quantities, such as COPC emission rates, concentrations in water, or human intake rates. It should be noted that, in constructing the profile of assumptions for an exposure scenario, each assumption is not taken at an extreme value, but rather the suite of assumptions is designed to produce a high-end exposure estimate that remains within plausible limits. Thus, there may be individuals who consume more of a particular type of food than what is assumed in the risk assessment, but there are likely to be few (if any) people who live at the location most affected by facility emissions, grow and raise the majority of their own food, and who live at this location and in this manner for thirty years. Hence it is unlikely that there are any individuals who receive a greater degree of exposure to compounds from the Maine Energy facility than is estimated for the hypothetical exposure scenarios. Some of the important constructs and assumptions used in the risk assessment are as follows: •

Risks are assessed for hypothetical "most-exposed individuals" (MEIs), rather than for a real population with a range of potential exposures. Because MEIs are designed to receive improbably high exposures, risk estimates deemed acceptable for MEIs suggest that the source of contaminants (the Maine Energy facility, in this case) will not cause unacceptable risks for actual persons.



Exposures of MEIs to COPCs in air, soil, and homegrown vegetables are evaluated at the locations where impacts from the Maine Energy facility to each of these media are projected to be highest, despite the fact that the maximums are in different locations.



Exposures of MEIs to COPCs in drinking water are evaluated using bounding estimates of the maximum possible concentrations that could be present in the Saco River.



MEIs in the farming and fishing scenarios are assumed to produce or obtain practically all of their vegetables, milk, meat, and fish from these locations, even though actual land use in the Biddeford/Saco area suggests far lower consumption rates are likely to occur at the locations where the exposures are assessed.



MEIs are assumed to have the same exposure to COPCs for a period of 30 to 40 years, in accordance with U.S. EPA guidance for the evaluation of high-end exposure estimates.



Both adult and child MEIs are considered. Separate consideration of a child MEI is important because exposure rates, when normalized by body weight, are often higher for children than adults.



In evaluating lifetime potential cancer risks caused by the MEI’s exposures to small doses of emitted compounds, the carcinogenic potency of the contaminants is based on

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–6

the response of laboratory rodents (in most cases) or humans (for some metals) to extremely high doses. In fact, it is not known if there would be any carcinogenic response at all at such low doses. •

For evaluating the potential chronic adverse effects of various (non-carcinogenic) contaminants, the doses to the MEIs are compared with reference doses. These reference doses have been derived by applying margins of safety to the lowest doses observed to cause harm to humans or laboratory animals. There is no reasonable expectation of adverse effect from doses near and below these reference doses.

Table ES-2

Summary of Risk Assessment Exposure Scenarios

Receptor

Exposure Pathways

Location

Human health risk assessment — adults and children Resident

Inhalation Ingestion of vegetables, drinking water, and soil

Maximum impact point

Recreational Farmer

Inhalation Ingestion of vegetables, drinking water, and soil Ingestion of home-raised meats and eggs

Maximum impact at locations of cultivatable land

Recreational Fisher

Inhalation Ingestion of vegetables, drinking water, and soil Ingestion of locally-derived fish

Maximum impact water body

Once the exposure scenarios for the MEIs are defined and the algorithms for estimating the COPC concentrations in relevant environmental media are applied, individual MEI exposure rates are estimated for each of the compounds in Table ES-1 and for each of the hypothetical exposure scenarios listed in Table ES-2. To evaluate whether these exposures might result in significant risks to an exposed individual’s health, the exposure rates are evaluated with respect to compound-specific toxic potency and health effects benchmark exposure values. Two types of health-based evaluations are made within the human health risk assessment. First, the potential for each compound to increase an exposed individual’s lifetime cancer risk is assessed. Second, the likelihood that each compound might cause adverse health effects other than cancer is evaluated.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–7

The incremental, or excess, lifetime cancer risk for an exposed individual is calculated by multiplying each compound’s predicted exposure rate with its estimated potency to cause cancer in humans. The resulting cancer risk estimate is the exposed individual’s additional risk of getting cancer in his or her lifetime, above and beyond the background level that people get cancer from all causes, which is 1 in 2 for men and 1 in 3 for women. This excess risk is compared with regulatory benchmark levels to evaluate whether the estimated risk is acceptable. Historically, the Maine Department of Human Services has established an acceptable incremental cancer risk level of 1 in 100,000 (or 10 in 1,000,000). This risk level may be expressed in scientific notation as 10–5 or 1 E-5, and which represents an increase in cancer risk above the background level of 0.003% for a woman and 0.002% for a man. The potential for emitted compounds to cause noncancerous health effects is evaluated by comparing the predicted level of exposure for each compound with a level of exposure that is believed to be safe, i.e., a level that can be tolerated without risk to health (unlike incremental cancer risk, where a risk is assumed for any level of exposure). The ratio of the estimated exposure to the safe, or reference, exposure level is referred to as the compound’s hazard quotient (HQ). If a compound’s HQ is less than 1, the exposure level is less than the reference exposure level, and no adverse health effects are expected to occur. For any given scenario, the sum of all the HQs is referred to as the hazard index (HI). If the HI is less than 1, then, overall, no adverse effects are expected. Although the health effects evaluated using the hazard index include diseases that affect different organs which differ among compounds, these broad categories of potential health effects are grouped because they are evaluated in a similar manner. If the hazard ratio is greater than one, the level of exposure exceeds the level thought to be potentially harmful, and the possibility of adverse health effects might exist. However, since the reference doses and concentrations used to characterize safe values frequently embody safety factors, it is incorrect to conclude that hazard ratios greater than one will in fact correspond to the actual incidence of health effects. Rather, hazard ratios exceeding one are indicators of the possibility of adverse health effects. Two types of hazard quotients are assessed to reflect different types of exposures to compounds emitted from the Maine Energy facility. Chronic hazard quotients are calculated to assess health effects that might be associated with exposure to compounds that could occur over extended periods of time. Acute hazard quotients are evaluated to gauge the nature of exposure to elevated concentrations of compounds in air that are predicted to possibly occur on an occasional basis. The overall results of the risk assessment of the Maine Energy facility are summarized in Tables ES-3 and ES-4. The total estimated lifetime incremental risks of cancer are listed in Table ES-3. These values reflect the sum of the estimates for all known or potentially carcinogenic compounds found in the Maine Energy facility emissions. The compounds and exposure pathways that contribute principally to each cancer risk estimate are also provided in Table ES-3. The incremental risk levels due to Maine Energy facility emissions are larger for the recreational farmer and fisher scenarios, reflecting the additional indirect exposures included in these

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–8

scenarios. Objectively, the lifetime incremental cancer risk estimates are quite small, especially when compared with the background (overall) risk of getting cancer. As can be seen from the values in Table ES-3, the highest excess lifetime cancer risks associated with emissions from the Maine Energy facility total an incremental risk of 3 in 1,000,000 for the recreational farmer. This estimated risk level is more than a factor of three smaller than the regulatory benchmark of 10 in 1,000,000, and it represents an increase of about only 0.001% above background cancer incidence levels. Table ES-4 presents risk estimates for compounds that, at sufficient levels of exposure, could cause adverse health effects other than cancer. The highest overall hazard index is well below 1 for both chronic (long-term) and short-term risks. Potential short-term risks have been evaluated based on both the facility’s emission levels under normal and upset operating conditions. The greatest HI is 0.09 for the fishing scenario as evaluated in the unnamed pond on the Goosefare Brook. This value is far below a level at which adverse effects might occur. Additionally, these values represent the sum of all of the hazard ratios for the individual compounds, and hazard ratios should, strictly, be separated into categories of specific health effects. More detailed information on these risk estimates, including risk estimates for each COPC under each exposure scenario, is presented in Chapter 7 of the risk assessment report. Most of the risk estimates presented in Tables ES-3 and ES-4 correspond to the estimates of emissions from the Maine Energy facility when it is operating under normal operational conditions, at full capacity, continuously throughout the year. Since the facility does not always operate at full capacity (e.g., it is shut down for periods of maintenance each year), the emission rates, and hence risk estimates, are overestimated, even accounting for potential upset conditions when emissions might be higher over short periods. Even so, a series of risk estimates is presented in the uncertainty section of the risk assessment report (see Chapter 8) based upon the highest emission rates measured during facility testing. These risk estimates tend to be about twice as large as the best-estimate values (at full operational loading) summarized in Tables ES3 and ES-4. This factor of two does not alter conclusions relative to typical regulatory risk criteria, as incremental cancer risks would remain well below 10 in 1,000,000, and hazard indices well below one. Thus, basing risk estimates on the highest measured emission rates would not lead to risk estimates of significant concern. Chapter 8 also contains risk estimates that have been calculated using somewhat different modeling assumptions than have been applied in the baseline estimates. Some of these sensitivity and uncertainty analyses result in slightly higher potential risk estimates, but none of them produce estimated risk indices that exceed the health-based criteria levels. Table ES-4 also presents short-term risk estimates that account for occasional “upset” conditions when operations of the Maine Energy facility deviate outside of their normal ranges. As described in Chapter 2, the Maine Energy facility is designed and operated to minimize the effects of process upsets, and some “upset” conditions that occur in practice (such as facility shutdowns) actually lead to decreased long-term emissions. Consequently, the risk assessment evaluates potential acute risks associated with short-term increases in facility emissions. The upset scenarios summarized in Table ES-4 indicate a worst–case hazard index 0.01 over a 1-hour

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–9

period, and a maximum sum of ratios of ambient concentrations to Biddeford 24-hour AALs of 0.03, indicating overall safety factors of 30 to 100 between (1) the ambient concentrations of COPCs that might result during a facility upset and (2) levels of potential concern. As another gauge of potential health risks due to emissions from the Maine Energy facility, the highest modeled concentrations of COPCs due to emissions from the Maine Energy facility were compared with applicable Ambient Air Limits (AALs) established by the City of Biddford’s Air Toxics Ordinance. No predicted COPC concentrations exceed any 24-hour or annual-average AALs at any location. At the worst-case, the COPC nearest its AAL is more than 100 times smaller than the permissible level. When one considers the incremental cancer risk estimates and non-cancer hazard indices in Tables ES-3 and ES-4, it is important to recall that these values are based on parameters and methods that intentionally overestimate the likely risks that will actually occur. This is done so that, if the modeled risks and hazard indices are below regulatory levels of significance, then the actual risks and hazards will definitely be below these levels of significance. Among the specific portions of the risk assessment that lead to overestimation of exposures and risks are the following: •

Exposures to each pollutant are evaluated at the locations where the impacts from the Maine Energy facility are the greatest despite the fact that these maxima may be in different locations.



Exposures due to consumption of drinking water from the Saco River are based on the pollutant concentrations that would exist in the water if all of the facility’s emissions entered the river directly. This is a very significant overestimate of these exposures (see Section 8.2.2).



The consumption rates for homegrown produce, meat, and dairy products for the farming scenario are collectively higher than would likely occur at the maximum impact location over the 40-year exposure period (see Table 8.11 for values).



The consumption rates for fish caught in local ponds are higher than is likely possible for these waterbodies over the long-term (i.e., for adults an average consumption rate of 66 pounds of fish per year for 24 years, see Table 8.11).



Short-term exposures that might occur during off-normal operation of the facility were evaluated as if all of the system upset conditions occurred simultaneously. Because these upsets only occur at most a few times a year on an individual basis, it is very unlikely that they would all happen at once.



The toxicological data used to evaluate whether these exposures would cause adverse health effects are based on animal and/or human exposures at much higher levels than would be experienced due to emissions from Maine Energy. Because of the safety

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–10

factors built into these data, the chances that the the modeled exposures might cause actual adverse health effects are likely to be much lower than the cancer risk and noncancer hazard indices suggest. Thus, while the estimated cancer risks and non-cancer hazard indices shown in Tables ES-3 and ES-4 are well below the regulatory levels of concern, the values are based on significant overestimations of the impacts from Maine Energy’s emissions. If the pollutant exposure levels and potential health risks that occur due to emissions from the Maine Energy facility were evaluated using methods and parameters that more closely reflect actual conditions, the risks and hazard indices would likely be much lower than those shown below (e.g., the actual incremental lifetime cancer risks would be well below the 1 to 4-in-a-million levels shown). This intentional and significant overprediction of exposures and risks is a fundamental part of the risk assessment process which is designed to provide a wide margin of safety and to allow for the assessment of a wide range of facilities using standard and well-reviewed methods.

Table ES-3

Summary of Incremental Cancer Risk Estimates a

Receptor

Incremental cancer risk estimate (Target limit = 10 in 1,000,000) b

Principal exposure pathways

Principal COPCs and fraction of total risk

Resident

2 in 1,000,000

drinking water homegrown produce

tetrachloroethene 34% vinyl chloride 22% PCDD/Fs 20%

Recreational Farmer

4 in 1,000,000

homegrown animal products/produce

PCDD/Fs 63% tetrachloroethene 16% vinyl chloride 10%

PCDD/Fs 39% locally caught fish homegrown produce tetrachloroethene 25% drinking water vinyl chloride 16% a The risk estimates shown here include risks due to both direct (report Table 7-10) and indirect (report Table 7-9) exposures. The estimates are based on continuous operation of the facility, using compound emission rates measured under stressed operating conditions, and for the exposure pathways shown in Table ES-2. b Incremental cancer risks shown here are reported in the body of the report in scientific notation; a risk of 8 in 100,000,000 may be also shown as 8 × 10–8 or 8 E-8. Recreational Fisher

2 in 1,000,000

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–11

Table ES-4

Summary of Hazard Indices and Total Ratios to Biddeford 24–hour Ambient Air Limits Used to Evaluate Risks of Non-Cancer Health Effectsa

Receptor

Hazard Index (Acceptable limit = 1)

Principal exposure pathways

Principal COPCs and fraction of total risk

Chronic (Long-Term) Exposure Scenarios Residentb

0.08

inhalation drinking water soil ingestion

n-butanol 68% mercuric chloride 7% 1,3 dichlorobenzene 6%

Recreational Farmerb

0.06

inhalation drinking water soil ingestion

n-butanol 72% 1,3 dichlorobenzene 7% 1,2,4 trimethylbenzene 6%

Recreational Fisher

0.2

locally caught fish drinking water

methyl mercury 76% n-butanol 17% dichlorobenzene 2%

Short-Term Exposure Scenarios 1-Hour Basis Hazard Index normal operation 1-Hour Basis Hazard Index upset conditions Total of Ratios to 24-Hour AALs normal operation Total of Ratios to 24-Hour AALs upset conditions

0.003

0.01

0.02

0.03

Inhalation

chloroform 20% methanol 17% propanol, 2- (isopropyl alcohol) 16%

Inhalation

arsenic 23% lead 16% hydrogen chloride 13%

Inhalation

benzene 37% methanol 18% hydrogen chloride 17%

Inhalation

benzene 33% lead 21% hydrogen chloride 17%

a

The risk estimates shown here include risks due to both direct (Table 7.1) and indirect exposures (Tables 7.3, 7.4, and 7.6). The estimates are based on continuous operation of the facility, using compound emission rates measured under stressed operating conditions, and for the exposure pathways shown in Table ES-2. b The maximum Hazard Indices for these scenarios are for the child receptors

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–12

In summary, • • • •

Emissions of a wide range of compounds from the Maine Energy facility have been measured; The highest expected personal exposures to these compounds by direct and indirect pathways have been modeled using methods that, in general, significantly over-predict actual exposure levels; The modeled exposures are estimated to produce less than a 0.001% increase in the risk of cancer and are well below the U.S. EPA’s reference dose and concentration levels for non-cancer effects; and The worst-case predicted concentrations of COPCs due to Maine Energy facility emissions are well below the Ambient Air Limits established by the City of Biddeford to protect public health.

Based on these findings, emissions of the Maine Energy facility present no significant risks to people living in its vicinity. The reader is encouraged to explore additional portions of the risk assessment report. The report is organized in the logical progression of the risk assessment, starting with the description of compound emission rates, and followed subsequently by air dispersion modeling, environmental fate-and-transport modeling, exposure estimation, and culminating with the calculation of risk estimates to human health and the environment, along with a discussion of uncertainties. Sufficient detail is provided in the main body of the report and its appendices to reproduce the calculations if desired, but the qualitative descriptions of the underlying approaches, assumptions, and philosophies are likely of greater value to the general reader. Through these descriptions, a sense for the risk assessment process can be gained from the report without the need for deriving the mathematical details of its numerous equations, tables, and numerical values.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

ES–13

1

Introduction and Background

This section provides background information on the operations of the Maine Energy Recovery Company’s Biddeford facility (Maine Energy facility), including diagrams, plan maps, and other information. The area surrounding the Maine Energy facility is also described since topography and land use have an effect on the estimation of atmospheric dispersion and deposition of facility emissions, concentrations of compounds of potential concern in environmental media, and relevant exposure scenarios. A description of the facility and surroundings also helps to establish a contextual sense for the risk assessment.

1.1

Basic facility information and site description

The Maine Energy facility is located in the central downtown area of Biddeford. Figure 1.1 depicts a topographic map of the area, centered roughly at the location of the facility. The pinkcolored portions at the center of the map indicate the urbanized areas of the Cities of Biddeford and Saco, located to the south and north, respectively, of the Saco River. Generally, terrain elevations increase in directions to the north and south of the Saco River valley. A relatively large hill is located to the southeast of central Biddeford. The two small ponds included in the modeling are also identified. Figure 1.2 shows an aerial photograph of the study area in the immediate vicinity of the Maine Energy facility; Figure 1.3 shows a wider aerial photograph of the study area indicating the locations of the Southern Maine Medical Center and local schools and educational institutions identified by Maine GIS (http://apollo.ogis.state.me.us). The Maine Energy facility is designed to process approximately 1200 tons of municipal solid waste (MSW) each day. Mechanical equipment is used to separate metal and other noncombustible materials that are recycled or disposed. The remainder of the MSW, which includes primarily paper, plastic, and food waste, constitutes the refuse-derived fuel that the facility combusts in two boilers. The steam produced by the boilers is fed through turbines to produce up to 22 megawatts of electricity. The byproducts of MSW combustion include residual ash, which is trucked to a landfill, and flue gases that are emitted to the air and subject to regulations enacted and enforced by the Maine Department of Environmental Protection (DEP). Prior to atmospheric release, flue gas from each boiler is treated by a series of air pollution control devices to reduce the levels of pollutants. First, a high efficiency cyclone separator removes most of the dust particles. Second, a spray dryer/absorber injects a lime slurry into the flue gas to remove most of the sulfur dioxide and reduce the level of acid gases. Last, a multi-compartment fabric filter, or baghouse, captures the calcium sulfate particles formed in the spray dryer, as well as unreacted lime and small particles

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–1

not initially collected by the cyclone collector. After treatment in the baghouse, the cleaned flue gas from the two combustion units is vented through a common stack at a height of 244 feet above ground.Figure 1.1. Topographic map of the vicinity of the Maine Energy facility.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–2

Figure 1.2

Aerial photograph of the study area in the immediate vicinity of the Maine energy facility. The orange circle is drawn at a 1 km distance (radius) from the facility’s boiler stack. Orthophotograph from Maine GIS.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–3

Figure 1.3

Aerial photograph of the lands around the Maine Energy facility (located at the center of the 1 km radius circle) that depicts the locations of the Southern Maine Medical Center (red cross) and local schools and educational institutions.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–4

In 2000 the Maine Energy facility installed an odor control and scrubbing system to reduce odors that could otherwise escape the facility’s buildings to the outdoors. The system uses fans to draw air into the building (and conversely to prevent the escape of air through doors and windows). Air is collected from the areas in which MSW is stored and processed. A portion of the air is used for combustion by the boilers, and the remainder is treated by filtering through a two-stage particulate removal system, and through activated carbon. Air from the Boiler Building, which is a mixture of ambient air, air from the boiler building, and treated air from the other sections of the odor control system, is treated through scrubbing with a water mist to reduce odors and pollutant levels. The treated air is released through three stacks located on the roof of the boiler building. The release points of the odor scrubbing stacks are 120 feet above the ground, a height roughly one-half that of the boiler stack.

1.2

Risk assessment methods and study area characteristics

The goal of a multi-pathway risk assessment is to determine whether the emissions from a particular facility pose significant risks to public health or the environment. The Maine Department of Human Services, Bureau of Health has historically defined significant risk as an excess (incremental) cancer risk estimate of 10–5 (1 in 100,000). This is a relatively conservative limit compared with the background cancer incidence (50,000 in 100,000 for men, and 33,000 in 100,000 for women).1 Risk assessment methodologies focus on protecting the health of all people. Reference doses (RfDs) and concentrations (RfCs) for compounds of potential concern (COPCs) are derived from toxicologic data with specific consideration of individuals that might be susceptible to adverse health effects at lower levels of exposure compared with the general population. RfDs and RfCs represent levels of exposure that are believed to be safe for all members of the public, including children, the elderly, pregnant women, asthmatics, and other groups of people potentially more sensitive to exposure to COPCs. Figure 1.3 depicts the locations of schools, hospitals, and other locations of special interest in the near vicinity of the Maine Energy facility. A multi-pathway risk assessment for a combustion facility (such as that described herein) focuses only on the risks due to compounds emitted from the facility, and does not consider compounds already present in the environment for other reasons (e.g., due to natural background or emissions from other anthropogenic sources). As such, the risk assessment addresses only the incremental risk due to emissions from the particular facility under evaluation, not the

1

In the U.S., men have a 1 in 2 chance of developing cancer in their lifetime, and women have a 1 in 3 chance (American Cancer Society, 1996). It should also be noted that regulatory risk thresholds (such as a 1 in 100,000 incremental cancer risk allowable for a combustion facility) are based on projected, or modeled, risks of contracting cancer that tend to be estimated in a manner that is believed to overpredict actual risk. These factors should be kept in mind when comparing to background cancer incidence rates, which are measured, actuarial risks. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–5

cumulative risk due to all sources of environmental exposure to compounds (i.e., the facility combined with background and all other sources). The first analyses of waste-to-energy facilities considered only the hazards of airborne COPCs. Subsequent efforts demonstrated the potential importance of indirectly contacting COPCs through more complex routes that, for example, involve deposition to the ground and incorporation into the food chain. Such considerations have identified a variety of potential exposure pathways that trace the movement of COPCs in the environment and their availability to humans. Today, a risk assessment addresses a wide range of pathways by which humans might be exposed to COPCs originating at the Maine Energy facility. Figure 1.4 is a conceptual representation of the multi-pathway exposure assessment. Fate and transport models utilize mathematical algorithms to predict the travel of COPCs emitted from the Maine Energy facility. COPC emission rates are developed from stack test reports from the facility. Air dispersion models combine information about (1) the plant (such as the height of the stack and the properties of the flue gas), (2) the terrain of environs surrounding the plant, and (3) hourly measurements of meteorological parameters, to predict the dispersion of COPCs in the atmosphere. A subsequent algorithm predicts the rates at which airborne COPCs are deposited to soil, water, and vegetative surfaces. Upon deposition, relevant physical and chemical processes are modeled in order to predict the behavior of COPCs in each of these media. Additional models predict the transfer and accumulation of COPCs in locally-produced vegetables, meats, fish, and dairy products.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–6

Figure 1.4

Conceptual model of a multi-pathway risk assessment. Adapted From the U.S. EPA’s Mercury Report to Congress (1997).

The arrows between the different environmental compartments indicate the pathways through which COPCs are assumed to travel and reach humans. Some relationships are fairly simple; for example, COPC concentrations in air are estimated in a straightforward manner by modeling the dispersion of emissions from the facility stack. Other routes are much more complex and require the pursuit of COPCs through several (sometimes connected) environmental media. Consider, for example, the estimation of COPC concentrations in cow’s milk that is produced within the area affected by the facility. Empirical measurements from other studies can be used to estimate the concentrations of COPCs in milk that will result from a cow’s exposure to COPCs in her environment. In this case, the cow’s intake of COPCs is derived from eating food and (incidentally) soil. As indicated in Figure 1.4, COPC concentrations in soil accumulate by deposition of COPCs from air. The cow’s food supply (vegetation) can also become contaminated through two mechanisms: airborne COPCs can deposit to (or are absorbed by) the

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–7

surface of vegetation, and plants can translocate COPCs from soil (which, again, receives COPC deposition from the air). Clearly, modeling impacts through the food chain is multi-dimensional. Figure 1.4 illustrates the three mechanisms through which humans are exposed to COPCs — inhalation of vapors and particles, ingestion (both purposeful and incidental) of a variety of media, and dermal contact with soil. Incorporating COPC uptake by these three routes allows the estimation of a total intake for each COPC.

1.3

The concept of a ‘most exposed individual’

Within this assessment we attempt to conservatively estimate worst-case risks — that is, some overestimates of adverse effects that might result from exposure to chemicals that may be released from the stacks of the Maine Energy facility. In order to do so, we construct reasonable worst-case exposure scenarios. In this context, ‘reasonable’ does not imply average or expected exposure, but rather indicates something plausible, even though not probable. We are well aware that notions of plausibility are subjective and debatable; we have tried to be prudent, perhaps overly so, even as we have tried to be sensible. It is certainly possible to be more extreme in some assumptions; one can make an overestimate arbitrarily large, just by concatenating more layers of conservatism. We believe, though, that concatenations of implausible scenarios yield estimates that are not so much conservative as they are useless. Overall, we base the risk estimates on a higher degree of exposure to plant-related COPCs than is likely to occur from actual operations of the Maine Energy facility. Many of the methods and assumptions we adopt are typical of risk assessments conducted for other waste-to-energy facilities. As examples, exposures are evaluated at the locations of highest impact (as predicted by the air dispersion and deposition studies), and individuals are assumed to live in this impacted area for 40 years, and to partake heartily in activities that lead to exposure to facility-related COPCs (for example, to raise essentially all of their own farm and dairy products – including the grain used to feed livestock – and, moreover, to do so for this period of up to 40 years). Most models and assumptions derive from published guidance (e.g., U.S. EPA, 1989; Maine DEP & DHS, 1994; and NYSDOH, 1991), and thus are typical of human health risk assessments. To the degree justifiable, we have tailored the risk assessment to the environs of the Maine Energy facility, and in particular to the area in which the projected effects from the plant are the greatest.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–8

1.4

Uncertainty and conservatism

Two elements pervading risk estimates are uncertainty and conservatism. In the context of risk assessment, uncertainties are the inevitable byproduct of abstracting complicated current and future real-world phenomena into mathematical approximations, and conservatism is a method of compensating for these uncertainties. For example, variations in the composition of trash and some variability in the day-to-day operations of the waste-to-energy plant cause COPC emission rates to fluctuate. To address this uncertainty, measurements from a number of stack tests are averaged to provide best estimates of long-term emission rates. A high degree of conservatism is incorporated by the assumption that the facility runs continuously with no down time. Uncertainties also accompany fate and transport modeling. The models used here are necessarily simple in nature, including only the essential processes that influence the environmental destiny of COPCs, because no set of models can capture the full complexities of the physical environment. Fortunately, simple models can be very reliable. For example, one of the models used to predict air dispersion incorporates empirically derived estimates of the time-averaged rate at which the plume widens as it travels from the stack. The parameters that characterize this dispersion are estimated from hourly meteorological measurements of atmospheric stability, wind speed, and wind direction. The model also accounts for the effects of terrain and potential aerodynamic interferences of buildings in the vicinity of the stack. The mathematical modeling greatly simplifies the actual, turbulent atmospheric processes, but produces estimates of long-term average air concentrations that can match observations reasonably closely over most conditions. This degree of accuracy is quite adequate. Perhaps the greatest uncertainty lies in the models used to predict the toxicologic potencies (especially the carcinogenic potencies) of the COPCs of interest. In order to gauge whether a chemical is a human carcinogen, groups of laboratory rodents are exposed, typically for most or all of their lifetimes, to very large doses of the chemical. If the doses induce an increased incidence of any type of cancer, compared to the rate observed in unexposed control animals, then the chemical is deemed a carcinogen. Two or more of such tests with positive results suffice to label the chemical a ‘probable human carcinogen,’ even if no actual or useful data from exposed humans are available. This qualitative designation of carcinogenicity is, in many cases, entirely appropriate. Rats, mice, and humans are all mammals that develop cancer from a variety of exposures, and while there are abundant differences among the three species, these differences are not so large as to suggest that chemicals carcinogenic to one species will not be carcinogenic to others. But while the qualitative extrapolation from rodents to humans may be reasonably straightforward, the quantitative extrapolation required for risk assessment is highly uncertain. This is because the doses at which the rodents are tested are typically many thousands of times larger than doses experienced by humans. The central question is, are carcinogenic responses always proportional to dose, such that even at extremely low levels of exposure there is some risk of cancer, and that risk becomes zero only at zero dose? The answer is largely unknown. Knowledge of how

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–9

specific chemicals cause cancer may by helpful on a case-by-case basis, but such information is in most cases still too rudimentary to drive regulatory decision-making. Uncertainties in risk assessment can lie in both directions, since any fate, transport, exposure, or toxicity model can either over-predict or under-predict the variable of interest. In general, risk estimates are conservatively biased — that is, models and parameters are selected intentionally in a manner that tends to increase risk estimates. With regard to cancer, it is assumed that all rodent carcinogens are also human carcinogens, and that all chemicals carcinogenic at high doses are also carcinogenic at vanishingly small doses. Clearly, these are conservative assumptions: if, in actuality, a minuscule dose of a substance poses no risk of cancer to humans, then our assumption of non-zero risk at these doses is an infinite overestimate. Not every assumption in the risk assessment is chosen in a conservative manner (for example, body weight, used in the calculation of doses, is assigned as an average value), but, overall, the bottom-line estimates of risk err on the high side.

1.5

The meaning of risk estimates

The results of the risk assessment are expressed as numbers that represent quantitative estimates of risk. As discussed previously, we believe that the methods employed almost certainly overestimate the actual risks that would result from the operation of the Maine Energy facility. The risk estimates derived herein are intentional overestimates and are not actual risks. The incremental risk of cancer of 1–3 in a million derived for the MEIs does not guarantee that the real, most-exposed individuals will incur this additional risk from exposure to emissions from the Maine Energy facility. Because the risk assessment is constructed with a conservative (health protective) bias, the actual excess risk (of cancer) will almost certainly be lower, and could even be zero. On a different but related issue, we emphasize that an estimated cancer risk of 3 in a million for the recreational farmer does not imply that 3 additional cancers will occur in a population of one million. First, as stated above, the risk of cancer is almost certainly overstated, and may be as low as zero. Second, and more importantly, the risk estimate is derived for a theoretical construct — the most exposed individual. The exposure profile designed for the recreational farmer is not intended, by definition, to fit the general population. The recreational farmer is assumed to live at the location that the EPA models project to be most impacted by facility emissions and consume generous amounts of vegetables, beef, milk, eggs, chicken, and pork (all raised from feed grown at this most-impacted location) for a period of forty years. The individual risk estimate for the MEI applies only to a very limited population — perhaps only to one or two real people; more likely to no one real person. Risk estimates to the general population surrounding the plant — which would consider geographic differences in plant impacts and demographic differences in exposure patterns — are not herein presented. From previous analyses of waste-to-energy plants, we know that once population-weighted risks are assessed, they are considerably lower than the estimates derived for MEIs.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–10

A final consideration requires comment. Cancer from all causes is quite common: about one-third of all Americans develop some form of cancer in their lifetimes, and about one-fourth of all Americans die of cancer. The background risk of cancer death, then, for the MEI (as for all of us) is about 25% (or 250,000 in a million). This overall cancer risk applies to the conceptual MEI irrespective of the existence of the Maine Energy facility. If the estimated incremental cancer risk to the adult MEI is, at most, 4 in one million, the overall cancer risk would increased only slightly to 25.0004% (or 250,004 in a million) — an imperceptible increase.

1.6

Risk assessment basis and organization

Much of the risk assessment for the Maine Energy facility is based on the U.S. EPA’s draft Human Health Risk Assessment Protocol for Hazardous Waste Combustion Facilities (hereafter HHRAP, U.S. EPA, 1998a). The HHRAP is quite detailed and builds upon previous U.S. EPA guidance. The HHRAP is currently available in draft form, including an addendum document issued to correct errors and omissions (U.S. EPA 1999). The HHRAP is similar to methodology used to develop the 1996 risk assessment for the Maine Energy facility, although it has added important enhancements (e.g., wet deposition whereby compounds are removed from the atmosphere by precipitation and deposited to land and water) and refined numerous assumptions and parameters. The HHRAP is used as the framework for the updated Maine Energy facility risk assessment because it is consistent with contemporary risk assessment guidance and sound scientific knowledge. Comprehensive stack testing serves as the primary source of information on compound emissions from the Maine Energy facility. The multi-pathway risk assessment predicts, through the use of modeling, compound properties, and site-specific information, the disposition of compounds in the environment, and hence estimates how they may be contacted by people and animals, and whether such contact presents significant risks to health. A number of refinements to the HHRAP guidance are included in the Risk Assessment Protocol (RAP, included here as Appendix I), and additional improvements are described in the body of the risk assessment report. As stated above, the overall goal of these refinements is to achieve better consistency with scientific knowledge. Peer-review of the risk assessment protocol was also incorporated to provide a comprehensive assessment of potential risks to both human health and the environment. This risk assessment report follows the HHRAP’s suggested outline and content. Subsequent sections describe elements of the risk assessment as discussed in the HHRAP and (as appropriate) in the context of conditions specific to the Maine Energy facility. The report is organized into a series of chapters that describe the sequential steps of the multi-pathway risk assessment, with each step built upon those that preceded. Emissions of compounds of potential concern (COPCs) from the Maine Energy facility are described and quantified in Chapter 2. Chapter 3 describes the detailed modeling study designed to estimate the levels of COPCs in air and in wet and dry deposition that result from emissions from sources at the Maine Energy Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–11

facility. Chapters 4 through 6 describe the procedures used to estimate the levels of COPCs that could be contacted by people in their environment and diet, focusing on categories of individuals likely to receive the highest levels of exposure. Potential risks to human health that could result from such exposure are estimated in Chapter 7, focusing (as is traditional in risk assessments) on the incremental chance that exposure to the facility’s emissions might lead to the development of cancer or other adverse health effects. Uncertainties of the human health risk assessment are discussed in Chapter 8. Finally, the technical appendices contain the detailed information needed to reproduce the calculations of the risk assessment.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

1–12

2

Facility emissions characterization

The Maine Energy Recovery Company’s Biddeford facility (Maine Energy facility) has two sources that emit chemicals of potential concern (COPCs) to the atmosphere. The main source is the boiler stack effluent associated with the combustion of refuse-derived fuel. The second source is the effluent of the odor control and treatment system. Both of these sources utilize control technologies that reduce the levels of COPC emissions to the atmosphere.

2.1

Facility process information

The Maine Energy facility is designed to process approximately 1200 tons of municipal solid waste (MSW) each day. Mechanical equipment is used to separate metal and other noncombustible materials that are recycled or disposed. The remainder of the MSW, which includes primarily paper, plastic, and food waste, constitutes the refuse-derived fuel that the facility combusts in two boilers. The steam produced by the boilers is fed through turbines to produce up to 22 megawatts of electricity. The byproducts of MSW combustion include residual ash, which is trucked to a landfill, and flue gases that are emitted to the air and subject to regulations enacted and enforced by the Maine Department of Environmental Protection (DEP). Prior to atmospheric release, flue gas from each boiler is treated by a series of air pollution control devices to reduce the levels of pollutants. First, a high efficiency cyclone separator removes most of the dust particles. Second, a spray dryer/absorber injects a lime slurry into the flue gas to remove most of the sulfur dioxide and reduce the level of acid gases. Last, a multi-compartment fabric filter, or baghouse, captures the calcium sulfate particles formed in the spray dryer, as well as unreacted lime and small particles not initially collected by the cyclone collector. After treatment in the baghouse, the cleaned flue gas from the two combustion units is vented through a common boiler stack at a height of 244 feet above ground. In 2000 the Maine Energy facility installed an odor control and scrubbing

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–1

system to reduce odors that could otherwise escape to the outdoors, as described in Section 1.1.

2.2

Compounds of Potential Concern (COPCs)

The main pollutants emitted from the Maine Energy facility are particles (more commonly called particulate matter), sulfur dioxide, oxides of nitrogen, total (unspeciated) volatile organic compounds, and carbon monoxide. These pollutants are known as criteria pollutants, and are regulated by DEP to ensure that emissions from the Maine Energy facility do not lead to exceedances of the National Ambient Air Quality Standards promulgated by the U.S. Environmental Protection Agency (EPA) (and adopted by the Maine Legislature as Maine Ambient Air Quality Standards) to protect human health. In fact, the Maine Energy facility is subject to continuous monitoring requirements for three of these five criteria pollutants. Since criteria pollutants are regulated by the DEP and the U.S. EPA to protect human health, the risk assessment focuses on other pollutants released by the Maine Energy facility. Collectively, these pollutants are sometimes called air toxics, and many are designated as Hazardous Air Pollutants in the context of the Clean Air Act regulations (including specific volatile organic compounds). Air toxics tend to be released in much smaller quantities and are hence not amenable to the continuous emission methods developed for criteria pollutants. Instead, air toxics are typically measured on a periodic basis in stack tests using methods developed and specified by the U.S. EPA. The list of air toxics is conceptually infinite, but through research and study regulatory agencies have developed a knowledge base of the pollutants released by waste-to-energy facilities (U.S. EPA, 1993). The 1996 health risk assessment for the Maine Energy facility focused on a select number of compounds of potential concern (COPCs) known to be released from waste-to-energy facilities. Only boiler stack emissions were considered, and the 1996 risk assessment focused on • • • • • • • •

arsenic; beryllium; cadmium; chromium; lead; mercury; nickel; and polychlorinated dioxins and furans (PCDD/PCDFs).

All of these same COPCs are evaluated in this updated risk assessment of the Maine Energy facility. In addition, however, the following additional COPCs are also evaluated for boiler stack emissions based on information gathered in recent testing of boiler stack emissions: • •

hydrogen chloride; six additional metals (copper, selenium, silver, tin, vanadium, and zinc);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–2



ten semi-volatile organic compounds (phenol, naphthalene, 2-methylnaphthalene, diethyl phthalate, di-n-butyl phthalate, bis(2-ethylhexyl)phthalate, benzyl alcohol, 2methylphenol, 3&4-methylphenol, and benzoic acid); and polychlorinated biphenyls (PCBs1, created as products of incomplete combustion).



The second source of emissions is the odor scrubbing system, which differs fundamentally in character since it is not a combustion source. The principal COPCs from the odor scrubbing system are various organic compounds that have been identified in source testing. The specific compounds detected in two monitoring studies of the odor scrubbing system include: • •

ethanol (the organic compound detected consistently at the highest concentration); acetone, benzene, bromomethane, 2-butanone (methyl ethyl ketone), carbon disulfide, chloromethane, chloroform, cyclohexane, 1,4-dichlorobenzene, ethylbenzene, freon 11, freon 12, heptane, hexane, methylene chloride, 2-propanol, styrene, tetrachloroethylene, toluene, 1,1,1-trichloroethane, 1,2,4-trimethylbenzene, vinyl chloride, and xylenes.

Not all of the organic compounds identified in effluent samples of the odor scrubbing system were found in all samples. Some, in fact, were detected in only a few samples. All of the chemicals detected in any sample, however, are considered in the risk assessment. Additionally, since combustion air for the boilers is derived from within the Maine Energy facility buildings, it presumably enters the boilers with some of the same COPCs contained in the odor handling system inlet. The majority of these hydrocarbon compounds are likely destroyed in the combustion of the refuse-derived fuel. However, a small percentage might escape destruction and hence be released from the boiler stack. Therefore, compounds that were detected in the scrubber inlet testing but which were not measured as part of the stack testing are included in the stack emissions, with an assumed destruction efficiency of 99.9%. Finally, because the odor scrubbing system collects air from all the enclosed portions of the facility, there is the potential that fugitive dusts generated within the plant may be emitted from the scrubber. To evaluate the possible impacts of these emissions all of the particulate phase metals and PCDD/PCDFs that are assessed as part of the stack emissions will also be included as part of the scrubber system emissions.

1

PCBs in the Maine Energy stack emission were measured as Aroclor 1248, the HHRAP guidance recommends that the fate and transport of PCB mixtures with greater than 0.5% congeners of more that 4 chlorines be modeled using the properties of Aroclor 1254. Because Aroclor 1248 contains approximately 75% congeners of more that 4 chlorines, PCBs emissions from the Maine Energy facility are modeled as Aroclor 1254. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–3

2.3

COPC emission rates

COPC emission rates are derived directly from the results of recent testing conducted at the facility. Data are considered from the last three years of boiler stack testing (2002 to 2004) and detailed air toxics testing studies conducted in 2002 and 2003 that evaluated emissions from the odor scrubbing system (the 2002 air toxics study also included boiler stack testing). Test results from both the boiler stack and odor scrubbing system are reported as concentrations present in the flue gas or effluent. These concentrations (in units of mass per unit volume) are multiplied by effluent discharge rates (in units of volume per unit time) to estimate mass emission rates (in units of mass per unit time) used as input to the air dispersion modeling analysis. Three sets of emission rates are considered in the risk assessment; (1) baseline long-term emission rates calculated as the average of available recent emissions data for each COPC, (2) maximum or high-end long-term emission rates calculated as the maximum of available recent emissions data, and (3) upset condition, short-term emission rates calculated as the maximum of available recent emissions data multiplied by a process upset factor. The measurements and calculation used to determine these emission rates for the purposes of risk assessment modeling are described in the sections that follow. The baseline risk assessment is developed using best estimates of emission rates based on full facility operation, calculated with the average values of measured COPC concentrations and operating conditions typical of full facility capacity for effluent flow rates. The use of average measured COPC emission rates for the baseline risk assessment actually results in an overestimate of the long-term facility impacts because the air dispersion modeling assumes that these COPC emission rates occur without interruption while each of the combustion units at the Maine Energy facility does in fact experience a significant amount of downtime over the course of a year. Data on the extent of the facility downtime are described below in relation to the assessment of facility upset conditions on COPC impacts. To test the sensitivity of the risk assessment results to uncertainties in the measured emission rates, risk estimates will also be performed using high-end estimates of the COPC emission rates. Following HHRAP guidance, these emissions estimates are calculated based on the lesser of (1) the maximum COPC concentrations detected in sampling and (2) the average concentration plus two standard deviations of the average, and also using continuous operation of the facility at the designed maximum capacity. Because the COPC emissions testing was conducted under normal facility operating conditions, there is the possibility that short-term, off-normal facility operations might lead to higher COPC impacts than would be predicted from the model using even the maximum measured COPC emission rates. Therefore, potential short-term elevated COPC emission rates have been estimated for several process upset conditions based on facility records of the occurrences of such upsets and continuously measured criteria pollutant measurements. Facility upset and downtime data are also used to assess the degree to which the modeling assumption that the facility operates continuously at its tested COPC emission rates produces an overestimate of COPC long-term impacts. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–4

In addition to the direct use of measured data for estimating COPC emissions from the Maine Energy facility, some specific COPCs require the application of additional data or assumptions to provide emission rates needed for the multi-pathway modeling. The first set of assumptions concerns the methods used to estimate emission rates of COPCs that were not present in the tested exhaust gasses at levels high enough to be reliably measured. The second and third set of assumptions concern the chemical and/or physical forms of the COPCs chromium and mercury. Chromium is assumed to be present in the facility’s emissions in either the trivalent or hexavalent forms (notated as Cr+3 and Cr+6 respectively), and mercury is assumed to be present as either elemental vapor-phase mercury, divalent vapor-phase mercury, or divalent particulatephase mercury (notated as Hg0, Hg2+(v), and Hg2+(p) respectively). Because the different forms of these metals have significantly different transport and/or toxicological properties, but they are not distinguished from each other in standard exhaust gas testing, additional information is necessary to allow for the assessment of their impacts in the multi-pathway modeling. The following sections describe the methods used to derive the COPC emission rates that have been included in this report’s direct and multi-pathway risk estimates. Detailed data from the facility’s emission testing programs, process upset logs, and continuous emission monitors that was used to calculate the emissions rates used in the risk assessment are included in Appendix III. The COPC emission rates that are used in the risk assessment models are summarized in Tables 2.2 and 2.3 which follow the descriptions of the methods used to derive these rates. The baseline emission rates of COPCs are similar to those considered in the 1996 health risk assessment, as described in Section 2.3.7. Additional details of the derivations are contained in the sections that follow; measured concentrations are contained in Appendix III.

2.3.1 Procedures for estimating stack COPC emission The modeled COPC emission rates from the Maine Energy facility’s boiler stacks are based on the average of the measured stack gas and exhaust flow rate data for each COPC as collected during the stack testing programs since 2002. Because several different types of COPCs are included in the modeling, and different rounds of stack testing have been conducted for different purposes over the past few years, the sources of COPC emission rate data differ from one compound to another. Additionally, some COPCs may be present in minor amounts in stack exhaust gases, but have never been part of any of the stack gas testing programs. The modeled stack emission rates for these compounds are based on concentrations that have been measured in other parts of the Maine Energy facility. The most complete set of recent stack gas test results are contained in the Maine Energy 2002 Air Toxics Test Program report (Eastmount, 2002). Full details of this testing program are found in the test program report — test procedures and conditions are summarized here. Stack testing for this report included the measurement of trace metals, hydrogen chloride, volatile organic compounds (VOCs), and PCDD/PCDFs in the boiler stack gases. Sampling of these gases was conducted during the week of July 29, 2002 and August 5, 2002. All pre-test preparation, testing, analysis, and calculations were conducted in accordance with procedures approved by the Maine DEP; the Code of Federal Regulations Chapter 40, section 60, Appendix A; the U.S. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–5

EPA Quality Assurance Handbook (Vol.III), and a pre-test protocol. Trace metals were measured using EPA Reference Method 29; hydrogen chloride was measured using EPA Reference Method 26/26A; VOCs were measured using EPA Reference Method 0040; and PCDD/PCDFs, PCBs (reported as Aroclor 1248), and semi-volatile organic compounds were measured using EPA Reference Method 23/0010. During the test program both of the facility’s boiler units were operated at or near their rated capacity. Two subsequent rounds of sampling were performed in July 2003 and August 2004 to measure emission rates of selected metals (As, Cd, Cr, Pb, Hg, and Ni in 2003; and Cd, Pb, and Hg in 2004) and hydrogen chloride. In addition to the 2002 Air Toxics testing, for PCDD/PCDFs concentrations were measured in September 2001, July 2003, and August 2004. The averages of the recent test results for each COPC were used to calculate emission rates for the risk assessment. Comparisons among test results for COPCs that have recently been measured more than once, and those which were included in the 1996 risk assessment, are included in section 2.3.8. The VOCs that are included as COPCs due to their presence at the inlet to the scrubber system but which were not measured as part the stack emissions test programs are: n-butanol, cyclohexane, 1,2-dichlorobenzene, 1,3-dichlorobenzene, 1,4-dichlorobenzene, ethanol, ethylbenzene, freon 11, freon 12, heptane, hexane, methane, methanol, propane, 2-propanol, 1,1,1-trichloroethane, and 1,2,4-trimethylbenzene. Because the combustion air for the Maine Energy boilers is taken from within the facility’s buildings, these compounds might be emitted from the boiler stacks if they were not destroyed in the combustion process. Therefore, the estimated stack emission rates for these compounds were calculated by multiplying their concentrations at the scrubber system inlet by the typical facility combustion air flow rate of 55,000 cubic feet per minute, and a factor of 0.001 which corresponds to a combustion system destruction and removal efficiency (DRE) of 99.9% for these compounds. This is a fairly conservative (i.e., probable underestimated) DRE for estimating organic COPC emissions from the boiler stack. For example, in a comparable context from a different industry, the U.S. EPA’s Resource Conservation and Recovery Act (RCRA) hazardous waste regulations require that boilers and industrial furnaces achieve a DRE of 99.99% for the hazardous organic constituents of the waste, and a DRE of 99.9999% for dioxin-bearing wastes (40 CFR Part 266, Subpart H).

2.3.2 Procedures for estimating odor scrubber system COPC emission rates The Maine Energy facility odor scrubber system is designed to collect gases that are present in the facility’s waste handling, processing, and combustion buildings, and to pass these gases through a series of filters and wet scrubbers before emitting them from three venting stacks on the top of the building. The flow rate through the system is high enough to prevent odorous gases from exiting the facility’s buildings through any open doorways, thus preventing emissions of untreated odorous compounds and fugitive at ground level.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–6

The gases at the inlet and outlet of the scrubber system were sampled and analyzed for VOCs using EPA Compendium Method TO-15 in August 2003. Details of the sampling program are contained in the report “Emission testing from three scrubbers at MERC” (APCC, 2003). Two sets of samples were collected and two flow rate measurement were taken at the inlet and outlet of each of the three scrubber units. The scrubber outlet COPC concentrations were multiplied by the flow rates for each of the six tests, the two results for each scrubber were averaged, and the averages were summed to produce overall average emission scrubber system COPC emission rates. Maximum total scrubber system COPC emission rates were calculated by summing the greater of the measured emission rates for each scrubber unit. A few COPCs that were not measured as part of the scrubber system sampling program were measured in three rounds of tests performed in August and September 2004 on the gases present in the tipping room of the facility. These data were used to supplement the scrubber system data by multiplying the tipping floor concentrations by the total scrubber system outlet flow. Average emission rates were calculated using average concentrations and flow; maximum emission rates were calculated using maximum concentrations and flow. This is a highly conservative estimate of the emission rates for these COPCs from the scrubber system vents because it does not account for any concentration reductions that might occur due to the action of the scrubber system or the significant dilution that occurs as the gases from the tipping room are mixed with additional air while being transported between the buildings. In addition to the gases that are collected and processed by the odor scrubber system, fugitive dusts that are present within the facility’s buildings may also be collected, filtered and emitted by the system. The fugitive dust within the facility that has the highest particulate-phase COPC content and the greatest potential for fugitive dust emissions is the ash collected from the combustors. This material has significantly higher particulate-phase COPC concentrations than the waste entering the facility and combustors because the metals and low-volatility compounds present in the waste and combustion products are less diluted by combustible and volatile compounds initially present in the waste. The ash has a higher tendency to be collected and emitted by the scrubber system because of the ashes elevated fraction of very fine particles which are more likely to be suspended as fugitive dust within the facility than are larger particles. An estimate of the fugitive ash emission rate from the facility was performed by Maine Energy in February 2003 and a report on this estimate submitted to the City of Biddeford in the Maine Energy Air Toxic Control Application. The fugitive ash report indicates that 46,322.32 tons of ash were produced in 2003. The report then applies an emission factor of 1.5 pounds of emissions per ton of material loaded into an open truck based on data from the U.S. EPA Emission Factors Handbook (AP42, Table 11.17.4, U.S. EPA, 1998c). An emission control efficiency of 80% is applied to account for the fact that the wetting of the ash reduces its tendency to be emitted as a fugitive dust. A second control efficiency of 90% is applied to account for the fact that the fugitive dusts are generated in an enclosed building with the only emissions having to pass through the collection and treatment system. It is believed that both of

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–7

these factors underestimate the true control efficiencies. Taken together these data yield an average ash emission rate of 1390 pounds per year or 0.020 g/s. These fugitive ash emission estimates were combined with COPC concentrations measured in the ash as part of the 2002-2004 MERC Annual Solid Waste Reports. The COPCs included in this report are: arsenic, cadmium chromium (total, and hexavalent), copper, lead, mercury, nickel, selenium, silver, vanadium, zinc, and PCDD/PCDFs. Concentrations of the COPCs beryllium and tin in the ash were not included in the analyses for solid waste reports, therefore these concentrations were estimated based on the ratios of these metals’ concentration to those of other metals as measured in the boiler stack emissions. Beryllium levels in the ash were estimated based on the chromium level in the ash and the ratio of beryllium to chromium in the stack emissions because both of these metals are classified as low volatility metals for the purposes of MACT compliance. Tin levels in the ash were estimated based on the lead level in the ash and the ratio of tin to lead in the stack emissions based on their relatively similar melting points, sources, and chemistries. Because the metal emission from the odor control system are very small relative to the boiler stack, these approximations have very little effect on the overall results of the risk assessment. The COPC ash concentrations (in units of mass per mass, ng/kg in the report) were multiplied by the estimated ash emission rate to obtain an estimate of the COPC emission rates.

2.3.3 Data and procedures for estimating stack COPC emission rates under process upset conditions A waste-to-energy plant does not always operate under normal conditions: the plant goes through startups and shutdowns, and may experience upset conditions if a portion of the air pollution control system malfunctions or the combustion control system is disturbed. Emissions that occur during these periods will be evaluated in the risk assessment only with respect to the potential impacts of direct short-term exposures. Emissions at the increased upset condition rates are short-lived because once the upset is detected, waste feed to the combustion unit experiencing the upset is automatically stopped (per facility design), or other actions are taken to eliminate the cause of the upset. For example, waste feed cutoff is required when the opacity of the visible emission (as monitored by the facility’s continuous emission monitor) is equal to or greater than 15%. The possible effect of upset emissions on long-term average emission rates and exposures is evaluated in Section 2.3.3.5. The potentially higher than normal facility emissions during system upsets are evaluated using upset factors developed for specific types of upset conditions, which are applied to the COPCs that would be affected by these conditions. To the extent possible, the upset factors used in the risk assessment models are based on site-specific information. This has been done by matching logged information on facility upsets with data from the facility’s continuous emission monitoring systems (CEMS) which log hour measurements of opacity, carbon monoxide (CO), and sulfur dioxide (SO2), among other pollutants. The facility data used to calculate the upset factors are included in Appendix III. Because the upset factors are employed to assess maximum

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–8

potential one-hour COPC exposures, the factors are based on one-hour increases in emissions. Because the highest potential short-term emission rates would occur when the facility is running at its maximum operating conditions, maximum non-upset emission rates are multiplied by the upset factors to estimate the maximum one-hour upset emission rates. By combining the estimates of the upset factors with maximum facility emission rates and maximum modeled air dispersion impacts, the model is very likely to significantly overpredict actual maximum shortterm impacts. This is because the three events being modeled: (1) operation at maximum initial COPC emission rates, (2) a process or control system upset, and (3) the presence of atmospheric dispersion conditions that lead to the maximum short term impact levels, are all fairly uncommon. It is therefore very unlikely that all three would occur during the same hour, as modeled under the upset condition short-term risk calculations. Upset emissions may be caused by equipment malfunctions in three of the Maine Energy facility’s systems: • • •

upsets to the boilers’ combustion control system that result in incomplete combustion and increased emissions of organic compounds, upsets to the baghouses that result in increased emissions of particulate pollutants, and upsets to the spray dryer system that result in increased acid gas emissions.

Although facility startup and shutdown may not be strictly classified as upset conditions, the short-term effects of operations during startup and shutdown will be considered as part of the upset analyses. While COPC mass emission rates during startup and shutdowns may be lower than during normal plant operation because of reduced throughput, it is possible for the maximum ground level COPC concentrations to be higher than normal due to lower than normal dispersion of the emissions. This condition is evaluated through the use of a special air dispersion modeling run as described in Section 2.3.3.4. To simplify the modeling of upset condition emissions, a composite, worst-case, set of COPC emission rates for the combustion control, baghouse, and spray-dryer upsets has been compiled. This overall set of emission rates uses the highest upset factor for each COPC, thus the COPCspecific risk and hazard estimates for the upset modeling are each valid for the upset condition that leads to the greatest impact for that compound. However, because the modeled upset condition approximates all the upsets occurring simultaneously (a highly unlikely occurrence), the overall risk and hazard estimates for the upset condition modeling are overestimated. The possibility that these upsets might occur during the low stack flow conditions of boiler startup and shutdown is modeled using the composite upset COPC mass emission rates with a special dispersion and deposition modeling run based on low stack flow conditions. The high stack flowrate upset condition is referred to as a “normal upset”; the low stack flowrate upset condition is referred to as a “startup upset.”

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–9

2.3.3.1

Combustion control upsets

Disturbances of the combustion control system may result in incomplete combustion of organic COPCs. Large increases in organic emissions during such upsets are partially controlled by the spray dryer/fabric filter system, which will retain some organics. The combustion control upset factor has been calculated based on Maine Energy records of boiler outages caused by fuel feed problems or boiler malfunctions or operating problems and hourly carbon monoxide concentrations measured before and during the period when the upset occurred. Carbon monoxide levels are used as an indication of incomplete combustion conditions. A total of 22 such upsets occurred between October1, 2003 and September 31, 2004; suitable CO measurements are available for 12 of these episodes (because of the timing of the upsets and the hourly CO measurements definitive CO concentrations before and/or after were not available for all of the events). The ratio of CO concentrations measured during the period when the upset occurred to those measured before the upset ranged from 2.3 to 3.9, with a mean ratio of 2.9. These values are consistent with short-term organic compound emission increases that have been observed at other facilities that have experienced combustion control upsets. For example, during the testing of the Marion County facility’s startup procedures, the overfire air fan failed for a brief time. Total hydrocarbon emissions during this period were elevated by a factor of 3–5 (U.S. EPA, 1988). The mean combustion control upset factor of 2.9 will be applied to assess potential short-term emission rate increases in the facility’s organic COPC impacts.

2.3.3.2

Spray dryer absorber upsets

Increased emissions of hydrogen chloride may result from the malfunction of the spray dryer absorber system. Most malfunctions of this system are short term, requiring less than about 10–15 minutes to address and return to normal operation. In addition, effects of partial or full spray dryer absorber malfunction are mitigated because residual unreacted lime on the fabric filter bags will continue to remove acid gases. Malfunctions of the system are very uncommon; from July 1, 2001 through March 31, 2005, only 3 spray dryer upsets occurred. To estimate the spray dryer upset factor, the sulfur dioxide levels measured before and during two of these upset periods were compared (the other upset period did not have suitable SO2 data available). The ratios of SO2 concentrations during the upset to those before the upset were 2.2 and 2.5. Because it was possible to calculate only two values, the higher of the two, 2.5, will be used to evaluate short-term spray dryer upset emissions. This value is comparable to the one estimated in the RAP (2.1) which was based on a malfunction that leads to a 15-minute, tenfold increase in emissions from one of the facility’s two spray dryers.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–10

2.3.3.3

Baghouse/fabric filter upsets

Although total failure of a combustion unit’s entire fabric filter system is not plausible, rupture of one or more filter bags in the unit’s baghouse system may occur. When this occurs, a portion of the flue gas stream is untreated. Operators quickly isolate the appropriate cell and replace the ruptured bags. If such a rupture results in an increase in opacity to a level equal to or greater than 10% for 15 minutes, feeding of waste into that unit automatically ceases. As with the spray dryer system, significant baghouse system malfunctions are very uncommon; since July 1, 2001 only four such upset have occurred and all of them were between July 18, 2001 and November 14, 2001. Unfortunately, sufficient data are not available to establish a site-specific upset factor for baghouse malfunctions. Therefore, the method suggested in the RAP is applied. It is assumed that the fabric filter bag has a particle mass collection efficiency of 99.5%, thus a total failure of the bag causes the emission rate through that bag to increase by a factor of 200. If 5% of the unit’s flue gas passes through the ruptured bags2 and one hour is required to isolate the cell containing the ruptured bag, the increase in PM emissions from that unit is: 0.95 (flow through intact units) + 200 × 0.05 (flow through ruptured units) = 10.95. Because there are two independent baghouses, the overall facility upset factor for such a filter bag rupture (averaged between the two units) is thus: 1 + 10.95 ≈6 2

2.3.3.4

Combustion startup/shutdown conditions

The conditions that exist during periods of combustor startup and shutdown will be included in the estimation of upset factors. Since startup and shutdown procedures include reduced feedrates and result in lower operating temperatures, the only COPC stack gas concentrations likely to increase are those for volatile organic compounds due to incomplete combustion of the feed material. Therefore, all VOCs will be considered in the startup/shutdown upset scenario. Although VOC concentrations in the boiler exhaust gases may be elevated during startup and shutdown periods relative to during periods of normal operation, the VOC mass emission rates may not increase by a proportional amount due to reduced throughput during these intervals. However, the lower feed rates that occur during startup and shutdown conditions also result in a

2

The estimate that 5% of the baghouse flow would pass through the ruptured bag is based on an analysis by American Ref-Fuel for a waste to energy plant for which Cambridge Environmental has previously performed a multipathway risk assessment (Cambridge Environmental, 1992). Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–11

lower gas flow rate at the stack exit which may lead to less atmospheric dispersion of emitted VOCs and thus potentially higher maximum ground level concentrations. Therefore, to estimate ground level short-term VOC concentrations during startup and shutdown operations, a special air dispersion model run was performed. This model run was based on a facility stack exit velocity of half the normal to account for the lower total gas throughput that occurs during startup. Because this model run was used only to evaluate the maximum one-hour COPC concentrations and direct exposure levels, COPC deposition was not included in the short-term upset condition modeling. The atmospheric dispersion results from this model run were combined with upset condition VOC emission rates to estimate maximum short-term startup/shutdown ambient VOC concentrations. Although CEMS data are available during periods when the facility’s boilers are operating under startup conditions as well as during planned and unplanned shutdowns, the interactions between variations in the fuel feed rates, combustion air flow rates, and exhaust flow rates and temperatures on these measurements preclude a simple estimate of VOC upset factors from the available data. Because the condition being assessed in the startup/shutdown upsets models is one of elevated VOC levels due to incomplete combustion, the VOC emission concentrations for the startup/shutdown conditions are modeled as equal to those used for the combustion control upset, or 2.9 times the maximum measured concentrations.

2.3.3.5

Effects of upset emissions on long-term average

The effects of process upset conditions are not included in the long-term multi-pathway risk assessment calculations because the inclusion of upset condition emission rates would not increase actual long-term emissions above the levels used in these calculations. The long-term average stack COPC emission rates used in the multi-pathway portions of the risk assessment are based on actual stack test results, with the assumption that these emissions occur during every hour of the year. However, each of the Maine Energy boilers experiences a significant amount of downtime or outage periods each year; therefore, the actual long-term average emission rates for normal facility operation are lower than the long-term emission average emission rates used in model. To assess whether the elevated emissions that are believed to occur during periods of upset operation might be large enough to compensate for this over estimation, the facility’s upset and outage records between July 1, 2001 and March 31, 2005 have been examined. For the periods when total operating and outage hours are available for either or both of the facility’s boilers, there are a total of 35,429 boiler-hours of operation and 3997.5 boiler-hours of outage. This corresponds to the facility’s boiler operating during 90% of the available hours. During the same period a total of 107 of system upsets (including system startup/shutdowns) were identified. Even if the maximum calculated system upset factor (a factor of 3.9 for combustion control upsets) were to be applied for each of these occurrences, this would only add an amount of COPC emissions equivalent to 310 hours or normal operation ((3.9 - 1) V 107). These added emissions are far less than those that are assumed to occur during 3997.5 boiler-hours when the facility’s boiler are not operating, and the actual emissions are zero. Therefore, the long-term emission rates used in the multi-pathway portions of the risk assessment are higher than those that would be calculated by taking the facility’s actual long-term average emission rates Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–12

(adjusted for facility downtime) and adding the effects of short-term elevated emission rates caused by upset operations.

2.3.4 Procedures for non-detected compounds As described in the Maine Energy RAP, and in Section 2.2 of this report, the only compounds that are included as COPCs in the risk assessment are those that have been detected in either boiler stack emission gas samples, odor scrubber system inlet or outlet samples, or in facility ash samples. The RAP indicated that an exception would be made for specific PCDD/PCDF congeners, and for hexavalent chromium. However, each of the 17 carcinogenic CDD/PCDF congeners have been detected in at least one of the stack gas samples, and hexavalent chromium was detected in some of the facility ash samples, so the potential exception for non-detects is irrelevant. The estimation of emission rates for COPCs that are detected in some (but not all) of the boiler stack or scrubber emission tests can be a source of uncertainty in a risk assessment. To produce estimates of the emission rates for these compounds, two different methodologies are applied. For those COPCs that have not been detected in the recent stack or scrubber tests (but have been detected in older tests), the assumed baseline emission rates are taken as one-half the detection limit of the most recent testing program. For those COPCs that have been detected in some (but not all) of the most recent tests, the test results in which the COPC is not detected are averaged with the detected results at the full detection limit. The specific detection limit for each COPC and test result that is used in emission rates calculations is dependent on the data that are available for each testing program. The goals for this portion of the risk assessment are to use as much data as is available, maintain a conservative bias (designed to overestimate actual emissions, but not introduce overly conservative methodologies to compensate for uncertainty). The detection limits used to estimate concentrations of non-detected COPCs are the maximum concentration level for which the COPC would not be indicated as a detected compound. Therefore, if method detection limits (MDLs) are given, and COPCs present at levels just above the MDL would be shown as being detected, the MDL value is used as the detection limit. If COPC concentrations are reported in the test results at just above MDLs (perhaps with a qualifier indicating an estimated concentration), the value is used as given (consistent with the treatment of estimated values in the Superfund program). In this case, the use of the HHRAP recommended MDL-derived reliable detection limit (RDL) values for non-detected POCs could result in test results with nondetects being averaged into the emission calculations at higher concentrations than detected tests. Detailed, COPC-specific measurements and the COPC-specific treatments of non-detected values are included in the tables in Appendix III.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–13

2.3.5 Chromium speciation Chromium can exist in two forms in environmental compounds, bonding either in trivalent or hexavalent forms. Because these two forms have significantly different toxicological properties, they should ideally be measured and modeled as two different COPCs in the assessment of direct and multi-pathway potential health risks. However, the measurement of hexavalent chromium in exhaust gases is difficult due to both limitations of the analytical methods and the low concentrations of hexavalent chromium in combustion exhaust gases, so stack testing for hexavalent chromium levels has not been attempted at the Maine Energy facility. Data from other similar facilities are scarce. Because the fraction of chromium present in the hexavalent form can have a significant effect on the overall estimated risks caused by a facility’s emissions, it is necessary to estimate this fraction from other available data. Based on previous risk assessments of municipal waste combustors, theoretical chemical equilibria and kinetics at typical municipal waste combustion conditions, measurements of hexavalent chromium in combustion exhaust gases, and the maximum fraction of hexavalent chromium found in ash from the Maine Energy facility, the anticipated level of hexavalent chromium in the Maine Energy facility’s emissions is believed to be negligible. It should be noted that the 1996 Maine Energy Risk Assessment (Cambridge Environmental, 1996) evaluated chromium in trivalent form only (i.e., it was assumed that there was no hexavalent chromium emitted from the facility). In a recent evaluation of emissions from the Harrisburg, Pennsylvania Waste to Energy facility Jones (2003) conservatively assumed a hexavalent chromium fraction of 2%. This value was based on a risk assessment of the Falls Township Waste to Energy facility submitted to the Pennsylvania Department of Environmental Protection, and test data from the Bridgeport, Connecticut Waste to Energy facility which found no detectable levels of hexavalent chromium, with a conservative estimate of the detection limits of 3–5% of the total chromium level. The hexavalent chromium fraction of 2% was assumed to be an upper confidence limit (percentile unspecified) of the actual mean fraction. The conservative nature of the 2% estimate is supported by the theoretical and laboratory combustion studies of Persson et al. (2000) and Sandelin et al. (2001) which found that the potential and measured fraction of hexavalent chromium in combustion exhaust gases is negligible (less than few percent) for gas temperatures up to approximately 1300°C, well above typical municipal waste combustion temperatures of 1000°C. Although hexavalent chromium measurements have not been performed in the Maine Energy exhaust stack, they have been performed in the facility’s collected combustion ash in eight quarterly measurements of ash collected in 2002 and 2003. Hexavalent chromium was only detected in the four tests conducted in 2002. Using the detection limits for the non-detected hexavalent chromium concentrations, the fraction of hexavalent chromium in the ash is approximately 1% of the total chromium concentration. Therefore, the baseline risk estimates will be performed with an assumed hexavalent chromium fraction of 2% (a conservative estimate based on the facility-specific ash measurements). Sensitivity analyses are described in Section 8.1.5 using other hexavalent chromium fractions of 1, 5, and 10%.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–14

2.3.6 Mercury speciation Because the potential health effects caused by exposures to emitted mercury often dominate the non-cancer portion of the risk assessment, and because the modeling of mercury emissions and transport is very sensitive to the selection of several modeling parameters, the measurement or estimation of the speciation of the mercury emissions (i.e., the distribution of mercury among its various chemical and physical forms) is critical in the multi-pathway modeling of this pollutant. Within the HHRAP guidance and this risk assessment it is assumed that mercury is emitted in three forms: vapor-phase elemental mercury, vapor-phase divalent mercury, and particulatephase divalent mercury. The divalent mercury species are further assumed to all be present as mercuric chloride (HgCl2). It is assumed that a fourth form of mercury, methyl mercury, is not emitted from the facility, but instead is formed in soils and surface waters from the emitted divalent mercury. Methyl mercury is the form that is found in fish tissues, and it is in the consumption of mercury in fish that the majority of the potential human health effects occur. The assumption that all of the divalent mercury is present as mercuric chloride results in an overestimate of the mercury impacts because mercuric chloride is among the most reactive forms of mercury, and it thus deposits from the atmosphere and can transform into methyl mercury far more rapidly than less reactive yet common divalent forms such as mercuric oxide (HgO). Mercury measurements in stack testing at the Maine Energy facility use U.S. EPA Reference Method 29. The components of the sampling train used for Method 29 are shown in Figure 2.2. Although the method is not explicitly designed for determining the speciation of mercury emissions, the U.S. EPA Report to Congress (U.S. EPA, 1997a) notes that the distribution of mercury in the Method 29 sampling train can be used to infer the form and speciation of mercury in MWC stack gas. The probe and filter of the sampling train collect mercury adsorbed onto particulate matter. Vapor-phase ionic mercury compounds that are soluble in water are collected in the nitric acid/hydrogen peroxide impingers, and elemental mercury is collected in the potassium permanganate/sulfuric acid impingers. The details of the most recent mercury test results at the Maine Energy facility are given in Table 2.1, with the mercury speciation fractions assumed for stack emissions in the multi-pathway risk assessment modeling summarized in the final column. Although there are uncertainties in the use of these test results to model mercury speciation in the Maine Energy stack emissions, the speciation values shown in Table 2.1 and used in the risk assessment are more likely to overestimate rather than underestimate the mercury impacts. The form that was found to comprise that majority of the emissions, Hg2+ vapor, is also the form that is deposited most rapidly from the atmosphere. Further, if it were assumed that the amount of mercury in the catches where none was detected was equal to the detection limit (rather than at one-half the detection limit as assumed in Table 2.1), then the fraction of assumed vapor-phase elemental mercury, which has very little impact on modeled health effects, would increase. Therefore any changes in the assumed distribution would likely decrease the estimated mercury impacts of the Maine Energy facility’s emissions, and so the modeled mercury speciation likely provides a conservative estimate of potential mercury related health risks.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

2–15

It should also be noted that the fraction of mercury assumed in divalent vapor form (77%) exceeds the value of the HHRAP default value (60%) suggested for municipal solid waste combustors. Adopting the speciation estimates based on facility-specific sampling thus results in a larger estimated fraction of mercury emissions from the Maine energy facility depositing in the local Biddeford/Saco vicinity. Table 2.1

Results of the August 2004 mercury stack concentration tests at the Maine Energy Facility. The catch is the amount of mercury measured in each stage of EPA Reference Method 29 that is analyzed for mercury. The overall mercury speciation fractions shown in the last column are based on catches with no detected mercury having mercury present at one-half the detection limit.

Sample Probe rinse and filter HNO3/H2O2 impinger Empty impinger (between HNO3 and MnO4 impinger) KMnO4/H2SO4 impinger HCl rinse of KMnO4/H2SO4

Assumed Hg species

Catch (:g)

Hg Particulate Hg2+ Vapor

test 1 0.19 1.26

test 2 0.29 1.37

test 3 0.286 1.25

Hg2+ Vapor

10–4 atm-m3/mol); vapors with moderate Henry’s Law constants; (10–6 < H < 10–4 atm-m3/mol) vapors with low Henry’s Law constants (H < 10–6 atm-m3/mol); vapor-phase ionic mercury.

Each emission type was assigned a nominal emission rate (e.g., a generic emission rate of 10 g/s) so that model predictions could be scaled easily to COPC-specific emission rates. It was necessary in some cases to assign large nominal emission values (e.g., cases with low vapor scavenging rates) so that the models provided output with sufficient numbers of significant figures. Several COPCs such as PCBs, dioxins/furans, and several semi-volatile organic compounds are present in both the particle and vapor phase. The vapor fraction for each COPC is given by the parameter Fv in Appendix II. The modeled dispersion and deposition parameters for for these compounds are the weighted sum of the values for their vapor and particulate phases. Figures 3.9 to 3.20 typify model predictions. Each figure depicts the pattern of modeled concentrations in the vicinity of the Maine Energy facility (located at the center of each figure). Each figure provides two projections: a domain-wide projection that covers the entire 15 km radius study area, and an expanded view of the 6 km by 6 km area closest to the Maine Energy facility. The color-coded shading in each figure indicates the spatial variation of predictions. Representative figures of all of the modeling runs are provided in Appendix VI. A cross at the center of each figure indicates the location of the Maine Energy facility, and the Goosefare Brook and Wilcox Pond watersheds are outlined. Figures 3.9 to 3.14 depict results from the ISCST3 modeling of boiler stack emissions of particulate-phase COPCs that have volume-weighted (rather than surface area-weighted) distributions among the particle size classes. These figures show estimated concentration and deposition levels for a nominal emission rate of 1 g/s. Figure 3.9 shows the maximum 1-hour average concentrations peaking in the near vicinity of the Maine Energy facility, although a distinct “donut-hole” of lower concentrations is predicted immediately adjacent to the facility. The pattern is similar in Figure 3.10 for the maximum 24-hour average predictions, although the “donut-hole” effect is somewhat smaller. The annual average concentration predictions in Figure 3.11 exhibit well-defined maxima over two areas just to the north and to the southeast of the Maine Energy facility. As expected, the magnitude of the predicted concentrations decrease as the averaging period lengthens, with peak 1-hour maxima roughly a factor of thirty greater than the annual average predictions (Figure 3.9 v. Figure 3.11). Figures 3.12 to 3.14 depict modeled annual deposition rates. The pattern of total (wet plus dry) deposition (Figure 3.14) is very similar to the predicted pattern of wet deposition (Figure 3.12) because for these emissions wet deposition predictions are much greater than those of dry deposition (Figure 3.13). Figures 3.15 to 3.20 display the AERMOD model predictions of dispersion and deposition of emissions from the three odor scrubbing stacks. The figures depict the series of runs for low

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3–31

Henry’s Law constant COPCs, based on a nominal emission rate of 300 g/s (100 g/s for each stack). Figure 3.15 shows the maximum 1-hour average concentrations peaking in the near vicinity of the Maine Energy facility, although a distinct “donut-hole” of lower concentrations is predicted immediately adjacent to the facility. The pattern is similar in Figure 3.16 for the maximum 24-hour average predictions, although the “donut-hole” effect is less distinct. The annual average concentration predictions in Figure 3.17 exhibit well-defined maxima over limited areas just to the north and to the southeast of the Maine Energy facility. As expected, the magnitude of the predicted concentrations decrease as the averaging period lengthens, with peak 1-hour maxima roughly a factor of thirty greater than the annual average predictions (Figure 3.15 v. Figure 3.17). Figures 3.18 to 3.20 depict modeled annual deposition rates. The pattern of total (wet plus dry) deposition (Figure 3.20) is very similar to the predicted pattern of dry deposition (Figure 3.19) because dry deposition predictions are much greater than those of wet deposition (Figure 3.18), which are predicted to peak in the immediate vicinity of the Maine Energy facility. This is in contrast to the deposition results for volume-weighted particulate COPC emissions from the facility stacks where the rates for dry deposition (Figure 3.13) are smaller than those for wet deposition (Figure 3.12). Table 3.10 summarizes the set of modeled concentrations identified as maximum values over the modeling domain for the two short-term and one long-term averaging period. One-hour average concentrations for stack emissions were modeled using the normal operating stack flow rate and at a flow rate of one-half normal so that upset emissions that occur during facility startup/shutdown conditions can be evaluated. Table 3.11 summarizes the long-term maximum modeled deposition rates over the domain. Vapor-phase dry deposition is not modeled separately, but is calculated from concentration levels and an assumed dry deposition velocity of 1.4 cm/s. Annual average concentration and deposition values over the watersheds of interest are shown in Table 3.12. Domain-wide deposition and watershed concentration and deposition are only modeled for assessment of long-term effects. These values are used in subsequent chapters along with modeled deposition rates to develop the multi-pathway exposure and risk estimates. Table 3.13 summarizes Annual average concentration and deposition values at the location identified as having the maximum potential health risks for the farming exposure scenario. The farming scenario is evaluated only at locations further that 1 km from the facility (see Chapter 4 for details). To find the maximum impact location that is at least 1 kilometer from the Maine Energy facility, a Microsoft Excel macro was written to calculate the overall potential cancer and non-cancer health effects for the farming scenario at all of the receptor locations, and to generate a list of overall risk indices and receptor numbers. Because the highest impacts for air concentration, wet deposition and dry deposition do not coincide, and these points also differ by contaminants based on fate and transport properties, each potential exposure point had to be tested using its specific combination of parameters. Once the receptor location at which the highest potential health impacts for the farming scenario was identified, the full analysis was documented using the atmospheric dispersion and deposition results for that location.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3–32

Figure 3.9

Maximum 1-hour concentration, ISCST3 modeling of boiler stack emissions of volume (mass) weighted particles (normalized to 1 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-33

Figure 3.10

Maximum 24-hour concentration, ISCST3 modeling of boiler stack emissions of volume (mass) weighted particles (normalized to 1 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-34

Figure 3.11

Annual average concentration, ISCST3 modeling of boiler stack emissions of volume (mass) weighted particles (normalized to 1 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-35

Figure 3.12

Annual average wet deposition, ISCST3 modeling of boiler stack emissions of volume (mass) weighted particles (normalized to 1 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-36

Figure 3.13

Annual average dry deposition, ISCST3 modeling of boiler stack emissions of volume (mass) weighted particles (normalized to 1 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-37

Figure 3.14

Annual average total deposition, ISCST3 modeling of boiler stack emissions of volume (mass) weighted particles (normalized to 1 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-38

Figure 3.15

Maximum 1-hour concentration, AERMOD modeling of odor scrubbing system emissions of small particles and vapors with low Henry’s Law constants (normalized to 300 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-39

Figure 3.16

Maximum 24-hour concentration, AERMOD modeling of odor scrubbing system emissions of small particles and vapors with low Henry’s Law constants (normalized to 300 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-40

Figure 3.17

Annual average concentration, AERMOD modeling of odor scrubbing system emissions of small particles and vapors with low Henry’s Law constants (normalized to 300 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-41

Figure 3.18

Annual average wet deposition, AERMOD modeling of odor scrubbing system emissions of small particles and vapors with low Henry’s Law constants (normalized to 300 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-42

Figure 3.19

Annual average dry deposition, AERMOD modeling of odor scrubbing system emissions of small particles and vapors with low Henry’s Law constants (normalized to 300 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-43

Figure 3.20

Annual average total deposition, AERMOD modeling of odor scrubbing system emissions of small particles and vapors with low Henry’s Law constants (normalized to 300 g/s emission rate). Top projection depicts entire modeling domain, lower projection a 6-km by 6-km region around the Maine Energy facility. Color-coded legend indicates relative values. Cross indicates facility location, and black outlines pond watersheds.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3-44

Table 3.10

Maximum modeled air dispersion concentrations of COPCs emitted from the Maine Energy facility. Concentrations are in units of :g/m3 normalized to a COPC emission rate of 1 g/s. To obtain actual modeled concentrations these values are multiplied by the COPC specific emission rates. The X and Y values give the location of the maxima in units of meters to the east (X) and north (Y) of the facility stack; negative X and Y values indicate locations to the west and south of the stack respectively. Particulate-phase COPC emissions classified as either mass-weighted (m) or surface-weighted (s); vapor-phase COPC emissions are classified based on their Henry’s Law constant (high, medium, or low), with vapor-phase mercury (Hg) classified separately. See Table 3.7 for details of these classifications.

Stack emissions COPC type

1-hour average normal flow Conc.

X

Y

Part.(m)

2.64

0

Part.(s)

3.07

-689

122

Vap.(h)

3.27

-689

Vap.(m)

3.27

Vap.(l) Vap.(Hg)

1-hour average startup flow Conc.

X

Y

24-hour average Conc.

X

1400 3.20 -5909 -1042 0.212

0

Y

1-year average Conc.

X

Y

1400 0.014

0

2200

3.45

-953 -550 0.302 -689 -2349 0.022

0

4000

122

3.75

-308 -846 0.413 -855 -2349 0.030

0

4000

-689

122

3.73

-308 -846 0.371 -608 -3447 0.031

0

4000

3.27

-689

122

3.75

-308 -846 0.412 -855 -2349 0.030

0

4000

3.27

-689

122

3.75

-308 -846 0.412 -855 -2349 0.030

0

4000

Scrubber emissions Part.(m)

27.2

376

-137

27.2

376

-137

11.0

-43

25

1.67

-49

9

Part.(s)

27.2

376

-137

27.2

376

-137

11.0

-43

25

1.67

-49

9

Vap.(h)

27.2

376

-137

27.2

376

-137

11.0

-43

25

1.67

-49

9

Vap.(m)

27.2

376

-137

27.2

376

-137

11.0

-43

25

1.67

-49

9

Vap.(l)

27.2

376

-137

27.2

376

-137

11.0

-43

25

1.67

-49

9

Vap.(Hg)

27.2

376

-137

27.2

376

-137

11.0

-43

25

1.67

-49

9

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3–45

Table 3.11

Maximum modeled deposition rates of COPCs emitted from the Maine Energy facility. Deposition rates are in units of g/m2-year normalized to a COPC emission rate of 1 g/s. To obtain actual modeled concentrations these values are multiplied by the COPC specific emission rates. The X and Y values give the location of the maxima in units of meters to the east (X) and north (Y) of the facility stack; negative X and Y values indicate locations to the west and south of the stack respectively. Particulate-phase COPC emissions classified as either mass-weighted (m) or surface-weighted (s); vapor-phase COPC emissions are classified based on their Henry’s Law constant (high, medium, or low), with vapor-phase mercury (Hg) classified separately. See Table 3.7 for details of these classifications.

Stack emissions COPC type

Vapor-phase wet deposition

Particulate-phase wet deposition

Particulate-phase dry deposition

Dep.

X

Y

Dep.

X

Y

Dep.

X

Y

Part.(m)







1.52

-43

-25

0.0667

0

2200

Part.(s)







0.839

-43

-25

0.0333

0

2200

Vap.(h)

0.187

-43

-25













Vap.(m)

0.0331

-43

-25













Vap.(l)

0.00034

-43

-25













0.148

-43

-25













Vap.(Hg)

Scrubber emissions Part.(m)







0.000132

-87

-50

0.017

-49

9

Part.(s)







0.000132

-87

-50

0.017

-49

9

Vap.(h)

0.000069

-87

-50













Vap.(m)

0.0069

-87

-50













Vap.(l)

0.000132

-87

-50













0.0069

-87

-50













Vap.(Hg)

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3–46

Table 3.12

Maximum modeled air dispersion concentrations and deposition rates at the two assessed watersheds (Goosefare Brook and Wilcox Pond) for COPCs emitted from the Maine Energy facility. Concentrations are in units of :g/m3; deposition rates are in units of g/m2-yr, both are normalized to a COPC emission rate of 1 g/s. To obtain actual modeled concentrations these values are multiplied by the COPC specific emission rates. Particulate-phase COPC emissions classified as either mass-weighted (m) or surface-weighted (s); vapor-phase COPC emissions are classified based on their Henry’s Law constant (high, medium, or low), with vapor-phase mercury (Hg) classified separately. See Table 3.7 for details of these classifications. Only the parameters shown are used for multi-pathway modeling of water sheds, see Chapter 5, Section 2 for details.

Stack emissions Goosefare Brook COPC type

Wilcox Pond

Concentration

Wet deposition

Total deposition

Concentration

Wet deposition

Total deposition

Part.(m)





0.0405





0.0311

Part.(s)





0.0208





0.0165

Vap.(h)

0.0190

0.0000027



0.0144

0.0000045



Vap.(m)

0.0190

0.000269



0.0144

0.000430



Vap.(l)

0.0190

0.00138



0.0144

0.00216



Vap.(Hg)

0.0190

0.00112



0.0144

0.00170



Scubber emissions Goosefare Brook COPC type

Wilcox Pond

Concentration

Wet deposition

Total deposition

Concentration

Wet deposition

Total deposition

Part.(m)





0.00029





0.00016

Part.(s)





0.00029





0.00016

Vap.(h)

0.0387

7.0 E-7



0.0219

1.3 E-6



Vap.(m)

0.0387

6.9 E-5



0.0219

1.3 E-4



Vap.(l)

0.0386

1.5 E-6



0.0219

2.5 E-6



Vap.(Hg)

0.0387

6.9 E-5



0.0219

1.3 E-4



Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3–47

Table 3.13

Modeled air dispersion concentrations and deposition rates for COPCs emitted from the Maine Energy facility at the site identified as having the greatest risks for the farming exposure scenario. Concentrations are in units of :g/m3; deposition rates are in units of g/m2-yr, both are normalized to a COPC emission rate of 1 g/s. To obtain actual modeled concentrations these values are multiplied by the COPC specific emission rates. Particulate-phase COPC emissions classified as either mass-weighted (m) or surface-weighted (s); vapor-phase COPC emissions are classified based on their Henry’s Law constant (high, medium, or low), with vapor-phase mercury (Hg) classified separately. See Table 3.7 for details of these classifications. Only the parameters shown are used for multi-pathway modeling of water sheds, see Chapter 5, Section 2 for details.

Stack emissions COPC type

Concentration

Wet deposition

Total deposition

Part.(m)





0.083

Part.(s)





0.044

Vap.(h)

0.029

3.2 E-3



Vap.(m)

0.029

6.3 E-4



Vap.(l)

0.029

3.2 E-3



Vap.(Hg)

0.029

2.6 E-3



Concentration

Wet deposition

Total deposition

Part.(m)





0.00052

Part.(s)





0.00013

Vap.(h)

0.063

1.3 E-6



Vap.(m)

0.063

1.3 E-4



Vap.(l)

0.063

2.6 E-6



Vap.(Hg)

0.063

1.3 E-4



Scubber emissions COPC type

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

3–48

4

Exposure scenario selection

The multi-pathway exposure assessment builds on the air dispersion and deposition modeling by estimating the concentrations of COPCs in a variety of environmental media to which humans and wildlife may be exposed. To determine which media need to be evaluated, the pathways which lead to exposures must be defined. The pathways that will be assessed in this risk assessment include the direct inhalation of airborne contaminants, and a variety of indirect pathways that consider the deposition of contaminants to soil, water, and vegetation, with possible transfer and accumulation in the food-chain. Exposure scenarios are defined as a combination of such exposure pathways evaluated for a receptor at a specific location. The locations to be evaluated are those where there is the potential for the reasonable maximum longterm human exposures to emitted COPCs to occur through a few specific pathways. As recommended in the HHRAP, the following exposure scenarios will be considered for the evaluation of chronic risks: • • • •

Residents (adult and child); Recreational farmers (adult and child); Recreational fishers (adult and child); and Nursing infants.

Children are distinguished from adults because their rates of exposure to chemicals (as expressed per unit body weight) are frequently higher. Recreational fishers and farmers represent individuals whose diet includes a substantial portion of food that they catch in local waters and raise on local lands. These scenarios have been renamed from the subsistence fisher and farmer scenarios that are described in the HHRAP to better characterize the habits of people living in the Cities of Biddeford and Saco and nearby areas. The term subsistence suggests that essentially all of a person’s food source derives from a single source. An examination of the HHRAP’s assumptions reveals that the rates of food ingestion for the subsistence scenarios are well below levels required for sustained existence. Hence, the term subsistence is an inappropriate descriptor for the scenarios characterized in the HHRAP. In order to avoid the erroneous implications that could be associated with the term subsistence, the updated risk assessment for the Maine Energy facility will use the descriptor “recreational” to refer to high-end exposure scenarios. A recreational fisher is intended to represent a person living in the Biddeford/Saco area who fishes frequently in local waters and regularly consumes a substantial portion of the catch. Similarly, a recreational farmer is intended to represent a person living within the area who raises vegetables and livestock to derive a significant portion of their food supply.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

4–1

Allowing for the change in descriptors, the exposure scenarios to be evaluated are essentially those recommended in the HHRAP. Table 4.1, reproduced and adapted from the HHRAP, delineates the exposure pathways to be evaluated for each exposure scenario. Evaluation of the surface water pathway as a source of drinking water is evaluated based on the Saco River, which serves as the source of municipal water to many area residents.

Table 4.1

Exposure scenarios to be evaluated in the human health risk assessment Exposure ScenariosA

Exposure pathways Inhalation of vapors and particlesB

x

Recreational Farmer x

Incidental ingestion of soil

x

x

x

Ingestion of drinking water

x

x

x

Ingestion of homegrownC produce

x

x

x

Resident

Ingestion of homegrown beef

x

Ingestion of milk from homegrown cows

x

Ingestion of homegrown chickens

x

Ingestion of eggs from homegrown chickens

x

Ingestion of homegrown pork

x

Recreational Fisher x

Ingestion of fishC

x

Infant ingestion of breast milkD

x

x

x

A

All of these exposure scenarios will be evaluated for adults and children. The acute risks of one-hour and 24-hour direct inhalation of COPCs will be evaluated at the residential receptor site. C Ingestion of homegrown produce and livestock as well as fish are evaluated at ingestion rates that reflect a substantial dietary intake of foods raised or caught locally in the Biddeford/Saco area. D Ingestion of breast milk will be evaluated for dioxin exposure to infants of mothers exposed to COPCs in each of the three exposure scenarios B

As indicated in Table 4.1, the ingestion of breast milk is evaluated as a special pathway of potential concern independent of the other exposure scenarios, as recommended in the HHRAP. Infants can be exposed to concentrated doses of pollutants that are transferred through their mother’s milk. The nursing infant pathway is based on the assumption that the mother receives exposure to contaminant emissions from the Maine Energy facility, and hence is a member of the

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

4–2

resident, recreational fishing, and/or recreational farming populations evaluated in the risk assessment. The first step in the multi-pathway exposure assessment for each of these scenarios involves the interpretation and use of the atmospheric modeling analysis. As described in Chapter 3 for both the boiler stack and odor scrubbing system emissions, the maximum projected ambient concentration and dry deposition impacts are found some distances beyond the property line at which the elevated plume touches down (on average), and the locations for the two sources will differ. However the maximum wet deposition impacts occur very close to the facility. Consideration has been given to land use and the prediction of impacts at specific receptor locations of interest. The residential receptor scenarios are evaluated at the locations of highest projected facility impacts outside of the facility property (as there are residential areas fairly close to the facility). For the recreational fishing scenario, the fate-and-transport modeling, however, considers the actual locations and characteristics of water bodies in estimating pollutant levels in fish. It is assumed that the individual fishers live at the worst-case residential location, as it is plausible that any resident can be a recreational fisher. The 1996 risk assessment focused on two small ponds because it is more likely that facility emissions could more substantially affect their water quality. As a sensitivity calculation, the updated health risk assessment also considers fish taken in the Saco River, a more significant fishing resource, to test the assumption that the river’s greater dilution volume reduces potential impacts from facility emissions (See Chapter 8). The recreational farming scenario has been evaluated under two sets of assumptions regarding consumption of home-grown animal products in the vicinity of the Maine Energy facility. First, that there may be residential locations where sufficient space exists for the growth of sufficient vegetables and fruits to support the HHRAP home-grown consumption rates at any location in the vicinity of the Maine Energy facility. Second, that there may be residential locations where sufficient space exists for the growth of sufficient vegetables and fruits, and animal products to support the HHRAP home-grown consumption rates at any location beyond 1 kilometer from the Maine Energy facility. The assumption that there is no substantial consumption of animal products (i.e., beef, pork, poultry, eggs, and milk) within 1 km of the facility is supported by Figures 1.1, 1.2, and 3.6. Figures 1.1 and 1.2 show the relatively dense urban setting in the area within 1 km of the facility. Examination of Figure 3.6 reveals that the nearest identifiable area of cultivated land is approximately 2 km to the east-southeast of the Maine Energy facility (in the area to the south of the laurel Hill Cemetery on Figure 1.1). The maximum estimated health risks for the recreational farming scenario are somewhat insensitive to the specific distance chosen for this cutoff (at distance between around 700 and 5000 meters) because the maximum COPC deposition rates do not change significantly once the receptors are outside the small area near the facility where the high wet deposition rates are at their highest. Even if there are locations with the 1 km distance from the facility where chickens may exist, it is unlikely that a sufficient number might be kept to sustain the assumed long-term consumption rates used to evaluate this scenario (roughly 10.5 ounces of poultry and eggs consumed per adult per week); it is highly unlikely that pigs, and dairy and beef cattle exist in this location. The method used to

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

4–3

identify the location of the maximum potential health risks under these assumptions for the farming scenario are described in Section 3.7. In addition to the resident, recreational farmer, and recreational fisher scenarios, an acute risk scenario will also be considered to evaluate the potential for facility emissions to adversely affect nearby residents over short time periods (e.g., one hour, one day) via the inhalation of chemicals of potential concern. This scenario will be evaluated at the location of the highest estimated onehour, off-site ambient air COPC concentrations. Because the updated risk assessment will consider two sources of emissions from the Maine Energy facility released from different heights, there will likely be different worst-case projected points of impact. Consequently, the resident, recreational farmer, and recreational fisher scenarios will be considered at the locus of each of the worst-case projections for emissions from the boiler stack and the odor scrubbing system. The potential for such acute health effects is evaluated using the maximum 1-hour average COPC concentrations from the air dispersion modeling per HHRAP guidance, and by comparing modeled 24-hour COPC concentrations with the Air Toxics Ordinance ambient air limits (AALs) promulgated by the City of Biddeford.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

4–4

5

Estimation of media concentrations

In order to evaluate the impact of atmospheric emissions from the Maine Energy facility, the concentrations of each COPC are estimated in a variety of media, specifically: air, soil, natural vegetation, fruit and vegetable crops, livestock and related farm products, surface and drinking water supplies, and fish. A wide range of fate and transport parameters are needed to conduct such a multi-pathway assessment, including compound-specific properties of COPCs, which were discussed in Chapter 2, and site-specific land-use characteristics, which will be described in the relevant sections below. Table 5.1 contains the required general default parameters from the HHRAP guidance for time durations, and air, water and soil properties, as well as site-specific parameters for the watersheds for Wilcox pond and the unnamed pond on the Goosefare Brook. For each medium addressed in the calculations, detailed algorithms and equations are applied, as described in the HHRAP’s Chapter 5 and its Appendix B. The HHRAP algorithms are primarily based on previous guidance set forth by the U.S. EPA and other regulatory agencies. The HHRAP algorithms and default assumptions are used except where site-specific considerations suggest the use of different models and assumptions. Deviations from the HHRAP guidance and its default parameters are noted in the descriptions of the calculations and the impacts of these deviations are addressed in the uncertainty evaluations in Chapter 8. The concentration estimates for a given medium are frequently passed on to another medium following the natural progression for the transport of compounds in the environment (e.g. soil concentrations progress to vegetation concentrations which progress to livestock concentrations). Eventually these concentrations are incorporated into human exposure estimates which then lead to estimates of possible health impacts. Air concentrations and depositions rates for each COPC have been described in Chapter 3, and have been used to calculate concentrations in other environmental media in the following order: soil, produce, animal tissue, water, and fish. Only the equations and parameters used to calculate COPC concentrations are given in the text below. The COPC-specific properties required for the calculations are given in full in Appendix II, and the calculated COPC concentrations for each medium are given in Appendix IV.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–1

Table 5.1 General and site-specific parameters and properties required for the calculation of COPC concentrations in various environmental media. Parameter name Time duration parameters Time period at beginning of combustion Length of exposure, child Length of exposure, adult period for resident and subsistence fisher Length of exposure, subsistence farmer Time period of combustion Standard air and water parameters Temperature ambient Dry deposition velocity Air density Water density von Karman's constant Viscous sublayer thickness Viscosity of water Viscosity of air Drag coefficient

Symbol

Value

Units

Reference

T1

0

yrs

HHRAP Table B-1-1

T2

6

yrs

HHRAP Table B-1-1

T2

24

yrs

HHRAP Table B-1-1

T2

40

yrs

HHRAP Table B-1-1

tD

100

yrs

HHRAP Table B-1-1

Ta Vdv Da Dw k 8z :w :a Cd

298.1 1.4 1200 1 0.4 4 0.0169 1.81 E-4 0.0011

HHRAP Eqn. 5-6A see RAP HHRAP Table B-2-8 HHRAP Eqn. 5-41B HHRAP Table B-4-20 HHRAP Table B-4-20 HHRAP Table B-4-20 HHRAP Table B-4-21 HHRAP Eqn. 5-41B

Ideal gas constant

R

8.21 E-5

Temperature correction factor Soil related parameters Precipitation, annual average Irrigation, annual average Evapotranspiration, annual average Surface runoff, annual average Soil mixing zone depth, untilled Soil mixing zone depth, tilled Soil bulk density Soil solids particle density

2

1.026

°K cm/s g/m3 g/cm3 unitless unitless g/cm-s g/cm-s unitless atm-m3/mol-K at 20°C unitless

P I

109 0

cm/yr cm/yr

Geraghty et al.,1973 conservative assumption

Ev

58

cm/yr

Geraghty et al.,1973

RO Zs Zs BD Ds

51 1 20 1.5 2.7

cm/yr cm cm g/cm3 g/cm3

Soil volumetric water content

2sw

0.2

unitless

Soil volumetric water content Soil bioavailability Rainfall factor Erodability factor Length slope factor

2v Bs RF K LS

0.24 1 115 0.19 0.55

unitless unitless yr–1 ton/acre unitless

Geraghty et al.,1973 HHRAP Table B-1-1 HHRAP Table B-1-1 HHRAP Table B-1-1 HHRAP Errata, page 18 HHRAP Table B-1-3, and USDA (1981) HHRAP Errata, page 18 HHRAP Table B-3-10 Wischmeire, 1978 Soil Conservation Service, 1982

HHRAP Eqn. 5-40 HHRAP Eqn. 5-40

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–2

Table 5.1 (continued) General and site-specific parameters and properties required for the calculation of COPC concentrations in various environmental media. Parameter name Symbol Value Units Reference Cover management factor C 0.1 unitless HHRAP Table B-4-13 Supporting practice factor PF 1 unitless HHRAP Table B-4-13 Sediment delivery ratio a 1.92 unitless coefficient — Wilcox Pond interpolated from values in Sediment delivery ratio HHRAP Table B-4-14 a 1.74 unitless coefficient — Goosefare pond Sediment delivery ratio exponent b 0.125 unitless HHRAP Table B-4-14 Sediment delivery ratio — SD 0.311 unitless Wilcox Pond HHRAP Table B-4-14 Sediment delivery ratio — SD 0.252 unitless Goosefare pond 0.269 kg/m2-yr HHRAP Table B-4-13 Unit soil loss Xe Watershed parameters 293 °K HHRAP Eqn. 5-30 Water temperature Twk Water body surface area — 3.4 E+4 m2 Aw Wilcox Pond Water body surface area — 3.4 E+4 m2 Aw unnamed Goosefare pond Determined from USGS Total watershed area — Wilcox topological map 7.5 minute A(L) 2.1 E+6 m2 Pond series Total watershed area — 2 A(L) 5.3 E+6 m unnamed Goosefare pond Impervious watershed area A(I) 1 % Volume flow through water body 2.0 E+6 m3/yr estimated from precipitation Vfx — Wilcox Pond recharge, Cambridge Volume flow through water body 5.0 E+6 m3/yr Environmental, 1996 Vfx — unnamed Goosefare pond* Suspended solids, total — TSS 2.72 mg/L HHRAP Eqn.5-36C Wilcox Pond Suspended solids, total — TSS 5.31 mg/L HHRAP Eqn.5-36C unnamed Goosefare pond 1825 m/yr HHRAP default Eqn.5-36C Deposition rate, suspended solids Dss averaged from 1986 - 1990 Wind speed, average annual W 4.177 m/s hourly data Water column depth — Wilcox Maine Department of Inland 1.07 m dwc Pond* Fisheries & Wildlife, 2001 0.03 m HHRAP Table B-4-15 Benthic sediment depth dbs 0.6 unitless HHRAP Table B-4-16 Porosity, bed sediment 2bs 1 g/cm3 HHRAP Table B-4-16 Concentration, bed sediment Cbs 0.07 unitless HHRAP Table B-4-28 Fish lipid content flipid Organic content of bottom 0.04 unitless HHRAP Table B-4-28 OCsed sediment

* The water column depth for Wilcox pond was also used in the modeling of the pond on Goosefare Brook and in the bounding estimates for the Saco River. The TSS concentration for the pond on Goosefare Brook was also used in the bounding estimates for the Saco River.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–3

5.1

COPC concentrations in soil

The incremental concentrations of COPCs in soils due to emissions from the Maine Energy facility are dependent on the COPC’s air-to-soil deposition rate, the rate at which the COPC is lost from or degraded in the soil, and the length of time over which these processes have occurred. Two equations are required to calculate the concentrations based on whether they are to be estimated at given moment in time, or averaged over a period of time. The former calculation is applied to COPCs being evaluated for noncancer health risks, which are not assumed to be dependent on cumulative exposures; the latter calculation is applied to COPCs being evaluated for cancer risks, which are based on lifetime average COPC exposures. Some COPCs are evaluated for cancer and noncancer risks, therefore both calculations have been performed for all COPCs. Because the deposition of COPCs to the soil is assumed to be constant over the operating lifetime of the Maine Energy facility and the loss of COPCs from the soil is proportional to their concentrations, the equations predict COPC levels in soil to increase over time asymptotically approaching a steady state value at which the deposition and loss terms would be equal. Therefore, in order to estimate the highest levels of COPCs in soil to which an individual might be directly or indirectly exposed, COPC concentrations are calculated for noncancer risks at the end of the facility’s predicted lifetime, and for cancer risks are averaged over an individual’s assumed exposure duration up to the end of the facility’s predicted lifetime. The equation used to calculate soil concentrations of COPCs evaluated for cancer risks is:

Cs =

 exp( − ks ⋅ T1 )   exp( − ks ⋅ tD)    ⋅   tD +  −  T1 +    ks ks ks ⋅ (tD − T1 )    Ds

and the equation used to calculate soil concentrations of COPCs evaluated for noncancer risks is:

CstD

Ds ⋅ [1 − exp( − ks ⋅ tD) ] = ks

where the terms are: Cs Ds T1 ks tD CstD

Average soil concentration over exposure duration (mg COPC/kg soil); Deposition term (mg COPC/kg soil/yr); Time period at the beginning of combustion (yr); COPC soil loss constant due to all processes (yr–1); Time period over which deposition occurs (time period of combustion) (yr); Soil concentration at time tD (mg/kg).

Default values for T1=0, and tD=100 years are taken from HHRAP Appendix Table B-1-1; the values for Ds, and ks are calculated below. The deposition term, Ds, is calculated from the COPC

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–4

atmospheric concentrations and deposition rates determined by the ISCST3 modeling described in Chapter 3. The ISCST3 values for unitized wet and dry deposition of particles and vapors are combined with the COPC emission rates (described in Chapter 2) and converted into a soil concentration deposition term Ds by including the soil mixing depth and density in the denominator:

 100 ⋅ Q  Ds =   ⋅ Fv ⋅ (0.31536 ⋅ Vdv ⋅ Cyv + Dywv ) + ( Dydp + Dywp) ⋅ (1 − Fv )  Z s ⋅ BD 

[

]

where the terms are: Ds 100 Q Zs BD Fv 0.31536 Vdv Cyv Dywv Dydp Dywp

Deposition term (mg COPC/kg soil-yr); Units conversion factor (mg-m2/kg-cm2); COPC emission rate (g/s); Soil mixing zone depth (cm); Soil bulk density (g soil/cm3 soil); Fraction of COPC air concentration in vapor-phase (yr–1); Units conversion factor (m-g-s/cm-:g-yr); Dry deposition velocity (cm/s); Unitized yearly average air concentration in the vapor-phase (:g-s/g-m3); Unitized yearly average wet deposition from vapor-phase (s/m2-yr); Unitized yearly average dry deposition from particle-phase (s/m2-yr); Unitized yearly average wet deposition from particle-phase (s/m2-yr).

The calculation of COPC emission rates, Q, were described in Chapter 2. Based on HHRAP Appendix Table B-1-1, two soil mixing zone depths have been modeled to account for different transport and exposure scenarios: 20 cm for tilled soils, and 1 cm for untilled soils. The soil bulk density, BD, is 1.5 g/cm3, and the vapor-phase deposition velocity, Vdv, is 1.4 cm/s. The fraction of each COPC concentration in air in the vapor-phase, Fv, is a COPC-specific parameter. The loss rate for COPCs from soils is the sum of several terms:

ks = ksg + kse + ksr + ksl + ksv where the terms are: ks ksg kse ksr ksl ksv

COPC soil loss constant due to all processes (yr–1); COPC loss constant due to biotic and abiotic degradation (yr–1); COPC loss constant due to soil erosion (yr–1); COPC loss constant due to surface runoff (yr–1); COPC loss constant due to leaching (yr–1); COPC loss constant due to volatilization (yr–1).

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–5

Based on HHRAP Appendix Table B-1-2, kse is taken as zero for all COPC’s at the point of maximum emissions impact, but is included in the soil concentration calculations within the watershed as discussed below in Section 5-4; ksg values are COPC-specific and are based on HHRAP Appendix Tables A-3 and Howard et al. (1991). The other loss terms are calculated as follows: Losses of COPCs due to surface runoff and leaching are dependent on the COPC’s soil-water partitioning coefficient and the amount of water available for these processes:

RO ksr = θ sw ⋅ Z s

ksl =

  1   ⋅   1 + ( Kd s ⋅ BD / θ sw ) 

P + I − RO − E v

[

]

θ sw ⋅ Z s 1 + ( BD ⋅ Kd s / θ sw )

where the terms are: ksr ksl RO P I Ev

2sw

Zs Kds BD

COPC loss constant due to surface runoff (yr–1); COPC loss constant due to leaching (yr–1); Average annual surface runoff from pervious areas (cm/yr); Average annual precipitation (cm/yr); Average annual irrigation (cm/yr); Average annual evapotranspiration (cm/yr); Soil volumetric water content (mL water/cm3 soil); Soil mixing zone depth (cm); Soil-water partition coefficient (mL water/g soil); Soil bulk density (g soil/cm3 soil).

Based on HHRAP Appendix Tables B-1-4 and B-1-5, and data from the Geraghty et al. (1973), site-specific values used are RO = 51 cm/yr, P = 109 cm/yr, I = 0 cm/yr, and E = 58 cm/yr; default values used for BD = 1.50 g/cm3 and 2sw = 0.2 mL/cm3, and COPC-specific values are used for Kds. The calculation of the COPC loss constant due to volatilization, ksv, is described in the HHRAP Errata memorandum (U.S. EPA, 1999, pages 17-19) as the product of the gas equilibrium coefficient, Ke, and the gas-phase mass transfer coefficient, Kt:

ksv = Ke ⋅ K t

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–6

The equilibrium coefficient, Ke, is given by:

31536 . × 10 7 ⋅ H Ke = Z s ⋅ Kds ⋅ R ⋅ Ta ⋅ BD where the terms are: Ke 3.1536×107 H Zs Kds R Ta BD

COPC gas equilibrium coefficient (s/yr-cm); Units conversion (s/yr); Henry’s Law constant (atm-m3/mol); Soil mixing zone depth (cm); Soil-water partition coefficient (mL/kg); Ideal gas constant (atm-m3/mol-K); Average ambient air temperature (K); Soil bulk density (g soil/cm3 soil).

The gas-phase mass transfer coefficient, Kt, is given by:

Kt =

  BD  Da  1−   − θ sw  Zs   ρs  

where the terms are: Kt Da Zs BD

Ds 2sw

Gas-phase mass transfer coefficient (cm/s); Diffusion coefficient of COPC in air (cm2/s); Soil mixing zone depth (cm); Soil bulk density (g soil/cm3 soil); Density of soil solids (g/cm3); Volumetric soil water content (unitless).

Based on HHRAP Appendix Table B-1-3 and the Errata memorandum (U.S. EPA, 1999), default values are used Zs = 1 cm (untilled) or 20 cm (tilled), Ta = 298 °K, BD = 1.50 g/cm3, 2sw = 0.2 mL/cm3, and Ds = 2.7 g/cm3. The ideal gas constant, R, is 8.205 ×10–5 atm-m3/mol-K. COPCspecific values are used for H, Kds and Da.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–7

5.2

COPC concentrations in produce, grain, and vegetation

Natural vegetation and agricultural produce are assumed to receive COPCs through three mechanisms: direct deposition of particulate COPCs, direct air-to-plant transfer of vapor-phase COPCs, and uptake of COPCs in soils through the plant’s root system. The wide range of vegetation and produce for which COPC concentrations need to be estimated is divided within the HHRAP guidance into the following categories: aboveground, belowground, and protected produce; forage; silage; and grain. For types of plants with edible portions below ground or otherwise protected from direct atmospheric deposition and transfer, only COPC uptake through the plant’s root system is modeled. The equations and parameters needed to calculate COPC concentrations by all three mechanisms have been evaluated in the HHRAP for each plant type, and are described below. The concentration of COPCs in the exposed and edible portions of vegetation due to direct atmospheric deposition of particles is calculated by:

Pd i =

[

]

[

]

1,000 ⋅ Q ⋅ (1 − Fv ) ⋅ Dydp + ( Fw ⋅ Dywp) ⋅ Rpi ⋅ 1 − exp( − kp ⋅ Tpi ) Ypi ⋅ kp

where the terms are: Pdi 1,000 Q Fv Dydp Fw Rpi kp Tpi Ypi

Plant (aboveground produce) concentration due to direct (wet and dry) deposition (mg COPC/kg DW); Units conversion factor (mg/g); COPC emission rate (g/s); Fraction of COPC air concentration in vapor-phase (unitless); Unitized yearly average dry deposition from particle-phase (s/m2-yr); Fraction of COPC wet deposition that adheres to plant surfaces (unitless); Interception fraction of the edible portion of plant for the ith plant group (unitless); Plant surface loss coefficient (yr–1); Length of plant exposure to deposition per harvest of the edible portion of the ith plant group (yr); Yield or standing crop biomass of edible portion of the ith plant group (kg DW/m2).

Values of Rp, kp, Tp, and Yp for the various plant groups are given in Table 5.2. Values for above ground produce are from HHRAP Appendix Table B-2-7, and the values for forage and silage from Table B-3-7. Based on HHRAP Appendix Table B-3-7, Fw is 0.2 for anions and 0.6 for cations and all organic COPCs examined in this report.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–8

Table 5.2

Parameters for calculating the direct particle-phase deposition of COPCs to exposed vegetation and produce. Rp (unitless)

kp (yr–1)

Tp (yr)

Yp (kg DW/m2)

Aboveground produce

0.39

18

0.164

2.24

Forage

0.5

18

0.12

0.24

Silage

0.46

18

0.16

0.8

The concentration of COPCs in the exposed and edible portions of vegetation due to direct transfer from the vapor-phase is calculated by:

Pv = Q ⋅ Fv ⋅

Cyv ⋅ Bv ag ⋅ VGag

ρα

where the terms are: Pv Q Fv Cyv Bvag VGag

Da

Concentration of COPC in the plant resulting from air-to-plant transfer (:g COPC/g DW); COPC emission rate (g/s); Fraction of COPC air concentration in vapor-phase (unitless); Unitized yearly average air concentration from vapor-phase (:g- s/g-m3); COPC air-to-plant biotransfer factor ([mg COPC/g DW]/[mg COPC/g air]) (unitless); Empirical correction factor for aboveground produce (unitless); Density of air (g/m3).

The parameter Bvag is COPC-specific; for aboveground produce the parameter VGag is 0.01 for COPCs with a log Kow greater than 4, and 1.0 for COPCs with a log Kow less than 4; for forage the parameter VGag is 1, for silage the parameter VGag is 0.5. The concentration of COPCs in vegetation due to transfer from soil through the roots of vegetation is calculated for exposed and protected aboveground produce, forage, silage, and grain by:

Pr = Cs ⋅ Br and for belowground produce by:

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–9

Pr =

Cs ⋅ RCF ⋅ VGrootveg Kd s ⋅ 1kg / L

where the terms are: Pr Br VGrootveg Kds Cs RCF

Concentration of COPC in produce due to root uptake (:g COPC/g DW); Plant-soil bioconcentration factor for produce (unitless); Empirical correction factor for belowground produce (unitless); Soil-water partition coefficient (mL/kg); Average soil concentration over exposure duration (mg COPC/kg soil); Root concentration factor (unitless).

The values for Br for each aboveground plant type, and Kds and RCF for belowground produce are COPC-specific and either derived from HHRAP Appendix Table A-3, or from the correlations in HHRAP Appendix A-3; for the parameter VGrootveg is 0.01 for COPCs with a log Kow greater than 4, and 1.0 for COPCs with a log Kow less than 4.. Although some types of vegetation, such as forage, are not likely to be grown in tilled soil, the HHRAP recommends (page 5-20) that the COPC concentrations in soil be derived from the tilled soil calculations to reflect the depth of the plants’ root zone.

5.3

COPC concentrations in livestock and related farm products

The COPC concentrations in animal tissues (beef, pork, chicken) and dairy products (milk, eggs) are determined by the COPC concentrations in the various parts of the animal’s diets (including soil), each component’s intake rate, and COPC-to-animal product biotransfer functions:

Aj =

( ∑ ( F ⋅ Qp i

ij

)

)

⋅ Pij + Qs j ⋅ Cs ⋅ Bs ⋅ Ba j

where the terms are: Aj Fi Qpij Pi Qsj Cs

Concentration of COPC in animal product j (mg COPC/kg); Fraction of plant type I grown in contaminated soil and ingested by the animal (unitless); Quantity of plant type I eaten by animal type j each day (kg DW plant/day); Concentration of COPC in plant type I eaten by the animal type j (mg/kg DW); Quantity of soil eaten by animal type j each day (kg soil/day); Average soil concentration over exposure duration (mg COPC/kg soil);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–10

Bs

Soil bioavailability factor (unitless).

The intake rates (Q’s) for each animal group are shown in Table 5.3 (HHRAP Appendix B-3-10 for beef, B-3-11 for milk, B-3-12 for pork, B-3-13 for eggs and B-3-14 for chicken). Both the fraction of plants grown in contaminated soil, Fi, and the soil bioavailability factor, Bs, are assumed to be equal to one. All soil ingested by farm animals is assumed to be untilled. Table 5.3

Intake rates of various feed types and soil required for the calculation of COPC concentrations in farm animals and animal products. Animal food material intake rate (kg/day) (dry weight basis)

Animal

forage

silage

grain

soil

beef (cattle)

8.8

2.5

0.47

0.5

milk (cows)

13.2

4.1

3

0.4

pork (pigs)

0

1.4

3.3

0.37

chicken/eggs (chickens)

0

0

0.2

0.022

5.4

COPC concentrations in surface water

As described in Chapter 3, potential COPC impacts are estimated for two small ponds with watersheds in areas that are projected to receive the highest deposition levels of contaminants from the Maine Energy facility. The two ponds were identified from topographic maps of the area, in conjunction with the patterns of projected impacts over the modeling domain. The first is an unnamed pond located about 3.3 km northwest of the Maine Energy facility. This pond lies just to the east of Goosefare Hill, and is fed by Goosefare and Innis Brooks. The second is Wilcox Pond, which is situated about 2.8 km to the south of the Maine Energy facility. The concentrations of COPCs in these ponds have been calculated using the following equation, which includes terms for the total input (loading) and loss (outflow and dissipation) of COPCs from the water body:

Cwtot =

LT

Vf x ⋅ f wc + k wt ⋅ AW ⋅ (d wc + d bs )

where the terms are:

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–11

Ctot LT VfX fwc kwt Aw dwc dbs

Total water body COPC concentration (including water column and bed sediment) (g COPC/m3 water body); Total COPC load to the water body (g/yr); Average volumetric flow rate through the water body (m3/yr); Fraction of total water body COPC concentration in the water column (unitless); Overall total water body COPC dissipation rate constant (yr–1); Water body surface area (m2); Depth of water column (m); Depth of upper benthic sediment layer (m).

The relevant parameters for Wilcox Pond and the unnamed Goosefare pond were estimated in the 1996 Maine Energy Risk Assessment Report (Cambridge Environmental, 1996). The average volumetric flow rate through Wilcox Pond is 2 × 106 m3/yr; the average volumetric flow rate through the unnamed Goosefare pond is 5 × 106 m3/yr. The surface area of both ponds is 3.4 × 104 m2; and the average depth of the water column is 1.07 m (Maine Department of Inland Fisheries & Wildlife, 2001). The depth of the upper benthic sediment layer is taken as 0.03 m based on HHRAP Appendix Table B-4-15. Because the terms for COPC loading, dissipation, and fractionation are each dependent on the calculation of several additional terms, the determinations of LT, kwt, and fwc are described separately below. Potential COPC impacts on surface water quality in the Saco River are estimated based on a simplified bounding model which assumes that all the COPCs emitted from the Maine Energy facility enter the Saco River directly, as was done in the 1996 Maine Energy Risk Assessment (Cambridge Environmental, 1996). The methods for these estimates are described in section 5.4.3 following the methods for estimating COPC impacts on the small ponds near the facility.

5.4.1

COPC loading to nearby ponds

The total COPC loading to the water body is comprised of direct wet and dry deposition of particulate-phase COPCs, direct wet deposition and diffusion of vapor-phase COPCs, runoff of COPCs from pervious and impervious surfaces within the watershed, erosion of COPC containing soils from within the watershed into the water body, and internal chemical or biological transformation of COPCs. The following equation gives the sum of these terms; the calculation of each of these terms are defined below.

LT = LDEP + Ldif + LRI + LR + LE + LI where the terms are: LT LDEP

Total COPC load to the water body (g/yr); Total (wet and dry) particle-phase and wet vapor-phase COPC direct deposition load to the water body (g/yr);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–12

Ldif LRI LR LE LI

Vapor-phase COPC diffusion (dry deposition) load to the water body (g/yr); Runoff load from impervious surfaces (g/yr); Runoff load from pervious surfaces (g/yr); Soil erosion load (g/yr); Internal transfer (g/yr).

The loading terms are calculated as follows: The direct wet deposition of atmospheric vapor-phase COPCs and the total wet and dry deposition of atmospheric particulate-phase COPCs is simply the product of their average watershed unitized deposition rates (as determined by the ISCST3 air modeling described in Chapter 4), their emission rates, and the surface area of the lake.

[

]

LDEP = Q ⋅ FV ⋅ D ywwv + (1 − FV ) ⋅ D ytwp ⋅ AW where the terms are: LDEP Q FV Dywwv Dytwv AW

Total (wet and dry) particle-phase and wet vapor-phase COPC direct deposition load to the water body (g/yr); COPC emission rate (g/s); Fraction of COPC air concentration in the vapor-phase (unitless); Unitized yearly (water body and watershed) average wet deposition from vapor-phase (s/m2-yr); Unitized yearly (water body and watershed) average total (wet and dry); deposition from particulate-phase (s/m2-yr); Water body surface area (m2).

Atmospheric vapor-phase COPCs are also transferred to surface waters by direct air-to-water diffusion as determined by the COPC’s emission rate, unitized air concentration, vapor fraction, and Henry’s law constant, by a COPC-specific mass transfer coefficient (calculated as shown below), and by the lake’s surface area.

Ldif

KV ⋅ Q ⋅ FV ⋅ C ywv ⋅ AW ⋅ 1 × 10 −6 = H R ⋅ Twk

where the terms are: Ldif KV Q FV

Vapor-phase COPC diffusion (dry deposition) load to the water body (g/yr); Overall COPC transfer rate coefficient (m/yr); COPC emission rate (g/s); Fraction of COPC air concentration in the vapor-phase (unitless);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–13

Cywv AW 10–6 H R Twk

Unitized yearly (water body and watershed) average air concentration from vapor-phase (:g-s/m2-yr); Water body surface area (m2); Units conversion factor (g/:g); Henry’s Law constant (atm-m3/mol); Universal gas constant (atm-m3/mol-K); Water body temperature (K).

The value for Kv, the overall COPC transfer rate coefficient, is determined using the following series of equations: −1   H   − 1   KV =  K L +  KG ⋅ R ⋅ T  wk   

−1

⋅ θ Twk − 293

where the terms are: KV KL KG H R Twk

2

Overall COPC transfer rate coefficient (m/yr); Liquid-phase transfer rate coefficient (m/yr); Gas-phase transfer rate coefficient (m/yr); Henry’s Law constant (atm-m3/mol); Universal gas constant (atm-m3/mol-K); Water body temperature (K); Temperature correction factor (unitless).

For quiescent lakes or ponds:

K L = (C

0.5 d

 ρa  ⋅ W) ⋅    ρw

0.5

k 0.33 ⋅ λz

 µw   ⋅  ρ w ⋅ Dw 

− 0.67

⋅ 31536 . × 10 7

where the terms are: KL Cd W Dw

Da Dw k

8z :w

3.1536×107

Liquid-phase transfer rate coefficient (m/yr); Drag coefficient (unitless); Average annual wind speed (m/s); Diffusivity of COPC in water (cm2/s); Density of air (g/cm3); Density of water (g/cm3); von Karman’s constant (0.4, unitless); Dimensionless viscous sublayer thickness (unitless); Viscosity of water corresponding to water temperature (g/cm-s); Units conversion factor (s/yr);

and

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–14

K G = (C

0.5 d

k 0.33  µ w   ⋅ W) ⋅ ⋅ λ z  ρ w ⋅ Dw 

− 0.67

⋅ 31536 . × 10 7

where the terms are: KG Cd W Dw

Da k

8z :w

3.1536×107

Gas-phase transfer rate coefficient (m/yr); Drag coefficient (unitless); Average annual wind speed (m/s); Diffusivity of COPC in water (cm2/s); Density of air (g/cm3); von Karman’s constant (0.4, unitless); Dimensionless viscous sublayer thickness (unitless); Viscosity of water corresponding to water temperature (g/cm-s); Units conversion factor (s/yr).

The loading of COPCs to the lake by direct surface runoff of rainwater from impervious surfaces within the watershed is calculated similarly to the direct deposition of COPCs but with the impervious surface area of the watershed substituted for surface area of the water body itself:

[

]

LRI = Q ⋅ FV ⋅ Dywwv + (1 − FV ) ⋅ Dytwp ⋅ AI where the terms are: LRI Q FV Dywwv Dytwv AI

Runoff load from impervious surfaces (g/yr); COPC emission rate (g/s); Fraction of COPC air concentration in the vapor-phase (unitless); Unitized yearly (water body and watershed) average wet deposition from vapor-phase (s/m2-yr); Unitized yearly (water body and watershed) average total (wet and dry) deposition from particulate-phase (s/m2-yr); Impervious watershed area receiving COPC deposition (m2).

The values for AI, the impervious watershed areas receiving COPC deposition, has been estimated as 2.1×104 m2 for Wilcox Pond and as 5.3×104 m2 for the unnamed Goosefare pond, using an estimate of 1% impervious areas in the watersheds based on the USGS. The next two loading terms (LR , the COPC loading due to water runoff from pervious surfaces, and LE, the COPC loading due to the erosion of soil) are proportional to the average COPC concentrations in the watershed soils. The COPC concentrations in the watershed soils were estimated using the same equations as were employed in Section 5.2 with the following input values, parameters, and additions:

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–15

The COPC air concentration and deposition terms (Cyv, Dywv, Dydp, and Dywp) are evaluated for the average values over the entire watershed rather than at the location of maximum impact, the soil mixing zone depth is assumed to be 1 cm, the value for untilled soil, based on the evaluation that the land within the watershed is not primarily used as cropland (IDEM, 2001b), and a non-zero term for the COPC loss constant due to soil erosion, kse, has been used as described below. For calculating kse, the loss rate for COPCs from the soil within an area due to the erosion of soil from the area, the HHRAP gives the following equation:

kse =

Kd s ⋅ BD 01 . ⋅ X e ⋅ SD ⋅ ER ⋅ BD ⋅ Z s θ sw + ( Kd s ⋅ BD)

where the terms are: kse 0.1 Xe SD ER Kds BD Zs

2sw

COPC soil loss constant due to erosion (yr–1); Units conversion factor (1,000 g-kg/10,000 cm2-m2); Unit soil loss (kg/m2-yr); Sediment delivery ratio (unitless); Soil enrichment ratio (unitless); Soil-water partition coefficient (mL water/g soil); Soil bulk density (g soil/cm3 soil); Soil mixing zone depth (cm); Soil volumetric water content (mL water/cm3 soil).

The unit soil loss, Xe, and the sediment delivery ratio, SD are calculated using equations described below. The HHRAP default values of 1.5 g/cm3 for BD, 1 cm for Zs, and 0.2 mL water/cm3 for 2sw are used. Kds is a COPC-specific property. The HHRAP guidance recommends that the constant kse should be set equal to zero based on the assumption that the amount of contaminated soil eroding off of the site being evaluated is countered by a roughly equal amount of contaminated soil eroding onto the site. While this is perhaps a valid assumption for receptor sites the size of a residential property or farm, it is not a valid assumption for the evaluation of a watershed as a whole because, by definition, there is no flow of water (or by extension soil eroded by flowing water) into a watershed from areas outside its boundary. COPCs emitted from the Maine Energy facility and subsequently bound to watershed soils that are eroded from the land areas of the watershed are not replaced by COPCs emitted from the Maine Energy facility eroding into the watershed. Additionally, if the loading of COPCs into the water body by the erosion of watershed soils is included in the transport modeling (by the water body loading term LE, described below), then this must be balanced by

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–16

the loss of COPCs from the same soils. For these reasons a non-zero value for kse has been used in the calculation of average COPC concentrations within the watershed. The loading to the water body caused by runoff of dissolved COPCs from pervious soils is calculated by:

LR = RO ⋅ ( AL − AI ) ⋅

Cs ⋅ BD ⋅ 0.01 θ sw + Kd s ⋅ BD

where the terms are: LR RO AL AI Cs BD

2sw

Kds 0.01

Runoff load from pervious surfaces (g/yr); Average annual surface runoff from pervious areas (cm/yr); Total watershed area receiving COPC deposition (m2); Impervious watershed area receiving COPC deposition (m2); Average soil concentration over exposure duration (in watershed soils) (mg COPC/ kg soil); Soil bulk density (g soil/cm3 soil); Soil volumetric water content (mL water/cm3 soil); Soil-water partition coefficient (mL water/g soil); Units conversion factor (kg-cm2/mg-m2).

The values for RO, AL, and AI are site-specific as described in Table 5.1; Cs is calculated as described in Section 5.1 for untilled soils, with the inclusion of a term for COPC loss by erosion, kse. The HHRAP default values of 1.5 g/cm3 for BD, and 0.2 mL water/cm3 for 2sw are used. Kds is a COPC-specific property. The loading due to COPCs associated with eroding soils entering the water body is calculated by:

LE = X e ⋅ ( AL − AI ) ⋅ SD ⋅ ER ⋅

Cs ⋅ Kd s ⋅ BD ⋅ 0.001 θ sw + Kd s ⋅ BD

where the terms are: LE Xe AL AI SD ER Cs

Soil erosion load (g/yr); Unit soil loss (kg/m2-yr); Total watershed area receiving COPC deposition (m2); Impervious watershed area receiving COPC deposition (m2); Sediment delivery ratio (watershed) (unitless); Soil enrichment ratio (unitless); Average soil concentration over exposure duration (in watershed soils) (mg COPC/ kg soil);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–17

BD

2sw

Kds 0.001

Soil bulk density (g soil/cm3 soil); Soil volumetric water content (mL water/cm3 soil); Soil-water partition coefficient (mL water/g soil); Units conversion factor (mg/g).

The soil enrichment ratio, ER, is 1 for inorganic COPCs and 3 for organic COPCs. HHRAP default values from Appendix Table B-4-11 of 1.5 g/cm3 are used for BD, and 0.2 mL water/cm3 for 2sw. Kds is a COPC-specific property. The value for unit soil loss for the watershed, Xe, is calculated by applying the Universal Soil Loss Equation (USLE):

X E = RF ⋅ K ⋅ LS ⋅ C ⋅ PF ⋅

907.18 4047

where the terms are: Xe RF K LS C PF 907.18 4047

Unit soil loss (kg/m2-yr); USLE rainfall (or erosivity factor) (yr–1); USLE erodability factor (ton/acre); USLE length-slope factor (unitless); USLE cover management factor (unitless); USLE supporting practice factor (unitless); Units conversion factor (kg/ton); Units conversion factor (m2/acre).

The value of 115 for RF was taken from Figure 1 of Wischmeire (1978), while values of 0.39 tons per acre for K, 1.5 for LS, 0.1 for C, and 1 for PF are based on default values found in HHRAP Appendix Table B-4-13. The sediment delivery ratio, SD is calculated as 0.082 by the empirical correlation:

SD = a ⋅ ( AL )

−b

where the terms are: SD a b AL

Sediment delivery ratio (watershed) (unitless); Empirical intercept coefficient (unitless); Empirical slope coefficient (unitless); Total watershed area receiving COPC deposition (m2).

Values for the coefficients a and b are derived from HHRAP Appendix Table B-4-14. The slope coefficient, b, is 0.125. The value of the intercept coefficient a has been logarithmically interpolated as 1.7 for Wilcox Pond and 1.9 for the unnamed Goosefare pond based on the values for the intercept coefficients from HHRAP Table B-4-14.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–18

5.4.2

COPC dissipation in nearby ponds

The dissipation rate of COPCs from the water body is the sum of losses through volatilization from the water surface and burial to bottom sediments:

k wt = f wc ⋅ k v + f bs ⋅ k b where the terms are: Overall total water body dissipation rate constant (yr–1); Fraction of total water body COPC concentration in the water column (unitless); Water column volatilization rate constant (yr–1); Fraction of total water body COPC concentration in the benthic sediment (unitless); Benthic burial rate constant (yr–1).

kwt fwc kv fbs kb

The two loss terms are given by:

kv =

Kv

d z ⋅ (1 + Kd sw ⋅ TSS ⋅ 1 × 10 −6 )

where the terms are: kv Kv dz Kdsw TSS 1×10-6

Water column volatilization rate constant (yr–1); Overall COPC transfer rate coefficient (m/yr); Total water body depth; Suspended sediments/surface water partition coefficient (L water/kg suspended sediments); Total suspended solids concentration (mg/L); Units conversion factor (kg/mg); and

 X e ⋅ AL ⋅ SD ⋅ 1 × 10 3 − Vf x ⋅ TSS   TSS ⋅ 1 × 10 −6   ⋅  kb =  Aw ⋅ TSS    C BS ⋅ d bs 

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–19

where the terms are: kb Xe AL SD Vfx TSS Aw CBS dbs 1×10-6 1×103

Benthic burial rate constant (yr–1); Unit soil loss (kg/m2-yr); Total watershed area receiving COPC deposition (m2); Sediment delivery ratio (watershed) (unitless); Average volumetric flow rate through the water body (m3/yr); Total suspended solids concentration (mg/L); Water body surface area (m2); Bed sediment concentration (g/cm3); Depth of upper benthic sediment layer (m); Units conversion factor (kg/mg); Units conversion factor (g/kg).

Values for the watershed’s unit soil loss, Xe, and sediment delivery ratio, SD, were calculated based on equations described above. Values for the total watershed area, AL, the water body area, AW, average volumetric flow rate, Vfx, and total suspended solids concentration, TSS, have been based on site-specific data also described above, and HHRAP default values from Appendix Table B-4-16 were used for the bed sediment concentration, CBS, and the depth of the upper benthic sediment layer, dbs.

5.4.3

COPC partitioning in nearby ponds

It is necessary to calculate the partitioning of COPCs within the water body in order to account for COPC loss to outflow (the Vfx @fwc term in the equation for Cwtot), and to employ the correct COPC concentrations in the calculations for COPC concentrations in fish and in drinking water. The partitioning divides concentrations of the compounds between the water column and the benthic sediments, and within the water column between concentrations in the dissolved-phase and bound to suspended sediments. The fraction of each COPC within the water column is given by:

f wc =

(1 + Kd

(1 + Kd

sw

sw

⋅ TSS ⋅ 1 × 10 −6 ) ⋅ d wc d z

⋅ TSS ⋅ 1 × 10 −6 ) ⋅ d wc d z + (θ bs + Kd bs ⋅ C BS ) ⋅ d bs d z

the balance of each COPC in the water body is contained within the benthic sediment:

f bs = 1 − f wc

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–20

where the terms are: fwc fbs Kdsw TSS 1×10-6 dz

2bs

Kdbs CBS dwc dbs

Fraction of total water body COPC concentration in the water column (unitless); Fraction of total water body COPC concentration in benthic sediment (unitless); Suspended sediments/surface water partition coefficient (L water/kg suspended sediment); Total suspended solids concentration (mg/L); Units conversion factor (kg/mg); Total water body depth (m); Bed sediment porosity (Lwater/Lsediment); Bed sediment/sediment pore water partition coefficient (L water/kg bottom sediment); Bed sediment concentration (g/cm3 [equivalent to kg/L]); Depth of water column (m); Depth of upper benthic sediment layer (m).

The average water column depth, dwc, of 1.07 m is based on data from the Maine Department of Inland Fisheries & Wildlife (2001) for Wilcox Pond; the same depth is assumed for the pond on Goosefare Brook. The values for 2bs of 0.6, CBS of 1 g/cm3, and dbs of 0.03 m, are default values for HHRAP Appendix Tables B-4-15 and B-4-16: Total water body depth dz is the sum of dwc and dzbs. The partitioning coefficients Kdsw, and Kdbs are COPC-specific. The concentration of total suspended solids (TSS) within the water column can be estimated using the following equation:

TSS =

X e ⋅ ( AL − AI ) ⋅ SD ⋅ 1 × 10 3 Vf x + Dss ⋅ AW

where the terms are: TSS Xe AL AI SD 1×103 Vfx Dss Aw

Total suspended solids concentration (mg/L); Unit soil loss (kg/m2-yr); Total watershed area receiving COPC deposition (m2); Impervious watershed area receiving COPC deposition (m2); Sediment delivery ratio (watershed) (unitless); Units conversion factor (L/m3); Average volumetric flow rate through the water body (m3/yr); Suspended solids deposition rate (a default value of 1,825 for quiescent lakes and ponds) (m/yr); Water body surface area (m2).

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–21

The concentration of each COPC within the water body (Cwtot) is apportioned between the water column (Cwctot), which has both dissolved-phase COPCs (Cdw) and COPCs bound to suspended sediments, and COPCs sorbed to the bed sediments (Cbs) as follows:

Cwctot = f wc ⋅ Cwtot ⋅

d wc + d bs d wc

where the terms are: Cwctot fwc Cwtot dwc dbs

Total COPC concentration in water column (mg COPC/L water column); Fraction of total water body COPC concentration in the water column (unitless); Total water body COPC concentration, including water column and bed sediment (mg COPC/L water column); Depth of water column (m); Depth of upper benthic sediment layer (m);

Cdw =

Cwctot 1 + Kd sw ⋅ TSS ⋅ 1 × 10 −6

where the terms are: Cdw Cwctot Kdsw TSS 1×10–6

Dissolved-phase water concentration (mg COPC/L water); Total COPC concentration in water column (mg COPC/L water column); Suspended sediments/surface water partition coefficient (L water/kg suspended sediment); Total suspended solids concentration (mg/L); Units conversion factor (kg/mg); and

   d wc + d bs  Kd bs  ⋅  Csb = f bs ⋅ Cwtot ⋅   θ bs + Kd bs ⋅ C BS   d bs  where the terms are: Csb fbs Cwtot Kdbs

2bs

COPC concentration sorbed to bed sediment (mg COPC/kg sediment); Fraction of total water body COPC concentration in benthic sediment (unitless); Total water body COPC concentration, including water column and bed sediment (mg COPC/L water column); Benthic sediments/sediment pore water partition coefficient (L water/kg sediment); Bed sediment porosity (Lwater/Lsediment);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–22

CBS dwc dbs

Bed sediment concentration (g/cm3); Depth of water column (m); Depth of upper benthic sediment layer (m).

The distribution fractions, fwc and fbs, have been calculated based on equations described above. The depth of the water column, dwc, and the total suspended solids concentration, TSS, are sitespecific parameters. The depth of the upper benthic sediment layer, dbs, the bed sediment porosity, 2bs, and the bed sediment concentration, CBS,, are HHRAP default parameters from Appendix Table B-4-16. Finally, the partitioning coefficients, Kdsw and Kdbs, are COPC-specific properties.

5.4.4

Bounding estimates of COPC impacts on Saco River water

The Saco River originates roughly 190 km (120 miles) from Biddeford in the White Mountains of New Hampshire and its watershed covers approximately 4,400 km2 (1,700 square miles). Because the EPA generally considers steady-state Gaussian plume models such as ISC and AERMOD to be applicable only up to a distance of 50 km (40 CFR 51, App. W), using the programs to estimate model COPC dispersion and deposition to the whole area is neither practical nor recommended. Therefore, a conservative screening analysis has been performed to assess the maximum potential impacts emissions from the Main Energy facility might have on COPC concentrations in the Saco River. As a worst-case, bounding estimate, it is assumed that the impacts can be no larger than if all of the COPCs emitted from the Maine Energy facility entered the Saco River directly. This method of estimating COPC impacts in the Saco River is a significant overestimation, because only a small fraction of the facility’s emissions will actually likely enter the Saco River. The Saco River watershed includes only a small portion of the area around the Maine Energy facility, and only emissions that are transported to west from the facility can even potentially enter the watershed. Moreover, this bounding analysis assumes that precipitation scavenging effectively deposits COPCs close to the facility. Even if 100% effective, however, precipitation scavenging could only be important the fraction of time that it snows or rains. Under this assumption, the concentration of COPCs in water is determined by dividing the rate of COPC emissions by the rate at which water flows in the river:

C =

Q VSaco

where the terms are: C COPC concentration in river water (mg/l); Q total emission rate of COPC from Maine Energy (mg/s); and VSaco volume flow of water in the Saco River (l/s). As in the 1996 Maine Energy Risk Assessment (Cambridge Environmental, 1996), the average discharge of the Saco River (VSaco) at Biddeford is estimated to be 96,600 l/s based on the

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–23

average measured discharge of the river at Cornish Maine (2,695 ft3/s = 76,300 l/s, based on long-term gauging) multiplied by the ratio of drainage area at Biddeford (1636 mi2) to that at Cornish (1293 mi2). To model the partitioning of COPCs in the Saco River among the dissolved, suspended sediment, and benthic sediment phases, the same water column depth and total suspended sediment concentrations as for the Goosefare Pond were used. Although the Saco River is a very different water body than the Goosefare Pond, based on available data, the values of these parameters are fairly similar. Based on the USGS map for Biddeford Maine, the depth of the Saco River down stream of the facility varies from areas shown as tidal flats, to sections with depths of only 1 to 2 feet (0.3 to 0.6 meters), to central sections with depths from about 8 to 22 feet (2.4 to 6.7 meters). The depth of the river in this area also varies with the tide. Upstream of the facility, the river is significantly narrower, so the deep sections present downstream do not occur. Depths at the upstream gaging station vary over the seasons, with an average depth over the past 5 years of approximately 4.3 feet (1.3 meters, http://nwis.waterdata.usgs.gov/ nwis/measurements, site17 number 01066000). The average water column depth for Goosefare Pond used in the modeling is 3.5 feet (1.07 meters). Measured TSS levels in the Saco River vary over the course of a year ranging from 1 to 11 mg/L (USGS Water Resources Data, Maine, Water year 1994, U.S. Department of the Interior). A simple average of measured TSS values collected from November 1, 1993 through July 27, 1994 gives a TSS concentration of 5.2 mg/L. The calculated TSS concentration for Goosefare pond is 5.3 mg/L. Although the use of these values may over or underestimate the COPC partitioning to relevant compartments in the Saco River, the uncertainty in the partitioning is minor relative to the significant overestimates of COPC loadings to the river due to the use of the simple bounding estimate described above.

5.5

COPC concentrations in fish

The concentration of COPCs in fish is calculated using either a COPC-specific bioconcentration factor (BCF, for compounds with a log Kow less than 4.0), a COPC-specific bioaccumulation factor (BAF, for compounds with a log Kow greater than 4.0), or a biota-sediment accumulation factor (BSAF, for compounds which are extremely hydrophobic and listed as such in the HHRAP Appendix Tables A-3: PCDDs, PCDFs, and PCBs). Depending on which factor is appropriate, one of three equations has been used to calculate COPC concentrations in fish tissue:

C fish = Cdw ⋅ BCF fish C fish = Cdw ⋅ BAF fish C fish =

Csb ⋅ f lipid ⋅ BSAF fish OCsed

where the terms are:

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–24

Cfish Cdw Csb BCFfish BAFfish BSAFfish flipid OCsed

Concentration of COPC in fish (mg COPC/kg FW tissue); Dissolved-phase COPC concentration in water (mg COPC/L); Concentration of COPC sorbed to bed sediment (mg COPC/kg bed sediment); Bioconcentration factor for COPC in fish (l/kg); Bioaccumulation factor for COPC in fish (l/kg); Biota-to-sediment accumulation factor for COPC in fish (unitless); Fish lipid content (unitless); Fraction of organic carbon in bottom sediment (unitless).

The COPC concentrations dissolved in lake water and sorbed to bed sediments were calculated as described above; the BCFfish, BAFfish, and BSAFfish values are based on HHRAP Appendix Tables A-3 or guidance for their calculation in HHRAP Appendix A-3; a site-specific value for the BAFfish of mercury has been used in the calculations. Justification for and a derivation of the value used is described in below. An flipid value of 0.07, and a OCsed value of 0.04 were based on default values from HHRAP Appendix Table B-4-28.

5.5.1

The use of a site-specific value for the BAFfish for mercury

A critical parameter in the estimation of mercury levels in fish through the use of the HHRAP algorithms is the mercury bioaccumulation factor (BAF). This factor is the assumed ratio of the mercury concentration in fish tissue to the mercury concentration in surface water. Different BAFs may be derived for different types of fish (larger, older fish tend to have higher mercury concentrations, as do fish that are higher up on the food chain), and based on different sorts of mercury measurements (e.g., total, dissolved, or methyl mercury concentrations in surface waters). Because the potential noncancer health effects estimated to be caused by the Maine Energy facility emissions may be dominated by incremental mercury concentrations in local fish, it is essential that the value and form of this parameter provides as realistic an estimate of mercury levels in fish as possible. The transport and transformation of mercury from the ambient atmosphere to fish tissues involves a complex set of processes, many of which and not entirely well understood or well characterized in all but the most well studied systems. The bioaccumulation of Hg in fish is not a simple chemical partitioning but is dependent on a variety of parameters as the Hg progresses up the food chain to trophic level 4 fish. The transformation of mercury from inorganic species into methyl mercury and the subsequent biotransfer of methyl mercury up the food chain at a specific location is affected not only by the simple physical properties of the lake, but also by its surface and sediment chemistry and by the various species which make up its biological community. The U.S. EPA’s Mercury Report to Congress (U.S. EPA, 1997a; Vol. III, page 8-2), states that “the BAF value contains a substantial level of uncertainty.” The HHRAP guidance (Appendix Table B-4-27) specifically notes that “The COPC-specific BAF values may not accurately represent site-specific water body

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–25

conditions, because estimates of BAFs can vary, based on experimental conditions.” The Peer Review Comments on the HHRAP explicitly address this issue: “...major data gaps and limitations associated with fate and transport modeling are...noted [as]: mercury behavior in watersheds, and mercury bioaccumulation in fish...We believe that these data gaps are sufficiently large so that regulatory decisions should not be made without more detailed evaluations of these issues” U.S. EPA, 2000, page 100). The fact that mercury methylation and bioaccumulation are the subjects of major, current research efforts underlines both the scientific complexities and the modeling uncertainties associated with the process. The use of a single parameter to model the final step of this complex set of processes is a great oversimplification of the actual mechanisms that transform ionic, inorgainc mercury in surface waters into organic mercury in fish tissues. The bioaccumulation factor (BAF) recommended for methyl mercury, in Appendix Table A-3140 of the HHRAP is 6,800,000 l/kg. There are three reasons that this value and form of the mercury BAFfish will not be applied in this risk assessment. First, this value, according to the equation in Appendix Table B- 4-27, is to be applied to the total of the dissolved-phase concentrations of ionic and methyl mercury for estimating methyl mercury levels in trophic level 4 (piscivorous) fish. The reference for BAF is the 1997 U.S. EPA Mercury Study Report to Congress (U.S. EPA, 1997a). This BAF value is found specifically in Volume III, Appendix D in section D.3.4.1 “Bioaccumulation Factors Directly Estimated from Field Data – Methyl mercury in Piscivorous Fish.” The definition given for the BAF in the original document is: “average methyl mercury concentrations in piscivorous fish (trophic level 4) divided by average dissolved methyl mercury concentrations in water, accumulated by all possible routes of exposure.” This BAF is not derived from the methyl mercury concentration in trophic level 4 fish divided by dissolved total mercury concentration in water as described in the HHRAP. Hence the HHRAP-recommended value is inappropriate as used in the guidance, as it should not be applied to the sum of the methyl and ionic mercury concentrations in water. Second, the use of a mercury BAFfish based on methyl mercury concentrations in surface waters is problematic because the estimation of methyl mercury concentrations in the HHRAP is based on very simple default transformation and partitioning coefficients. Also, the measurement of methyl mercury levels in waterbodies is subject to a greater degree of analytic error than the measurement of total mercury and is more affected by seasonal and other environmental variations (U.S. EPA, 1997a). A mercury BAFfish based on total mercury levels in waterbodies provides a more reliable and stable value than one based on methyl mercury levels. Modeling of mercury bioaccumulation based on the total mercury BAF also allows easier benchmarking of estimated waterbody concentrations because these are most often measured as total mercury rather than methyl mercury. A BAF of 500,000 based on total dissolved mercury levels and trophic level 4 fish is derived in the Mercury Study Report to Congress(U.S. EPA, 1997a, Volume 3, Appendix D).

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–26

Third, and perhaps most importantly, because the transformation of mercury from inorganic forms into methyl mercury at a specific location is affected not only by the simple physical and chemical properties of the mercury compounds involved, but also by the site-specific watershed, surface water, and sediment chemistry, and by the various biological species which are present, it is preferable to use a BAF that is derived in waterbodies that are as close to the ones being evaluated as possible. The mercury BAFfish recommended in the HHRAP is derived from the results of four studies that were conducted in waterbodies that have very little in common with the two ponds evaluated in this risk assessment (U.S. EPA, 1997a). The waterbodies where the mercury BAFfish values were measured were either very large (i.e., Lake Michigan and Onondaga Lake, NY), or very far from Biddeford (i.e., Manitoba, Canada and Clear Lake, CA). As such, it is deemed prudent to use a mercury BAFfish that is based on total mercury measurements in waterbodies that are closer to those evaluated in this risk assessment than those used to develop the mercury BAFfish values found in the Mercury Study Report to Congress(U.S. EPA, 1997a). A recent study of mercury in water, sediment, and biota of small lakes in Vermont and New Hampshire (Kamman, et al., 2004) provides a very good dataset and methodology for deriving such a BAF. This study includes measurements of total mercury levels in yellow perch fillets and adjusted mercury levels to a mean fish age of 4.9 years. Measurements taken in fish filets are more representative of levels that humans would be exposed to than measurements of whole fish, and the adjustment to a standard fish age reduces the variability among the measurements due to sampling of fish with different ages. There were 49 measurements collected for total mercury in yellow perch and water in the epilimnion (the upper warmer layer of lakes), and 29 in the hypolimnion (the lower cooler layer of lakes). The mean log BAF for the epilimnion measurements was 5.25 (range: 4.72 to 6.01), and the mean log BAF for the hypolimnion measurements was 4.37 (range: 3.58 to 5.12). The overall mean log value of these samples is thus 4.91, which gives a geometric mean BAF of 82,000 for total mercury in the waterbody and suitable fish fillets. The range of total mercury BAFs from this dataset is from 3,800 to 1,020,000. The study authors note that the epilimnetic values were in excellent agreement with those found in other published studies. For this risk assessment’s baseline estimates of mercury concentrations in fish, the geometric mean BAFfish of 82,000 will be used. The effects of using other BAF values on the risk estimates will be evaluated in the uncertainty section of this report. It should be noted that the mercury BAFfish used in the 1996 Maine Energy Risk Assessment (Cambridge Environmental, 1996) was 5,500 based on a value in the 1986 U.S. EPA Superfund Public Health Evaluation Manual for total dissolved phase mercury concentration in water. This BAF will not be used because, like the value recommended within the HHRAP, it is based on national-level default data.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

5–27

6

Quantifying exposure

An exposure assessment builds upon the estimation of concentrations of COPCs in various media by defining, through various assumptions, the rates and frequencies at which receptors might breathe, ingest, and otherwise contact the media to which COPCs have migrated. Exposure assumptions are in general designed to estimate a high-end level of exposure (although not necessarily the highest potential degree of exposure). The calculation of human exposures to COPCs depends on (1) the COPC’s concentration in media relevant to the exposure scenario being considered; (2) the rate at which the individual consumes (i.e., ingests or inhales) the given medium; (3) the frequency and duration of the exposure; and in order to normalize the exposure for body size, (4) the individual’s weight. Exposures can be expressed in a generalized way by the following equation:

I gen =

C ⋅ CR ⋅ EF ⋅ ED BW ⋅ AT

where the terms are: Igen

C CR EF ED BW AT

Intake–the amount of COPC at the exchange boundary (mg/kg-day); for evaluating exposure to noncarcinogenic COPCs, the intake is referred to as average daily dose (ADD); for evaluating exposure to carcinogenic compounds, the intake is referred to as lifetime average daily dose (LADD); COPC concentration in media of concern (e.g., mg/kg for soil or mg/L for surface water); Consumption rate–the amount of contaminated medium consumed per unit time or event (e.g., kg/day for soil or L/day for water); Exposure frequency (days/year); Exposure duration (years); Average body weight of the receptor over the exposure period (kg); Averaging time–the period over which exposure is averaged (days); for carcinogens, the averaging time is 25,550 days, based on a lifetime exposure of 70 years; for noncarcinogens, averaging time equals ED (years) multiplied by 365 days per year.

Exposures to COPCs occur due to the direct inhalation of the compounds, and due to their ingestion within food, water and soils. The calculation of COPC intake rates by indirect pathways are contained in the HHRAP Appendices C-1-1 through C-1-5, with the total given in Appendix C-1-6:

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

6–1

I ind = I soil + I prod + I beef + I milk + I pork + I poultry + I eggs + I fish + I dw where the terms are: Iind Isoil Iprod Ibeef, Imilk, Ipork, Ipoultry, Ieggs, Ifish Idw

Total daily intake of COPC by indirect pathways (mg/kg-day); Daily intake of COPC from soil (mg/kg-day); Daily intake of COPC from above ground produce (mg/kg-day); Daily intake of COPC from beef, milk, pork, poultry, eggs, and fish respectively (mg/kg-day); Daily intake of COPC from drinking water (mg/kg-day).

Depending on the exposure scenario selected for analysis, some of these intake terms may be set to zero (see Table 4.1). Separate individual intake rates for adults and children are calculated using the equations below and parameters from Table 6.1. The intake of COPCs due to the incidental ingestion of soil is calculated by:

I soil =

Cs ⋅ CRsoil ⋅ Fsoil BW

where the terms are: Isoil Cs CRsoil Fsoil BW

Daily intake of COPC from soil (mg/kg-day); Average soil concentration over exposure duration (mg/kg); Consumption rate of soil (kg/day); Fraction of soil that is contaminated (unitless); Body weight (kg).

The COPC concentration in soil is calculated as described in Section 5.1 for untilled soil. Default values for CRsoil and BW for adults and children are listed in Table 6.1, and the parameter Fsoil is assumed to be 1 (HHRAP Appendix Table C-1-1). The intake of COPCs due to the ingestion of produce is calculated by:

I prod =

[(( Pd

ag

)

) (

) (

)]

+ Pv ag + Prag ⋅ CRag + Prpp ⋅ CR pp + Prbg ⋅ CRbg ⋅ F

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

6–2

where the terms are: Iprod Pd, Pv, Pr

CRag, CRpp, CRbg F

Daily intake of COPC from produce (mg/kg-day DW); Average COPC concentration in produce over exposure duration due to direct deposition, air-to-plant transfer, and root uptake respectively (mg/kg), with the subscripts ag, pp, and bg referencing above ground, protected and below ground produce respectively; Consumption rate of above ground, protected and below ground produce respectively (kg/kg-day DW); Fraction of produce that is contaminated (unitless).

The calculations of COPC concentrations in produce due to the listed mechanisms are described in section 5.2. Default values for the parameters CRag, CRpp, and CRbg for adults and children are listed in Table 6.1. The parameter F is assumed to be 1 for subsistence farmers and their children, and 0.25 for residents and subsistence fishers and their children (HHRAP Appendix Table C-1-2). The intake of COPCs due to the ingestion of beef, milk, pork, poultry, eggs, and fish is calculated by:

I i = Ai ⋅ CRi ⋅ Fi where the terms are: Ii Ai CRi Fi

Daily intake of COPC from animal tissue or product I (mg/kg-day); Average COPC concentration in animal tissue or product I over exposure duration (mg/kg FW); Consumption rate of animal tissue or product I (kg/kg-day FW); Fraction of animal tissue or product I that is contaminated (unitless);

The calculations of COPC concentrations in animal tissues are described in section 5.3. Default values for the parameters CRbeef, CRmilk, CRpork, CRchicken, and CReggs for adults and children are listed in Table 6.1, and the corresponding parameter Fi is assumed to be 1 and is applicable only to the subsistence farmer and child exposure scenarios (HHRAP Appendix Table C-1-3). The calculation of COPC concentrations in fish are described in Section 5.5. Default values for the parameters CRfish for adults and children are listed in Table 6.1, and the parameter Ffish is assumed to be 1 and is applicable only to the subsistence fisher and child exposure scenarios (HHRAP Appendix Table C-1-4). The intake of COPCs due to the ingestion of drinking water is calculated by:

I dw =

Cdw ⋅ CRdw ⋅ Fdw BW

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

6–3

where the terms are: Idw Cdw CRdw Fdw BW

Daily intake of COPC from drinking water (mg/kg-day); Dissolve-phase water concentration over exposure duration (mg/L); Consumption rate of drinking water (L/day); Fraction of drinking that is contaminated (unitless); Body weight (kg).

The COPC concentrations in surface waters are calculated as described in Section 5.4. Default values for CRdw and BW for adults and children are listed in Table 6.1, and the parameter Fdw is assumed to be 1, based on HHRAP Appendix Table C-1-5. Table 6.1 Parameters for calculation of human exposure to COPCs by indirect pathways. Parameter Body weight (kg) Consumption rates soil adult (kg/day) above ground produce (kg/kg-day DW) produce protected (kg/kg-day DW) below ground produce (kg/kg-day DW) beef (kg/kg-day DW) milk (kg/kg-day DW) pork (kg/kg-day DW) poultry (kg/kg-day DW) eggs (kg/kg-day DW) fish (kg/kg-day FW) drinking water (L/day)

BW

adult 70

child 15

CRsoil CRag CRpp CRbg CRbeef CRmilk CRpork CRchicken CReggs CRfish CRdw

0.0001 0.0003 0.00057 0.00014 0.00114 0.00842 0.00053 0.00061 0.00062 0.00117 1.4

0.0002 0.00042 0.00077 0.00022 0.00051 0.01857 0.000398 0.000425 0.000438 0.000759 0.67

The COPC exposure levels by indirect pathways which result from these calculations are used in Chapter 7 to estimate the incremental cancer risks and noncancer hazard quotients due to emissions from the Maine Energy facility. Exposures to COPCs by direct inhalation is calculated using the following equation from the HHRAP Appendix Tables C-2-1 and C-2-2:

ADI =

Ca ⋅ IR ⋅ ET ⋅ EF ⋅ ED ⋅ 0.001 BW ⋅ AT ⋅ 365

where the terms are: ADI Ca

Average daily intake of COPC via direct inhalation (mg COPC/kg-day); Total COPC air concentration over exposure duration (:g/m3);

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

6–4

IR ET EF ED BW AT 0.001 365

Inhalation rate (m3/hr); Exposure time (hr/day); Exposure frequency (days/yr); Exposure duration (yr); Body weight (kg); Averaging time (yr); Conversion factor (:g/mg); Conversion factor (days/year).

The total COPC concentrations in air are calculated as the sum of the 5-year average concentrations at the maximum impact locations using the ISCST3 modeling described in Chapter 3. Default values (as listed in HHRAP Appendix Tables C-2-1, and C-2-2) for inhalation rates, IR, are 0.30 m3/hr for children, and 0.63 m3/hr for adults; the exposure time, ET, is 24 hr/day; the exposure frequency, EF, is 350 days/year; the exposure durations, ED, are 6, 30, or 40 years (depending on the exposure scenario described in Chapter 4); the body weights, BW, are 15 kg for children, and 70 kg for adults; and the averaging times, AT, are 6, 30, or 40 years for noncancer hazard calculations, and 70 years for the calculation of cancer risks. The special exposure scenarios of nursing infants exposed to PCDDs and PCDFs through the ingestion of contaminated breast milk requires the calculation of COPC concentrations in the breast milk as described by the following equation from HHRAP Appendix Table C-3-1:

Cmilkfat

m ⋅ 1 × 10 9 ⋅ h ⋅ f 1 = 0.693 ⋅ f 2

where the terms are: Cmilkfat m h f1 f2 1×109 0.693

Concentration of COPC in milk fat of breast milk for a mother in a specific exposure scenario (pg/kg milkfat); Average maternal intake of PCDDs and PCDFs for each adult exposure scenario (mg COPC/kg BW-day); Half-life of PCDDs and PCDFs in adults (days); Fraction of ingested PCDDs and PCDFs that are stored in fat (unitless); Fraction of mother’s weight that is fat (unitless); Conversion factor (pg/mg); Conversion factor (ln2, to convert half-life to decay rate constant).

The average maternal intakes of PCDDs and PCDFs for each adult exposure scenario are calculated using the equations for adult COPC exposures described above. Default values (as listed in HHRAP Appendix Table C-3-2) for the half-life of PCDDs and PCDFs in adults, h, is 2,555 days; fraction of ingested PCDDs and PCDFs that are stored in fat, f1, is 0.9; and the fraction of mother’s weight that is fat, f2, is 0.3.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

6–5

The exposure of nursing infants to PCDDs and PCDFs is then evaluated using the following equation from HHRAP Appendix Table C-3-2:

ADDinf ant =

Cmilkfat ⋅ f 3 ⋅ f 4 ⋅ IRmilk ⋅ ED BWinf ant ⋅ AT

where the terms are: ADDinfant Cmilkfat f3 f4 IRmilk ED BWinfant AT

Average daily intake of COPC for an infant exposed to contaminated breast milk (pg COPC/kg-day); Concentration of COPC in milk fat of breast milk for a mother in a specific exposure scenario (pg/kg milkfat); Fraction of a mother’s breast milk that is fat (unitless); Fraction of ingested COPC that is absorbed (unitless); Ingestion rate of breast milk by the infant (kg/day); Exposure duration (yr); Body weight (kg); Averaging time (yr).

The concentration of COPC in the milk fat of breast milk is calculated using the equation above. Default values (as listed in HHRAP Appendix Table C-3-2) for the fraction of a mother’s breast milk that is fat, f3, is 0.04; the fraction of ingested COPC that is absorbed, f4, is 0.9; the ingestion rate of breast milk by the infant, IRmilk, is 0.8 kg/day; the exposure durations, ED, is 1 year); the body weight of the infant, BWinfant, is 10 kg; and the averaging time, AT, is 1 year.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

6–6

7

Risk and hazard characterization

The assessment of human health risks due to the main stack and odor system emissions from the Maine Energy facility are incremental in nature, and do not reflect a person’s total or cumulative risks from exposure to compounds in the environment, since there are background levels of compounds present in the atmosphere that are unrelated to operation of the Maine Energy facility. These incremental risks are examined in conjunction with the regulatory framework that treats facilities as independent entities. Two categories of incremental risks of chronic health effects have been considered: cancer and non-carcinogenic endpoints. Dose-response relationships for carcinogens are characterized by unit risk factors and potency slope factors. These factors are derived for assessing exposures via inhalation and ingestion pathways, respectively. Dose-response data for non-carcinogenic health effects are derived in a similar manner. Reference concentrations and reference doses are used to assess the likelihood of chronic (noncancer) health effects from inhalation and oral exposure, respectively. Both cancer and non-cancer risks have been evaluated for direct exposures (e.g. inhalation of directly emitted COPCs) and indirect exposures (e.g. food chain and drinking water related). The potential for long-term exposures to the COPCs to result in adverse chronic health effects has been evaluated using toxicological data principally obtained from U.S. EPA databases. A hierarchy of databases was used, with information from databases further down in the priority list being used only when data for a COPC is not available from a database higher up on the list. The preferred database is the Integrated Risk Information System (IRIS, available at: http://www.epa.gov/ngispgm3/iris/subst/index.html), followed by the Provisional Peer Reviewed Toxicity Values (PPRTVs, http://hhpprtv.ornl.gov/index.shtml) and Health Effects Assessment Summary Tables (HEAST, 1997). Some additional toxicological data has also been derived from the U.S. EPA Region III’s Risk-Based Concentration Table (U.S. EPA 2001b). Finally, the HHRAP itself recommends compound-specific toxicologic data that are derived from data in IRIS, HEAST, and other sources, by extrapolating from one route of exposure to another (e.g., deriving an ingestion potency from inhalation data). Carcinogenic risks are calculated as the product of the long-term average dose (concentration for inhalation exposure) and carcinogenic potency (unit risk). The exposure periods listed in Table 4.1 are used to assess incremental cancer risk. Individual risk estimates have been summed across compounds and exposure pathways to provide a total estimate of incremental cancer risk due to emissions from the Maine Energy facility.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–1

Estimated excess lifetime cancer risks are a conservative, high-end estimate of the incremental probability that an individual will develop cancer as a result of a specific exposure to a carcinogenic compound; they are estimated by multiplying an individual’s lifetime average daily dose (LADD) of a compound (mg/kg-day) by the compound’s cancer slope factor (CSF), (mg/kg-day)–1. The LADD for each COPC and exposure scenario has been calculated in Chapter 6 based on results contained in chapters 2 through 5; the COPC CSFs are contained in the COPC Properties Tables in Appendix III. An individual’s overall cancer risk due to a given facility’s emissions is the sum of the cancer risks from all of the compounds of concern, and thus includes potential cancers of all types.

ELCR = LADD × CSF where the terms are: ELCR LADD CSF

Excess Lifetime Cancer Risk (unitless), Lifetime Average daily dose (mg/kg-day), and Cancer Slope Factor ((mg/kg-day)–1.

Hazard quotients (HQs) have been calculated for noncarcinogenic endpoints as the ratio of the estimated dose (or concentration) due to facility-related emissions to the reference dose (or concentration) identified in the toxicity assessment. An overall hazard index (HI) has been constructed as the sum of all hazard ratios calculated for individual exposure routes and compounds. Because none of the hazard indices exceed a value of one, target-specific analyses were not conducted. Noncancer risk agents are assumed to exhibit a threshold below which no adverse effects are expected to be observed. As such, noncancer health hazards are evaluated by comparing an individual’s exposure to a compound against a reference dose (RfD) for oral exposures or a reference concentration (RfC) for inhalation exposures. The ratio of an individual’s exposure to a compound to the compound’s reference exposure level is the known as the hazard quotient for that compound and exposure:

HQ =

Ca ADD or HQ = RfD RfC

where the terms are: HQ ADD Ca RfD RfC

Hazard quotient (unitless) Average daily dose (mg/kg-day) Total COPC air concentration (mg/m3) Reference dose (mg/kg-day) Reference concentration (mg/m3)

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–2

The values for, ADD, and Ca in the above equation have been calculated in Chapter 6 based on results contained in chapters 2 through 5; the COPC-specific RfDs and RfCs are contained in the COPC Properties Tables in Appendix III. Because of the threshold assumption inherent in RfDs and RfCs, HQs below 1 are considered to be protective of human health. Additionally, an individual’s total noncarcinogenic hazard due to a given facility’s emissions is referred to as the hazard index (HI), which is calculated as the sum of the hazard quotients for all of the compounds of concern. As stated in the HHRAP guidance, the HI concept involves a considerable oversimplification of an individual’s potential to experience adverse health effects due to a facility’s emissions because it assumes that the effects of different COPCs are additive, even though they may include different (i.e., unrelated) health effects and compounds that may act synergistically or antagonistically with each other. Acute hazard quotients and indices have also been calculated as the ratio of modeled short-term exposure point concentrations in air to acute reference levels. The potential for acute effects caused by one-hour exposures have been evaluated using Acute Inhalation Exposure Criteria (AIEC), which are primarily referenced in Appendix A of the HHRAP guidance. Some additional AIEC values have been updated following the hierarchy described above. For example, some values are taken from a report issued by the Office of Environmental Health Hazard Assessment (OEHHA) at the California Environmental Protection Agency (CalEPA, OEHHA, 2000). Since the reference exposure levels (RELs) in this report supercede the formerly acute toxic effects levels (ATELs) that are currently used in the HHRAP guidance, they are also used to update the AIECs. In addition to these risk and potential hazard level calculations that are based on the HHRAP guidance, the estimated maximum average annual and 24-hour ambient air concentrations for each COPC have been calculated and compared with the City of Biddeford’s Air Toxics Ordinance ambient air limits (AALs, Biddeford, 2004). Because the Biddeford AALs are designed for direct comparison with ambient air concentrations rather than for use in calculating hazard quotients, the AALs and concentrations are both given in the tables below; HQs were not calculated based on these values. The acute hazard quotient and index evaluations have been performed using upset condition COPC emission rates (g/s) and dispersion modeling results for both normal operation “normal upset” and startup operation “startup upset” (i.e., at full and one-half normal stack flow rates respectively) as described in Sections 2.3.3 and 3.6. The 24-hour ambient air concentrations that have been estimated for comparison with the Biddeford 24-hour AALs have also been calculated based on Maine Energy emissions under normal operating conditions, as well as under upset conditions that occur during normal operation and upset conditions that occur during startup operation. For calculating the maximum 24-hour average ambient concentrations under upset conditions, it was assumed that the upset emissions and stack flows lasted for one hour and that normal emissions and flows existed for the other 23 hours. Because the upsets considered in the emissions estimations generally result in one or both of the Maine Energy boilers being shut down, the scenario essentially evaluates the average concentrations in the 24 hours leading up to,

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–3

and including the hour in which the upset occurs. During the hours after the upset, the concentrations are expected to be lower because of the shutdown of one or both of the boilers. Tables 7.1 through 7.8 contain the estimated cancer risks and hazard quotients for each COPC emitted from the Maine Energy facility and for which the appropriate toxicological data was available. For those COPCs which do not have the potential to cause both cancer or noncancer effects (based on the available toxicological data), only one set of values are given. Some COPCs are evaluated for their potential short-term effects only, therefore they are not included in the long-term effects tables. The health risk indices in Tables 7.1 and 7.2 are based on direct exposures (i.e., exposure by inhalation only) to COPC emissions. Table 7.1 evaluates potential chronic effects due to longterm direct exposures; Table 7.2 evaluates potential acute effects due to short-term exposures. The potential for long-term health effects to be caused by direct exposures to COPCs is exceedingly small, with the ELCRs below 1 × 10–8, HIs below 0.001, and all of the maximum annual average concentrations are more than 1000 times smaller than the Biddeford annual AALs. The potential for short-term health effects to be caused by direct exposures to COPCs is also small, with the overall HI equal to 0.021 for short-term exposures under upset operations, and with all of the maximum 24-hour concentrations more than 200 times lower than the Biddeford 24-hour AALs. Tables 7.3 through 7.7 contain the health risk indices estimated for indirect exposures to COPC emissions. The first of these, Table 7.3, is for the residential receptor scenario which includes exposures from homegrown vegetables. Table 7.4 contains the ELCRs and HQs for the recreational farmer scenario which, as described in Chapter 4, is evaluated at receptor locations further than 1 km from the Maine Energy facility. Tables 7.5 and 7.6 are for the recreational fisher scenarios evaluated for fish caught from Wilcox Pond and the unnamed pond on Goosefare Brook.1 Overall, the potential for indirect exposures COPCs emitted from the Maine Energy facility to cause adverse chronic health effects, is fairly small. The greatest estimated ELCR value is 3 × 10–6 for the adult recreational farmer, and the greatest HI is 0.09 for the adult recreational fisher evaluated at the unnamed pond on Goosefare Brook. Finally, Table 7.7 contains the estimated average exposure levels of adults, children, and nursing infants to PCDDs and PCDFs by indirect pathways. These exposures are on a Toxicity Equivalent Quotient (TEQ) basis. As described in the HHRAP guidance (Section 2.3.1.2), the potential for non-cancer health effects to be caused by exposure to PCDDs and PCDFs by indirect pathways is evaluated by comparing the estimated exposures with national average background exposure levels of 1 to 3 pg TEQ/kg-day for adults and 60 pg TEQ/kg-day for infants. As can be seen in Table 7.8, the incremental exposures caused by emissions from the Maine Energy facility are well below the typical national background exposures for all of the exposure scenarios. 1

As a sensitivity calculation, a screening-level analysis of Saco River fish consumption is provided in Chapter 8. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–4

Table 7.1

COPC

COPC-specific, potential chronic health risk indices based on long-term direct exposures (inhalation). Direct exposure levels are for the the maximum impact location and are used for all of the the exposure scenarios. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario. The rightmost two columns provide a direct comparison of the Biddeford annual AAL with the estimated annual average COPC concentrations (:g/m3). The sum of the ratios of the annual average concentration and the Biddeford annual AALs is given as well (in italics to indicate that it is not the sum of the concentrations above it). Annual Biddeford average ELCR ELCR HQ HQ annual concenAAL tration adult child adult child — —

Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Nickel Selenium Silver Tin Zinc Hydrogen chloride Acetone Benzene Benzoic acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Bromomethane Butanol, nButanone, 2Carbon disulfide Chloroform Chloromethane Cyclohexane Di-n-butylphthalate Dichlorobenzene, 1,2Dichlorobenzene, 1,3-

2.7E-09 7.2E-11 1.4E-09 — 3.1E-10 — 3.5E-10 — — 6.7E-10 — — — — — — 2.2E-08 — — 3.4E-11 — — — — 4.9E-08 6.5E-09 — — — —

9.5E-10 2.6E-11 5.1E-10 — 1.1E-10 — 1.3E-10 — — 2.4E-10 — — — — — — 7.9E-09 — — 1.2E-11 — — — — 1.7E-08 2.3E-09 — — — —

3.9E-05 2.8E-06 7.4E-06 6.5E-10 4.8E-07 2.4E-05 — 7.7E-07 1.7E-06 9.6E-05 1.3E-08 2.0E-08 1.2E-08 1.3E-07 3.0E-04 2.9E-05 1.8E-04 1.2E-09 1.1E-09 2.3E-07 6.3E-04 1.3E-02 3.3E-09 1.5E-05 1.3E-05 7.6E-05 1.2E-06 1.4E-09 2.4E-05 1.5E-03

8.6E-05 6.3E-06 1.6E-05 1.5E-09 1.1E-06 5.3E-05 — 1.7E-06 3.7E-06 2.1E-04 3.0E-08 4.4E-08 2.8E-08 3.0E-07 6.6E-04 6.4E-05 4.0E-04 2.6E-09 2.4E-09 5.1E-07 1.4E-03 2.8E-02 7.4E-09 3.3E-05 2.9E-05 1.7E-04 2.6E-06 3.2E-09 5.4E-05 3.4E-03

2.4E-02 2.0E-02 2.4E-02 1.2E+00 2.4E-02 — 1.2E-01 3.0E-01 6.0E-01 2.4E-01 4.8E-01 — — — 2.0E+01 — 3.8E+00 — — — — — 1.0E+03 7.0E+02 1.2E+02 2.5E+02 —— — —

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–5

1.6E-06 7.8E-08 2.0E-06 4.7E-06 6.7E-08 —7.6E-05 3.2E-07 2.5E-06 6.6E-06 3.7E-07 — — — 8.2E-03 — 7.4E-03 — — — — — 2.3E-05 1.4E-02 5.5E-03 9.4E-03 — — ——-

Table 7.1

COPC

COPC-specific, potential chronic health risk indices based on long-term direct exposures (inhalation). Direct exposure levels are for the the maximum impact location and are used for all of the the exposure scenarios. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario. The rightmost two columns provide a direct comparison of the Biddeford annual AAL with the estimated annual average COPC concentrations (:g/m3). The sum of the ratios of the annual average concentration and the Biddeford annual AALs is given as well (in italics to indicate that it is not the sum of the concentrations above it). Annual Biddeford average ELCR ELCR HQ HQ annual concenAAL tration adult child adult child — —

Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 Freon 12 Hexane Methanol Methylene chloride Methylnaphthalene, 2Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride Xylene, mXylene, oXylene, p2,3,7,8-TCDD Total PCDD/PCDF PCB Aroclor 1248 TOTAL Sum of Biddeford ratios

6.0E-08 — 5.9E-09 — — — — 2.3E-09 — — — — — — — — 3.5E-08 — — — 9.8E-09 — — — — 1.6E-09 2.7E-12 2.0E-07

2.2E-08 — 2.1E-09 — — — — 8.3E-10 — — — — — — — — 1.2E-08 — — — 3.5E-09 — — — — 5.6E-10 9.7E-13 7.1E-08

1.3E-05 1.8E-10 1.0E-05 1.3E-05 5.3E-05 1.5E-05 1.6E-04 3.2E-06 9.1E-08 5.7E-08 2.1E-08 2.1E-08 6.4E-07 9.8E-08 9.6E-06 1.1E-05 4.0E-05 8.1E-06 4.3E-06 1.7E-03 2.1E-05 2.0E-05 1.3E-05 2.0E-05 — — 1.3E-07 1.8E-02

2.9E-05 4.0E-10 2.2E-05 3.0E-05 1.2E-04 3.3E-05 3.5E-04 7.1E-06 2.0E-07 1.3E-07 4.7E-08 4.6E-08 1.4E-06 2.2E-07 2.1E-05 2.4E-05 9.0E-05 1.8E-05 9.5E-06 3.8E-03 4.6E-05 4.5E-05 3.0E-05 4.5E-05 — — 2.8E-07 4.0E-02

— — 1.0E+03 — — 2.0E+02 2.4E+01 4.1E+02

— — 1.4E-02 — — 1.4E-02 8.8E-01 1.3E-02

7.4E+01 7.4E+01 7.4E+01 3.0E+00 4.5E+01 — — — 4.0E+02 — — 1.0E+02 1.0E+03 1.0E+03 1.0E+03 1.0E-03 — 1.0E-01

4.8E-06 1.8E-06 1.8E-06 2.7E-06 2.8E-05 — — — 5.6E-02 — — 2.9E-03 2.0E-02 1.3E-02 2.0E-02 8.5E-12 — 1.2E-08 0.040

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–6

Table 7.2

COPC-specific, potential acute health risk indices based on direct short-term exposures. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. One-hour HQs are based on HHRAP Acute Inhalation Exposure Criteria; the total of the HQ values is the hazard index (HI) for the scenario. The rightmost four columns provide a comparison of the Biddeford 24-hour AALs with the estimated maximum 24–hour COPC concentrations (:g/m3) for various emission scenarios. The sum of the ratios of the 24hour average concentration and the Biddeford 24-hour AALs are given as well (in italics to indicate that they are not the sum of the concentrations above them). 1-hour HQ

COPC normal operation

Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Nickel Selenium Silver Tin Vanadium Hydrogen chloride Benzene Benzyl alcohol Bromomethane Butanone, 2Carbon disulfide Chloroform Chloromethane Dichlorobenzene, 1,2Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 Freon 12 Hexane Methanol

normal upset

24-hour estimated maximum Biddeford concentration AAL startup normal normal startup upset operation upset upset

3.3E-04 2.2E-03 2.6E-03 3.6E-02 1.2 E-05 1.4 E-05 1.5 E-05 1.2E-06 3.9E-06 2.3E-07 4.4E-08 2.2E-05 9.0E-05 1.9E-05 1.4E-04 7.9E-05 2.0E-05 8.9E-08 — 1.8E-06 4.2E-04 1.1E-04 4.8E-09 1.9E-05 1.8E-07 3.8E-05 5.9E-04 7.5E-07 3.6E-07 3.5E-07 5.1E-09 4.1E-07 7.4E-08 1.6E-08 — 5.1E-04

6.7E-06 6.7E-05 2.2E-06 1.0E-06 9.4E-05 1.6E-03 2.9E-05 4.0E-04 8.1E-04 1.5E-04 5.3E-07 — 9.7E-06 1.2E-03 3.3E-04 2.3E-08 3.1E-05 7.9E-07 5.8E-05 8.9E-04 1.3E-06 5.4E-07 5.4E-07 2.1E-08 6.1E-07 1.1E-07 2.3E-08 — 1.2E-03

8.0E-06 7.5E-05 2.7E-06 1.2E-06 1.1E-04 1.8E-03 3.4E-05 4.5E-04 9.8E-04 1.8E-04 6.4E-07 — 1.1E-05 1.4E-03 3.6E-04 2.6E-08 3.1E-05 9.0E-07 5.8E-05 8.9E-04 1.3E-06 5.4E-07 5.4E-07 2.4E-08 6.1E-07 1.1E-07 2.3E-08 — 1.2E-03

2.0E-02 3.6E-02 1.8E+00 3.6E-02 — 1.8E-01 3.0E-01 6.0E-01 3.6E-01 7.1E-01 — — — 2.7E+01 5.7E+00 — — 1.0E+03 7.0E+02 1.8E+02 3.7E+02 — — — 1.0E+03 — — 8.9E+02 1.3E+03

7.3 E-07 1.8 E-05 4.4 E-05 7.1 E-07 5.6 E-04 6.2 E-04 4.2 E-06 3.3 E-05 6.1 E-05 5.0 E-06 4.1 E-06 2.9 E-04 1.1 E-05 1.1 E-01 5.0 E-02 3.3 E-05 2.9 E-02 2.7 E-04 9.3 E-02 3.6 E-02 6.2 E-02 4.4 E-02 9.3 E-02 8.6 E-06 9.1 E-02 8.4 E-02 9.6 E-02 9.4 E-02 5.8E+00

8.5 E-07 2.6 E-05 5.8 E-05 1.2 E-06 5.9 E-04 8.8 E-04 4.3 E-06 3.5 E-05 7.7 E-05 6.3 E-06 4.7 E-06 3.0 E-04 1.3 E-05 1.2 E-01 5.2 E-02 3.8 E-05 2.9 E-02 3.1 E-04 9.5 E-02 3.7 E-02 6.3 E-02 4.5 E-02 9.5 E-02 9.7 E-06 9.3 E-02 8.6 E-02 9.7 E-02 9.5 E-02 6.1E+00

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–7

9.2 E-07 3.3 E-05 6.4 E-05 1.5 E-06 6.0 E-04 1.1 E-03 4.5 E-06 3.7 E-05 8.7 E-05 7.2 E-06 5.0 E-06 3.2 E-04 1.3 E-05 1.3 E-01 5.3 E-02 4.2 E-05 2.9 E-02 3.4 E-04 9.5 E-02 3.7 E-02 6.3 E-02 4.5 E-02 9.5 E-02 1.1 E-05 9.3 E-02 8.6 E-02 9.7 E-02 9.5 E-02 6.1E+00

Table 7.2

COPC-specific, potential acute health risk indices based on direct short-term exposures. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. One-hour HQs are based on HHRAP Acute Inhalation Exposure Criteria; the total of the HQ values is the hazard index (HI) for the scenario. The rightmost four columns provide a comparison of the Biddeford 24-hour AALs with the estimated maximum 24–hour COPC concentrations (:g/m3) for various emission scenarios. The sum of the ratios of the 24hour average concentration and the Biddeford 24-hour AALs are given as well (in italics to indicate that they are not the sum of the concentrations above them). 1-hour HQ

COPC

Methylene chloride Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride

Xylene, mXylene, oXylene, p2,3,7,8-TCDD PCB Aroclor 1248 TOTAL

normal operation 1.6E-05 — — — 3.7E-09 5.1E-07 4.7E-04 1.2E-05 3.8E-07 2.5E-05 3.1E-06 — 2.6E-07 1.5E-05 9.5E-06 1.5E-05 — — 3.0E-03

normal startup upset upset 2.9E-05 3.0E-05 — — — — — — 2.4E-08 2.7E-08 5.7E-06 6.5E-06 7.3E-04 7.3E-04 1.8E-05 1.8E-05 5.8E-07 5.8E-07 3.6E-05 3.6E-05 4.6E-06 4.6E-06 — — 4.2E-07 4.2E-07 2.4E-05 2.4E-05 1.5E-05 1.5E-05 2.4E-05 2.4E-05 — — — — 1.0E-02 1.1E-02

Sum of Biddeford ratios

Biddeford AAL 6.2E+02 1.1E+02 1.1E+02 1.1E+02 1.9E+02 6.8E+01 — — — 6.7E+02 — — 1.0E+02 1.6E+03 1.6E+03 1.6E+03 1.0E-03 1.0E-01

24-hour estimated maximum concentration normal normal startup operation upset upset 8.7 E-02 9.0 E-02 9.0 E-02 5.7 E-05 7.8 E-05 9.5 E-05 2.1 E-05 2.6 E-05 3.0 E-05 2.1 E-05 2.6 E-05 3.0 E-05 3.3 E-05 4.0 E-05 4.6 E-05 3.3 E-04 4.7 E-04 5.8 E-04 6.1 E-01 6.2 E-01 6.2 E-01 9.9 E-02 1.0 E-01 1.0 E-01 1.0 E-01 1.0 E-01 1.0 E-01 3.7 E-01 3.7 E-01 3.7 E-01 8.5 E-02 8.7 E-02 8.7 E-02 9.2 E-02 9.3 E-02 9.3 E-02 1.9 E-02 1.9 E-02 1.9 E-02 1.3 E-01 1.3 E-01 1.3 E-01 8.4 E-02 8.6 E-02 8.6 E-02 1.3 E-01 1.3 E-01 1.3 E-01 9.6 E-11 1.1 E-10 1.2 E-10 1.5 E-07 1.7 E-07 2.0 E-07 2.4 E-02 2.7 E-02 2.8 E-02

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–8

Table 7.3

Residential scenario, COPC-specific, potential chronic health risk indices due to direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Methyl mercury Nickel Selenium Silver Tin Vanadium Zinc Hydrogen chloride Acetone Benzene Benzoic acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Bromomethane Butanol, nButanone, 2- methyl ethyl ketone Carbon disulfide Chloroform Chloromethane Cyclohexane Di-n-butylphthalate

ELCR

ELCR

HQ

HQ

adult 1.5E-08 2.1E-08 1.5E-08 — 4.9E-08 — 1.5E-08 — — — 6.7E-10 — — — — — — — 9.0E-08 — — 4.8E-11 — — — — 5.1E-08 1.6E-08 — —

child 5.7E-09 1.3E-08 4.9E-09 — 1.8E-08 — 9.4E-09 — — — 2.4E-10 — — — — — — — 3.2E-08 — — 1.8E-11 — — — — 1.8E-08 5.7E-09 — —

adult 1.0E-04 9.6E-06 1.2E-04 6.2E-07 1.4E-06 6.4E-04 — 7.7E-07 1.7E-03 1.0E-04 1.1E-04 5.0E-06 2.5E-06 2.9E-06 1.7E-05 4.6E-05 3.0E-04 7.5E-05 9.0E-04 1.2E-07 1.2E-07 3.4E-07 1.2E-03 2.8E-02 3.3E-07 3.8E-05 1.0E-04 1.4E-04 1.2E-06 3.4E-09

child 2.1E-04 2.8E-05 1.9E-04 6.3E-07 2.8E-06 1.5E-03 — 1.7E-06 5.6E-03 3.0E-04 2.4E-04 8.0E-06 3.7E-06 8.2E-06 5.7E-05 7.4E-05 6.6E-04 1.4E-04 1.7E-03 1.9E-07 1.9E-07 7.7E-07 2.4E-03 5.3E-02 5.8E-07 7.4E-05 1.8E-04 2.9E-04 2.6E-06 7.0E-09

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–9

Table 7.3

Residential scenario, COPC-specific, potential chronic health risk indices due to direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Dichlorobenzene, 1,2Dichlorobenzene, 1,3Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 (trichlorofluoromethane) Freon 12 (dichlorodifluoromethane) Hexane Methanol Methylene chloride Methylnaphthalene, 2Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2- (isopropyl alcohol) Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride Xylene, mXylene, oXylene, pTotal PCDD/PCDF PCB Aroclor 1248 TOTAL

ELCR

ELCR

HQ

HQ

adult — — 8.7E-08 — 5.9E-09 — — — — 1.1E-08 — — — — — — — — 5.4E-07 — — — 3.4E-07 — — — 3.1E-07 6.7E-11 1.6E-06

child — — 3.0E-08 — 2.1E-09 — — — — 4.0E-09 — — — — — — — — 1.9E-07 — — — 1.2E-07 — — — 2.0E-07 2.4E-11 6.5E-07

adult 1.1E-04 2.5E-03 9.9E-05 7.8E-09 2.6E-05 1.9E-05 6.3E-05 3.1E-05 4.5E-04 5.0E-05 8.1E-07 1.1E-06 3.0E-07 2.7E-06 4.2E-07 9.3E-07 9.6E-06 2.2E-05 2.6E-04 1.1E-04 1.3E-05 1.7E-03 1.9E-04 2.2E-05 1.4E-05 2.2E-05 — 3.9E-06 4.0E-02

child 1.7E-04 5.0E-03 1.7E-04 1.4E-08 5.0E-05 4.0E-05 1.4E-04 6.2E-05 8.6E-04 8.9E-05 1.4E-06 1.9E-06 5.4E-07 4.9E-06 7.6E-07 1.6E-06 2.1E-05 4.3E-05 4.8E-04 2.0E-04 2.5E-05 3.8E-03 3.5E-04 4.7E-05 3.1E-05 4.7E-05 — 7.0E-06 7.8E-02

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–10

Table 7.4

Recreational farmer scenario beyond 1 km from the facility, COPC-specific, potential chronic health risk indices due to direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Methyl mercury Nickel Selenium Silver Tin Vanadium Zinc Hydrogen chloride Acetone Benzene Benzoic acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Bromomethane Butanol, nButanone, 2- methyl ethyl ketone Carbon disulfide Chloroform Chloromethane Cyclohexane Di-n-butylphthalate

ELCR

ELCR

HQ

HQ

adult 1.0E-08 2.2E-09 5.3E-09 — 1.2E-07 — 2.2E-09 — — — 6.7E-10 — — — — — — — 9.4E-08 — — 1.2E-10 — — — — 5.0E-08 6.5E-09 1.0E-08 —

child 2.6E-09 9.7E-10 1.4E-09 — 3.2E-08 — 9.6E-10 — — — 2.4E-10 — — — — — — — 3.2E-08 — — 3.5E-11 — — — — 1.8E-08 2.3E-09 3.4E-09 —

adult 6.8E-05 3.3E-06 3.2E-05 5.6E-07 2.2E-06 1.8E-04 — 7.7E-07 4.5E-04 1.8E-05 1.1E-04 2.8E-05 3.1E-05 7.3E-06 1.3E-06 9.5E-06 3.0E-04 5.1E-05 7.5E-04 5.8E-08 5.6E-08 7.7E-07 1.1E-03 2.1E-02 2.7E-07 3.3E-05 5.3E-05 7.6E-05 5.4E-05 4.6E-09

child 1.3E-04 7.8E-06 5.6E-05 2.8E-07 4.0E-06 3.2E-04 — 1.7E-06 1.1E-03 4.4E-05 2.4E-04 5.7E-05 6.8E-05 5.3E-06 3.8E-06 1.5E-05 6.6E-04 1.1E-04 1.7E-03 1.2E-07 1.1E-07 1.5E-06 2.4E-03 4.6E-02 5.9E-07 7.4E-05 1.2E-04 1.7E-04 1.2E-04 9.5E-09

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–11

Table 7.4

Recreational farmer scenario beyond 1 km from the facility, COPC-specific, potential chronic health risk indices due to direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Dichlorobenzene, 1,2Dichlorobenzene, 1,3Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 (trichlorofluoromethane) Freon 12 (dichlorodifluoromethane) Hexane Methanol Methylene chloride Methylnaphthalene, 2Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2- (isopropyl alcohol) Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride Xylene, mXylene, oXylene, pTotal PCDD/PCDF PCB Aroclor 1248 TOTAL

ELCR

ELCR

HQ

HQ

adult — — 7.7E-08 — 5.9E-09 — — — — 1.2E-08 — — — — — — — — 5.6E-07 — — — 3.6E-07 — — — 2.2E-06 9.1E-11 3.5E-06

child — — 2.7E-08 — 2.1E-09 — — — — 3.9E-09 — — — — — — — — 1.9E-07 — — — 1.2E-07 — — — 3.7E-07 2.8E-11 8.0E-07

adult 4.2E-05 2.1E-03 5.3E-05 5.9E-09 2.2E-05 1.8E-05 6.1E-05 2.8E-05 3.7E-04 3.9E-05 9.0E-07 7.3E-07 2.4E-07 2.2E-06 1.0E-06 7.4E-07 9.6E-06 1.8E-05 2.1E-04 8.7E-05 1.1E-05 1.7E-03 1.6E-04 2.1E-05 1.4E-05 2.1E-05 — 4.0E-06 2.9E-02

child 8.4E-05 4.6E-03 1.1E-04 1.3E-08 4.9E-05 4.0E-05 1.4E-04 6.2E-05 8.3E-04 8.8E-05 1.7E-06 1.6E-06 5.4E-07 4.9E-06 2.3E-06 1.6E-06 2.1E-05 4.1E-05 4.7E-04 1.9E-04 2.5E-05 3.8E-03 3.5E-04 4.7E-05 3.1E-05 4.7E-05 — 8.3E-06 6.4E-02

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–12

Table 7.5

Recreational fisher scenario at Wilcox Pond, COPC-specific, potential chronic health risk indices due to direct and indirect longterm exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Methyl mercury Nickel Selenium Silver Tin Vanadium Zinc Hydrogen chloride Acetone Benzene Benzoic acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Bromomethane Butanol, nButanone, 2- methyl ethyl ketone Carbon disulfide Chloroform Chloromethane Cyclohexane Di-n-butylphthalate

ELCR

ELCR

HQ

HQ

adult 2.2E-08 2.7E-08 3.8E-08 — 5.3E-08 — 1.5E-08 — — — 6.7E-10 — — — — — — — 9.0E-08 — — 3.7E-09 — — — — 5.1E-08 1.6E-08 — —

child 6.6E-09 1.4E-08 8.1E-09 — 1.9E-08 — 9.4E-09 — — — 2.4E-10 — — — — — — — 3.2E-08 — — 5.3E-10 — — — — 1.8E-08 5.7E-09 — —

adult 1.4E-04 1.2E-05 2.6E-04 3.0E-06 1.5E-06 6.4E-04 — 7.7E-07 1.7E-03 9.6E-02 1.3E-04 2.5E-05 1.4E-05 2.9E-06 1.7E-05 7.5E-04 3.0E-04 7.5E-05 9.0E-04 1.3E-07 1.3E-07 3.1E-05 1.2E-03 2.8E-02 3.3E-07 3.8E-05 1.0E-04 1.4E-04 1.2E-06 1.6E-07

child 2.3E-04 2.9E-05 2.9E-04 1.0E-05 2.8E-06 1.5E-03 — 1.7E-06 5.6E-03 6.7E-02 2.5E-04 2.2E-05 1.2E-05 8.2E-06 5.7E-05 5.6E-04 6.6E-04 1.4E-04 1.7E-03 2.0E-07 1.9E-07 2.2E-05 2.4E-03 5.3E-02 5.8E-07 7.4E-05 1.8E-04 2.9E-04 2.6E-06 1.1E-07

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–13

Table 7.5

Recreational fisher scenario at Wilcox Pond, COPC-specific, potential chronic health risk indices due to direct and indirect longterm exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Dichlorobenzene, 1,2Dichlorobenzene, 1,3Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 (trichlorofluoromethane) Freon 12 (dichlorodifluoromethane) Hexane Methanol Methylene chloride Methylnaphthalene, 2Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2- (isopropyl alcohol) Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride Xylene, mXylene, oXylene, pTotal PCDD/PCDF PCB Aroclor 1248 TOTAL

ELCR

ELCR

HQ

HQ

adult — — 8.7E-08 — 5.9E-09 — — — — 1.1E-08 — — — — — — — — 5.4E-07 — — — 3.4E-07 — — — 7.4E-07 7.1E-11 2.0E-06

child — — 3.0E-08 — 2.1E-09 — — — — 4.0E-09 — — — — — — — — 1.9E-07 — — — 1.2E-07 — — — 2.6E-07 2.5E-11 7.2E-07

adult 1.1E-04 2.5E-03 1.0E-04 5.2E-07 2.6E-05 1.9E-05 6.3E-05 3.1E-05 4.5E-04 5.0E-05 8.8E-07 1.4E-06 3.4E-07 3.0E-06 1.1E-06 1.1E-06 9.6E-06 2.2E-05 2.6E-04 1.1E-04 1.3E-05 1.7E-03 1.9E-04 2.2E-05 1.4E-05 2.2E-05 — 4.1E-06 1.4E-01

child 1.7E-04 5.0E-03 1.7E-04 3.7E-07 5.0E-05 4.0E-05 1.4E-04 6.2E-05 8.6E-04 9.0E-05 1.5E-06 2.1E-06 5.7E-07 5.1E-06 2.2E-06 1.9E-06 2.1E-05 4.3E-05 4.8E-04 2.0E-04 2.5E-05 3.8E-03 3.5E-04 4.7E-05 3.1E-05 4.7E-05 — 7.4E-06 1.5E-01

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–14

Table 7.6

Recreational fisher scenario at unnamed pond on Goosefare Brook, COPC-specific, potential chronic health risk indices due to direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Methyl mercury Nickel Selenium Silver Tin Vanadium Zinc Hydrogen chloride Acetone Benzene Benzoic acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Bromomethane Butanol, nButanone, 2- methyl ethyl ketone Carbon disulfide Chloroform Chloromethane Cyclohexane Di-n-butylphthalate

ELCR

ELCR

HQ

HQ

adult 2.4E-08 2.9E-08 4.4E-08 — 5.5E-08 — 1.6E-08 — — — 6.7E-10 — — — — — — — 9.0E-08 — — 4.4E-09 — — — — 5.1E-08 1.6E-08 — —

child 6.9E-09 1.5E-08 8.9E-09 — 1.9E-08 — 9.4E-09 — — — 2.4E-10 — — — — — — — 3.2E-08 — — 6.2E-10 — — — — 1.8E-08 5.7E-09 — —

adult 1.5E-04 1.2E-05 3.0E-04 3.0E-06 1.5E-06 6.4E-04 — 7.7E-07 1.7E-03 1.3E-01 1.4E-04 3.1E-05 1.7E-05 2.9E-06 1.7E-05 9.6E-04 3.0E-04 7.5E-05 9.0E-04 1.3E-07 1.3E-07 3.6E-05 1.2E-03 2.8E-02 3.3E-07 3.8E-05 1.0E-04 1.4E-04 1.2E-06 1.7E-07

child 2.4E-04 3.0E-05 3.1E-04 1.0E-05 2.8E-06 1.5E-03 — 1.7E-06 5.6E-03 8.9E-02 2.6E-04 2.6E-05 1.4E-05 8.2E-06 5.7E-05 7.1E-04 6.6E-04 1.4E-04 1.7E-03 2.0E-07 1.9E-07 2.6E-05 2.4E-03 5.3E-02 5.8E-07 7.4E-05 1.8E-04 2.9E-04 2.6E-06 1.2E-07

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–15

Table 7.6

Recreational fisher scenario at unnamed pond on Goosefare Brook, COPC-specific, potential chronic health risk indices due to direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.

COPC Dichlorobenzene, 1,2Dichlorobenzene, 1,3Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 (trichlorofluoromethane) Freon 12 (dichlorodifluoromethane) Hexane Methanol Methylene chloride Methylnaphthalene, 2Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2- (isopropyl alcohol) Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride Xylene, mXylene, oXylene, pTotal PCDD/PCDF PCB Aroclor 1248 TOTAL

ELCR

ELCR

HQ

HQ

adult — — 8.8E-08 — 5.9E-09 — — — — 1.1E-08 — — — — — — — — 5.4E-07 — — — 3.4E-07 — — — 8.4E-07 7.0E-11 2.2E-06

child — — 3.0E-08 — 2.1E-09 — — — — 4.0E-09 — — — — — — — — 1.9E-07 — — — 1.2E-07 — — — 2.7E-07 2.5E-11 7.4E-07

adult 1.1E-04 2.5E-03 1.0E-04 6.3E-07 2.6E-05 1.9E-05 6.3E-05 3.1E-05 4.5E-04 5.0E-05 9.6E-07 1.4E-06 3.7E-07 3.0E-06 1.1E-06 1.2E-06 9.6E-06 2.2E-05 2.6E-04 1.1E-04 1.3E-05 1.7E-03 1.9E-04 2.2E-05 1.4E-05 2.2E-05 — 4.0E-06 1.7E-01

child 1.7E-04 5.0E-03 1.7E-04 4.5E-07 5.0E-05 4.0E-05 1.4E-04 6.2E-05 8.6E-04 9.0E-05 1.5E-06 2.1E-06 5.9E-07 5.1E-06 2.2E-06 1.9E-06 2.1E-05 4.3E-05 4.8E-04 2.0E-04 2.5E-05 3.8E-03 3.5E-04 4.7E-05 3.1E-05 4.7E-05 — 7.3E-06 1.7E-01

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–16

Table 7.7

Exposures to PCDD/PCDFs by indirect pathways for evaluation of potential noncancer effects under six exposure scenarios. National-average, background, total daily TEQ exposures are: 1 to 3 pg/kg-day TEQ for adults, and 60 pg/kg-day TEQ for nursing infants. Nursing infant exposures are listed under the exposure scenario that describes the adult from which the infant is nursing. Indirect exposure levels resident adult

resident child

PCDD/PCDF total pg/kgday TEQ

0.019

0.0025

0.036

0.034

0.026

0.015

PCDD/PCDF total pg/kgday TEQ, nursing infants

0.072



0.96



0.41



farm adult farm child fishing adultfishing child

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

7–17

8

Uncertainty evaluation

The health risk assessment for the Maine Energy facility relies on a wide variety of data and procedures: • facility-, site-, and COPC-specific properties and parameters; • environmental transport, fate and exposure modeling assumptions; and • toxicological reference doses and slope factors. All of these elements are subject to varying degrees of uncertainty based on whether they are measured directly, calculated from basic physical and chemical principles, or extrapolated from indirect measurements. In general, the more assumptions required for determining each property or parameter, the more uncertain the resulting value. Uncertainties may arise due to a lack of basic information, the need to make predictions outside the realm of present or available knowledge, or the use of overly conservative assumptions in the calculations. The impact of each of these uncertainties on the overall risk estimates depends on the interactions among the parameters and model (e.g., fate-and-transport uncertainties for COPCs with that dominate risk estimates have a much greater impact on the overall conclusions than those for COPCs with low risk estimates). For this reason, and because the number of possible permutations is so great, the following sections are directed towards the evaluation of property, parameter, modeling, and toxicity uncertainties which have the greatest impacts on the overall risk estimates. Several specific sources of uncertainties that were expected to impact the overall risk assessment results were described in the Maine Energy RAP. Some of these did not have as large an impact on the risk assessment as anticipated because the use of facility or site-specific data resulted in modeling parameter values were not very different from those that were suggested as defaults by the HHRAP. Additional uncertainties found to have a significant influence on the final results will also be discussed. Where applicable, these discussions include quantitative comparisons of the range of possible values for critical properties and parameters and for various options aimed at reducing the overall uncertainty. Also, significant departures from HHRAP guidance methods or default assumptions are addressed with regard to their impact on the risk assessment’s overall level of uncertainty. The uncertainty evaluations in this chapter are arranged in the same basic order as in the assessment itself. The areas which are discussed are: • Facility characterization—emissions • Estimated long-term emission rates based on maximum rather than average measured COPC concentrations • Accounting for unmeasured organic compounds by a TOE factor Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–1

• • • •



• • •

Emission rates for COPCs emitted below detection limits Use of DREs to estimate some COPC emission rates Hexavalent chromium fraction in stack emissions Measured (rather than default) values for mercury speciation and partitioning fractions Air dispersion and deposition modeling • Superposition of maximum concentration and deposition values • Bounding estimates for COPC concentrations in the Saco River • Receptor grid spacing at far-field maximum impact locations Estimation of media concentrations • Non-zero values for kse in calculating average watershed soil concentrations • Site-specific, empirical BAFfish values for mercury Quantifying exposure • Use of HHRAP default ingestion rates for local animal products Risk and hazard characterization • Inherent uncertainties in toxicological data due to extrapolation from original research results • Toxicity of coplanar PCB congeners

8.1

Facility characterization—emission uncertainties 8.1.1 Estimation of long-term emission rates from maximum rather than average measured concentrations

The baseline, long-term health risk estimates given in Chapter7 are based on COPC emission rates from the Maine Energy facility that have been calculated using average measured COPC concentrations in the facility’s exhaust gases. Because the tests in which COPC concentrations were measured lasted at most a few hours, there is some uncertainty in extrapolating from shortterm test results to long term emission rates. To test the sensitivity of the risk assessment results to uncertainties in the measured emission rates and the extrapolation of short-term results to long-term estimates, the full risk calculations were performed using both baseline and high-end estimates of the COPC emission rates. Following HHRAP guidance (page 2-7), the high-end emissions estimates are calculated based on the lesser of (1) the maximum COPC concentrations detected in sampling and (2) the average concentration plus two standard deviations of the average, and also using continuous operation of the facility at the designed maximum capacity. Table 8.1 provides the results of these sensitivity calculations for the various risk estimates generated in the health risk assessment. In each case, the baseline risk estimates are paired with the sensitivity estimate developed for the maximum emission rates. All of the risk estimates increase (as would be expected), though none of them exceeds the target risk criteria ( hazard index of one, or an incremental cancer risk of ten in a million). The highest hazard index (noncancer risk) for the maximum emission rate scenario is 0.13 for the adult recreational fisher, a value seven times lower than the acceptable value. The highest incremental cancer risk of 4.9 E-06, also for the recreational fisher, is more than two times smaller than the acceptable limit. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–2

Table 8.1

Measured COPC concentration used for emission rate Average Maximum Increase Measured COPC concentration used for emission rate

Comparisons of the health risk indices as calculated using COPC emission rates based on the average measured COPC concentrations in exhaust gases and on maximum measured COPC concentrations in exhaust gases. Direct exposures for all scenarios are evaluated at the maximum impact location.

Residential scenario, indirect exposure cancer

Residential scenario, direct exposure

non-cancer

adult

child

adult

child

62%

65%

105%

100%

cancer

non-cancer

adult child adult child 2.0 E-07 7.1 E-08 1.8 E-02 4.0 E-02 1.4 E-06 5.8 E-07 2.2 E-02 3.9 E-02 2.2 E-06 9.6 E-07 4.5 E-02 7.7 E-02 2.6 E-07 9.2 E-08 3.5 E-02 7.7 E-02

Farming scenario, indirect exposure cancer adult

31%

adult

child

95%

cancer adult

non-cancer

child

adult

child

Average

3.3 E-06 7.3 E-07 1.1 E-02

Maximum

5.4 E-06 1.2 E-06 2.2 E-02 4.8 E-02 3.3 E-06 1.1 E-06 2.5 E-01 2.2 E-01

Increase

62%

57%

93%

2.5E-02

95%

Fishing scenario (Goosefare Brook), indirect exposure

non-cancer

child

31%

96%

2.0 E-06 6.7 E-07 1.5 E-01 1.3 E-01 68%

67%

65%

71%

8.1.2 Extrapolation of risks to account for un-analyzed compounds The estimated potential health risks due to emission of organic compounds from the Maine Energy facility are based on measured concentrations of these compounds in the facility’s combustion stack and odor control system outlet. However, organic compounds that are not identified by laboratory analysis (principally because they are not on the analyte lists of potentially hazardous substances) were not treated as COPCs in the risk calculations. Although there is no reason to suspect appreciable risks due to these compounds (since the lists of target analytes focus on the inclusion of potentially toxic compounds), these unknown compounds may still contribute to overall risks. The HHRAP guidance (section 2.2.1.3) recommends the calculation of a Total Organic Emission (TOE) test to account for the potential health effects of unidentified organic compounds. Although emissions from the Maine Energy facility have been analyzed for organic compounds on several occasions, because the facility does not combust hazardous waste, the specific tests for determining TOE described in the HHRAP have not been performed for the facility’s stack or odor control. Therefore a more general method is used to assess the uncertainty in the potential for organic compounds that could not be identified by the stack and odor control system testing to lead to significant estimated health risks.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–3

The first step in assessing this uncertainty is to determine the fraction of the overall estimated risk indices that are due to organic emissions from the Maine Energy stack and odor control system. The percentages of the multi-pathway risk indices due to emissions from these sources are given in Table 8.2. From examination of the values in the table, it is clear that organic emissions from the Maine Energy boiler stack have a very minor impact on the overall risk estimates, especially for the recreational fisher and farmer scenarios where indirect exposure pathways have a significant impact on the estimates. Although a full TOE test cannot be applied to account for the potential health effects of unidentified organic compounds, even if there were an equal amount of unidentified compounds as identified compounds in the stack emissions (i.e., if the TOE adjustment factor were 2), this would have only a minor impact on the overall risk assessment findings. In contrast, the organic emissions from the Maine Energy facility’s odor control system lead to a significant proportional contribution of the overall risk estimate in some cases. Therefore, an adjustment to the risk estimates to account for the possible health effects of unidentified organic compounds emitted from the odor control system might have a significant effect on the overall results. The potential level of unidentified organic compounds in the odor control system stack have been measured as part of the odor control system efficiency testing conducted in August and September 2004. Over three sampling periods, measurements of total hydrocarbon emissions were measured with a flame ionization detector (FID) and speciated organic measurements were made with a Fourier-transform infrared (FTIR) detector. Comparing a total of 931 one-minute measurements, finds that an average of 94.2% of the total hydrocarbons measured by the FID system were identified as specific compounds by the FTIR detector. It should be noted that the FTIR detector measured the concentrations of only eleven compounds (methane, propane, methanol, ethanol, propanol, butanol, o-, m-, and p-xylene, toluene, and acetic acid) rather than the 61 compounds included in the VOC testing performed in August 2003 using EPA method TO-15. If the concentrations of the extra compounds in the TO-15 analytical set were added to those in the FTIR set, then the fraction of identified compounds would be even higher. Nevertheless, using the FTIR to FID fraction of identified compounds, an adjustment factor of 1.06 (the inverse of 0.942) may be applied to the odor control system’s organic emissions to account for organic compounds that were not part of the system’s emission testing (assuming, as recommended in the HHRAP, that unidentified compounds are of equal toxicity on average to those identified). Even for the exposure scenario that had the highest fractional impact from these emissions (noncancer effects for the adult residential scenario) where odor control system organic compounds contributed 87% of the total risk, application of the adjustment factor would increase the overall estimate by only 5.2% (i.e., 0.87 × 0.06). Based on the facts that organic emissions from the Maine Energy combustion system stack contribute very little to the overall risk estimates, and the fact that the organic emissions from the odor control system are almost entirely identified, the potential health impacts of unanalyzed organic compounds are small relative to the overall risk estimates. The data that have been used to make this estimate of possible unidentified organic odor control system emissions are included in Appendix III.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–4

Table 8.2

Exposure scenario Health endpoint

Percentages of baseline risk indices that are due to organic emissions from the Maine Energy boiler stack and odor control system. Resident

Recreational fisher

Recreational farmer

adult child adult child adult child adult child adult cancer (ELCR)

non-cancer (HI)

cancer (ELCR)

non-cancer (HI)

child adult child

cancer (ELCR)

non-cancer (HI)

organic emissions from boiler stack

8.7% 7.2% 1.4% 1.5% 4.8% 5.7% 0.2% 0.5% 0.005% 0.007% 0.6% 0.5%

organic emissions from odor control system

37%

31%

87%

78%

20%

24%

12% 23% 0.032% 0.041% 44% 39%

8.1.3 Treatment of COPCs below detection limits in stack tests A frequent source of uncertainty in developing a multi-pathway risk assessment centers on the treatment of toxicologically important compounds that may be emitted from the facility being studied at concentrations below the detection limits of the analytical methods used to measure their presence. Various approaches call for assuming that such non-detected compounds (1) are not present at all in the emissions if they are never detected, (2) are present at one-half the detection limit, or (3) are present at the full detection limit. The method for treating non-detected compounds that was described in the Maine Energy RAP, and further detailed in responses to comments on the RAP, entails a combination of approaches (2) and (3) as described above. Specifically, for those COPCs that have not been detected in the recent stack or odor control system tests (but which are nevertheless included in the risk assessment), the assumed baseline emission rates are taken as one-half the detection limit of the most recent testing program. For those COPCs that have been detected in some but not all of the most recent tests, the test results in which the COPCs is not detected are averaged with the detected results at the full detection limit. Most of the compounds that were detected in some but not all of the emissions tests (and for which test results were reported as ‘non-detects’) were organic compounds emitted from both the facility’s boiler stack and odor control system. For emission estimates of metals from the facility combustion stack, only three non detects occurred among 66 measurements. For emission estimates of individual PCDD/PCDF congeners from the facility combustion stack, 87% (222 of 255) of the individual congener measurements were “detects.” For emission estimates of organic compounds from the facility combustion stack, among the compounds that were measured in recent tests, the detection rate was 69%. Emission estimates of metals and PCDD/PCDF

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–5

congeners from the odor control system were based on measured values for the facility ash, for which no substitution of detection limit-based concentrations were necessary. The high overall frequencies of detections among the various COPC categories generally result from the use of high quality (sensitive) test and analysis methods. The only COPCs that were assumed to be present at half their detection limit in the stack emissions were organic compounds: chloroform, 1,2- dichlorobenzene,, ethylbenzene, 1,1,1- trichloroethane, and oxylene. The only COPCs that were assumed to be present at half their detection limit in the odor control system emissions were also organic compounds, namely bromomethane, carbon disulfide, chloroform, chloromethane, dichlorobenzene, (1,2-, and 1,3- only), and vinyl chloride. The effects of employing different treatments of undetected COPCs on the final risk assessment results are shown in Table 8.3. The largest sensitivities are seen for the residential scenario for which contribution of organic compounds comprises a larger fraction of the risk estimates than for the other two scenarios. The cancer risk increases are primarily due to the assumed emission rates of vinyl chloride emissions from the odor control system, where none was measured. The non-cancer HI increases are primarily due to the assumed emission rates of dichlorobenzene emissions from both the stack and the odor control system. The small decreases seen with decreasing the assumed levels of COPCs that are detected in some but not all tests are due to the fact that, for the most part, the compounds with the largest risk estimates were nearly always detected. None of the percentage increases indicated in Table 8.3 make any overall risk estimates exceed or even approach target risk levels of concern. For example, the highest projected increase of 30% for the adult resident’s incremental cancer risk increases the baseline value from 1.2 E-06 to 1.6 E-06, a value that is still more than a factor of six times smaller than the target risk criterion. Table 8.3

Effects of different assumptions regarding COPC measurements in which the compound was not detected relative to the baseline assumptions described above.

Exposure scenario Health endpoint

Resident

Recreational fisher

Recreational farmer

adult child adult child adult child adult child adult child adult child cancer

non-cancer

cancer

non-cancer

cancer

non-cancer

increase with all non-detects at detection limit

30%

25%

8%

8%

16%

19%

17%

17%

10%

9%

decrease with all non-detects at one-half detection limit

-4%

-4%

-1%

-1%

-2%

-3% -0.1% -0.3% -3%

-3%

-1%

-1%

1%

2%

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–6

8.1.4 Use of a DRE to estimate some COPC emission rates Some of the COPCs included in the risk assessment were measured in the odor control system, but were not included in the list of analytes for as the combustion stack testing. Because the same source of air (from the MSW and RDF handling buildings) that is processed by the odor control system is also used for RDF combustion in the boiler, the compounds measured at the odor control system inlet are also likely present at the combustion system inlet, and may be present at the combustion stack outlet. Since most of these organic compounds are combustible, they are likely destroyed during combustion of the RDF. The stack emission rates for these organic COPCs were estimated from their concentrations at the odor control system inlet and an assumed combustion system destruction and removal efficiency (DRE) as described in section 2.3.1. The baseline DRE 99.9% was chosen as a fairly conservative value — a more realistic DRE might be as high as 99.99% (the U.S. EPA’s required DRE for hazardous waste combustors, as described in the HHRAP), of even 99.9999% (the required DRE for dioxinbearing wastes). These higher DRE values would result in lower estimated COPC emission rates, with each extra ?9” reducing the estimate by a factor of 10. Because the use of an assumed DRE was only employed in estimating emissions of some organic compounds from the combustion system stack, and because the estimated health impacts of organic emissions from this stack comprise a small portion of the overall health risks, changes in the assumed DRE value have very little impact on the risk assessment results. Table 8.4 shows the fractional changes in the overall health hazard and risk indices based on changing the assumed DRE to lower and higher values. The positive values in the table represent fractional increases in the overall multipathway ELCR and HI values that would result in assuming a lower DRE value; while negative values represent decreases in the health risk estimates due to higher assumed DRE values. For example, the first value in the table of 3 E-7 indicates that if the assumed DRE were changed from 99.9% to 90% (i.e., the modeled emissions of COPCs estimated using the DRE method from the combustion stack were to increase by a factor of 100), that the residential adult cancer risk would increase by 0.3-in-a-million. Only one higher DRE value is shown because, even at this change of DRE from 99.9% to 99.99%, the change in estimated risks is remarkably small. The largest change in health risk indices is for the noncancer effects in the residential exposure scenario where the HI increases by only 1.3% for a decrease in assumed DRE to only 90%. The reason the effect of changing the DRE is so small is that the estimated potential health impacts of organic compounds emitted from the combustion stack comprise only a minor portion of the overall risk (see section 8.1.2 above), and because emission rates for only a few of the organic compounds emitted from the stack are estimated using the DRE method. None of the potential changes associated with the DRE value would increase overall risk estimates to levels approaching or exceeding target risk criteria.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–7

Table 8.4

Exposure scenario Health

Fractional change in multipathway risk indices for different assumed values of organic compound combustion system destruction and removal efficiencies (DRE) relative to the baseline DRE value of 99.9%. Resident

Recreational fisher

adult child adult child adult child adult child cancer

non-cancer

cancer

non-cancer

Recreational farmer adult

child

cancer

adult

child

non-cancer

DRE = 90%

3 E-07 3 E-07 1 E-02 1 E-02 2 E-07 2 E-07 2 E-03 4 E-03 1 E-08 2 E-08 1 E-02 1 E-02

DRE = 99%

3 E-08 2 E-08 1 E-03 1 E-03 2 E-08 2 E-08 2 E-04 4 E-04 1 E-09 1 E-09 9 E-04 9 E-04

DRE = 99.99% -3 E-09 -2 E-09 -1 E-04 -1 E-04 -2 E-09 -2 E-09 -2 E-05 -4 E-05 -1 E-10 -1 E-10 -9 E-05 -9 E-05

8.1.5 Chromium speciation in emissions For the baseline risk assessment calculations the fraction of chromium that is assumed present in the hexavalent form is 2%. As described in section 2.3.5, this value is based on the chemical thermodynamics of chromium in combustion system, measurements of chromium speciation in laboratory scale testing, and the measured fraction of hexavalent chromium in the ash of the Maine Energy facility. Nevertheless, because hexavalent chromium concentrations have not been measured in the facility’s combustion stack emissions, there is some uncertainty regarding the actual fraction of emitted chromium present in the hexavalent form. To test the impact of different hexavalent chromium fractions on the overall health risk estimates, several other values were tested. Because the potential health impacts of hexavalent chromium are principally related to increased incremental cancer risks (i.e., the cancer risk estimates are more sensitive to changes in the hexavalent chromium fraction), the changes in estimated ELCR values have been examined for hexavalent chromium fractions of 1, 2, 5, and 10 %. Based on the analysis described in section 2.3.5, 2% is considered the baseline estimate, which is a plausible but still somewhat conservative (i.e., over predictive) estimate; 1% is the value that matches the mean measured fraction in the Maine Energy facility’s ash samples, 5% is a likely upper level fraction based on the data described in section 2.3.5, and 10 % is included in this uncertainty analysis as an extreme upper bound for the hexavalent fraction. The overall ELCR values for each exposure scenario and assumed hexavalent chromium fraction are shown in Table 8.5. The ELCR estimates are expanded to two decimal places so that some of the small changes in potential risk levels can be observed. Although hexavalent chromium is a fairly potent carcinogen, the overall estimated cancer risks for the most significantly impacted receptor (the adult farmer) would only increase by about 30% if the assumed hexavalent chromium fraction is increased by a factor of 5 relative to the baseline level. Even at this higher estimated risk level, the ELCR estimates are still well below the target criterion of 10 in a million (1 E-5).

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–8

Table 8.5

Overall excess lifetime cancer risks for different assumed fractions of emitted chromium that is present in the hexavalent form.

Exposure scenario 6

Recreational fisher (Goosefare Brook)

Recreational farmer

child

adult

child

adult

child

6.45 E-07

2.12 E-06

7.27 E-07

3.49 E-06

7.88 E-07

Resident

hexavalent chromium adult fraction 9 1% 1.54 E-06 2%

1.56 E-06

6.54 E-07

2.15 E-06

7.36 E-07

3.55 E-06

8.04 E-07

5%

1.64 E-06

6.81 E-07

2.23 E-06

7.64 E-07

3.73 E-06

8.52 E-07

10%

1.76 E-06

7.26 E-07

2.37 E-06

8.11 E-07

4.03 E-06

9.32 E-07

8.1.6 Mercury speciation and distribution in emissions The estimates of health risks due to mercury (Hg) exposure through ingestion are based on mercury speciation and distribution fractions in the facility’s stack emissions, which are in turn are calculated from detailed analytical results of the most recent mercury testing of the Maine Energy stack gas. The details of the speciation/ distribution data and associated calculations are described in section 2.3.6. Mercury is assumed to partition among three forms in stack emissions: (1) as elemental vapor-phase mercury, (2) as divalent vapor-phase mercury, and (3) as divalent particulate-phase (or particulate bound) mercury. The divalent forms are assumed to be present as mercuric chloride (HgCl2), which is among the most reactive of mercury species. This assumption is fairly conservative (health protective) because mercuric chloride is more likely to be transformed to the toxicologically important species methyl mercury than, for example, would a relatively unreactive divalent species such as mercuric oxide (HgO). The choice of mercury partitioning fractions among the three physical/chemical forms has a significant impact on the estimated health risks for recreational fishers. Because an individual’s exposure to methyl mercury is almost exclusively due to fish ingestion, the other exposure scenarios are not greatly effected by predicted Hg levels. The primary reason that the partitioning fractions have a large effect on the final hazard quotients is that divalent mercuric chloride deposits much more readily from the atmosphere than other forms of mercury. Although the baseline mercury speciation and distribution fractions in the facility’s stack emissions are based on the most recent measurements of the Maine Energy facility’s boiler (combustion) stack gases, the test used for these measurements is not officially promulgated by the U.S. EPA as a method for performing speciated mercury measurements. Therefore, some uncertainty exists relative to the validity of the modeling estimates. To test the impact of employing different mercury emission speciation fractions on the overall risk assessment results, mercury exposures and associated hazard potentials for the recreational fisher scenarios were estimated using two other mercury distributions (1) the default one from the HHRAP guidance, and (2) the distribution derived from among all of the individual Maine Energy stack tests that results in the highest fraction of HgCl2 vapor (the baseline estimates use averages of the Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–9

distributions inferred from stack testing). Table 8.6 shows the three distributions and the associated hazard indices for the recreational fisher scenarios. The differences among these HI values are not very large in part because the baseline, default, and maximum vapor-phase HgCl2 fractions are all fairly high. If the measured fraction of vapor-phase HgCl2 were much lower (fractions on the order of a few percent have been measured for some combustion sources such as coal-fired power plants), then the differences between the measured and default impacts would be much greater. It is interesting that the use of facility-specific data results in slightly higher risk estimates than those based on the HHRAP default recommendations, but even if maximum (worst-case) facility-specific data are used, the non-cancer risk estimates remain more than a ten smaller than the acceptable risk criterion. Table 8.6

Three mercury speciation distributions and the corresponding recreational fisher hazard indices. Recreational fisher scenario HI Fraction of mercury emitted as: Basis for mercury (Goosefare Brook) speciation distribution HgCl 2 adult child HgCl2 vapor Hg0 vapor particulate

Average of most recent measured values

0.08

0.15

0.77

0.15

0.13

HHRAP default distribution

0.20

0.20

0.60

0.13

0.12

Distribution with maximum measured HgCl2 vapor fraction

0.03

0.01

0.96

0.16

0.13

8.2

Air dispersion and deposition modeling uncertainties

Along with the estimation of facility emissions, the air dispersion and deposition modeling serves as one of the backbones of multi-pathway risk assessment, as all exposure routes depend on its predictions. As with all models, however, uncertainties are inherent to air dispersion modeling analyses. There is an oft-quoted statement that air dispersion models are accurate to within a “factor of two.” This statement, however, applies to the prediction of pollutant concentrations in air under fairly idealized conditions. Additional uncertainties arise in the atmospheric modeling with the introduction of terrain features, and algorithms to predict pollutant deposition, etc. While these uncertainties should be recognized in assessing the overall results of a dispersion and deposition model, the uncertainties are difficult to evaluate quantitatively, or to reduce through the use of additional site-specific data or detailed modeling. Among the uncertainties in the modeling of atmospheric dispersion and deposition is the degree to which the meteorological data used in the modeling corresponds to the conditions present at the location being modeled. On-site meteorological data are not available for the facility, so it is Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–10

not possible to directly evaluate the representativeness of the Portland Jetport data used in the modeling. From a geographic perspective, we feel that the Portland Jetport data are likely representative of conditions in Biddeford and Saco because both locations are located quite close to the coastline, which at both locations is oriented from the southwest to the northeast. Concern about the representativeness of meteorological data was raised by the Maine DEP in conjunction with the previous 1996 risk assessment of the Maine Energy facility. To address this issue, the Maine DEP provided a five-year (1989–1993) meteorological data set from the S.D. Warren facility located in Westbrook, located about 10 miles inland to the northwest of Portland. A comparison of wind roses was provided in the 1996 risk assessment. The general pattern of winds is similar between the Portland Jetport and S.D. Warren locations, although the S.D Warren site lacks the strong southerly component of the Portland Jetport data and generally exhibits lower wind speeds. Similar to the 1996 study, sensitivity modeling was conducted using the S.D. Warren meteorological data, with similar results. These runs do not consider contaminant deposition or depletion (as the S.D. Warren set lacks the parameters for deposition analysis). Peak estimates of annual average concentrations are about 17% lower using the S.D. Warren meteorological data, although peak short-term impacts are higher than those modeled using the Portland Jetport data. The highest one-hour impact prediction using the S.D. Warren meteorological data was 5% higher than that predicted using the Portland Jetport data (the basis of the risk assessment), and hence would have little impact on the overall short term risk estimate. The peak 24-hour prediction using the S.D. Warren data, however, was 1.7 times greater than the Portland Jetport-based value used in the risk assessment. Even so, the difference would have no effect on the risk assessment conclusions, as the 24-hour hazard ratio of 0.01 (including the upset scenario) is so small. Within the specific dispersion and deposition modeling of the Maine Energy facility, two conservative, simplifying assumptions have been applied in the dispersion and deposition modeling. The impact of these simplifying assumptions on the overall modeling results are discussed below.

8.2.1 Superposition of maximum concentration and deposition values To simplify the modeling of exposures of residential receptors, and the modeling of non-fish related exposures of recreational fishing receptors, the separate maximum impacts of COPCs emitted from the Maine Energy stack and odor control system and deposited by either dry or wet deposition were added together, despite the fact that these maxima occur at different physical locations. This simplification precludes the need to calculate COPC concentrations for each environmental medium at each receptor location (six interconnected media, some with several different compartments, and 1656 receptor locations). An examination of Figures 3.8 – 3.19 show the different patterns of estimated COPC atmospheric concentrations and deposition levels. The greatest differences among these figures are between the wet deposition, which is highest very close to the facility and drops off rapidly, and dry deposition, which is very low near the facility and reaches a maximum around 2 km away.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–11

Because of these differences in the geographic deposition patterns among sources (boiler stack vs. odor control system) and types of deposition, the risk estimates for any specific exposure scenario at any specific location will be lower than those estimated in the risk assessment simply because all of the maximum values are not predicted to occur at the same place. An evaluation of the implications of incorrectly superimposing modeled maxima in the health risk assessment was performed using a special receptor scenario similar to the baseline recreational farmer scenario, but simplified to exclude consumption of animal products, and placing no restriction on the distance of the receptor from the Maine Energy facility (the baseline risk estimates assume the recreational farmer lives at least 1 km away from the Maine Energy facility). Although this special scenario does not correspond to any of the HHRAP’s recommended exposure scenarios, it allows for the quantitative assessment of sensitivity to the “incorrect” superposition method used in the baseline risk assessment. The calculated cancer and non-cancer health risk indices for this scenario, as compared with the same calculations made by superposing maxima (as is done in the risk assessment), are shown in Table 8.7. The differences between the risk estimates are rather small (less that 10%) for the cancer risk indices because these risks are not heavily influenced by either wet deposition or emissions from the odor control system, both of which have their greatest impacts near the facility (i.e., with maxima only 50 m from the stack). For non-cancer indices, which are more influenced by emissions from the odor control system, the simplified addition of maximum impacts near the plant with other maxima further away produces approximately three times the more accurate receptor location-specific adding of impacts. Table 8.7

Differences in health risk indices for special receptor scenario based on the use of separate maximum COPC impact levels for each source and dispersion/deposition pathway, and based on the COPC levels at the location of the maximum total health risk impact from all of the sources and pathways.

Maximum exposure calculation method COPC exposures evaluated by superposing the separate maximum impact locations for each source dispersion/deposition Maximum COPC exposures evaluated with impacts at the same location

cancer (ELCR)

non-cancer (HI)

adult

child

adult

child

4.3 E-06

1.2 E-06

3.8 E-02

6.8 E-02

4.0 E-06

1.1 E-06

1.3 E-02

2.2 E-02

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–12

8.2.2 Bounding estimate for COPC concentrations in the Saco River As described in section 5.4.4, the dispersion models used in this risk assessment (ISC and AERMOD) are not approved to model COPC dispersion and deposition over the entire Saco River watershed (the models have only been approved for regulatory modeling applications out to a distance 50 km, the Saco River watershed extends 190 km from the facility). Therefore a very simple bounding calculation has been performed to estimate maximum possible COPC levels in the river based on the extreme assumption that all of the COPCs emitted from the facility enter the river, and that they are only diluted by the volume of water flowing in the river (i.e., not by atmospheric dilution, soil or surface water dissipation factors, etc.). This bounding estimate clearly overestimates the incremental COPC concentrations in the river that actually result from Maine Energy’s emissions because winds do not always blow the emitted compounds over the river’s watershed, and even those emissions that are blown over the watershed do may not wholly deposit to the land and water within it, and the fraction that that does deposit may not reach or remain within the water. A first order estimate of the extent to which these bounding calculations over predict COPC levels in the Saco River can be performed by examining the fraction of the time that winds blow from the Maine Energy facility towards areas within the Saco River watershed. Figure 8.1 shows an outline of the watershed and the location of the City of Biddeford (Biddeford & Saco Water Company, http://www.biddefordsacowater.com/water/). By comparing this figure with the windrose for the Portland Jetport (Figure 3.5) that shows the frequencies with which winds blow from various directions, an estimate can be made of the fraction of time emissions from the Maine Energy facility are transported over the Saco River watershed. Because the Saco River watershed covers a sector from roughly north-northwest to west-northwest, winds blowing from the south-southeast to east-southeast will send COPC emitted is Biddeford over the watershed. Summing the three windrose petals in this sector yields and estimated fraction of 11% for the time that these conditions exist. Therefore, the incremental COPC levels estimated in the Saco River are overestimated by at least a factor of 9. Additional modeling or application of other factors would further lower this percentage by accounting for the facts that not all of the emissions that are dispersed over the watershed are deposited there, and that, even when some fraction of the emitted compounds are deposited, not all of this amount would ultimately make its way into and or remain suspended within the river.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–13

Figure 8.1

The Saco River watershed. Note that the map is somewhat rotated counterclockwise from the usual position where a north/south line would be vertical; the border between Maine and New Hampshire, shown to the right of North Conway runs due north and south.

If the incremental COPC concentrations estimated in the Saco River are reduced to 11% of levels derived using the bounding estimate, the overall risk estimates decrease significantly for the residential exposure scenarios and for the recreational fisher scenario evaluated for fish caught from the river (see Section 8.3.2, further on). The decreases in the residential scenarios are due to the reduction in estimated risks due to COPCs in drinking water; the fishing scenario reductions also include concomitant reduction in estimated COPC levels in fish. Table 8.8 compares the total health risk indices for exposure scenarios in which estimated COPC levels in the Saco River have the most significant impacts. The greatest reduction is for the noncancer HIs for recreational fisher scenarios because these risk levels are dominated by exposure to methyl mercury in fish. As stated above, application of the same COPC loss algorithms to these estimates as have been used in the other watershed evaluations would further reduce the health risk estimates.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–14

Table 8.8

Comparisons of health risk indices for exposure scenarios most impacted by changes in estimated COPC concentration in the Saco River. The first row of results are based on the extreme bounding estimate for these concentrations; results in the second row are based on concentration estimates that include an additional factor to account for the fact that the wind only blows from the Maine Energy facility towards the Saco River watershed 11% of the time.

Exposure scenario Health endpoint Bounding COPC estimate

adult

Resident child adult

cancer (ELCR)

child

non-cancer (HI)

Recreational fisher (Saco River) adult child adult child cancer (ELCR)

non-cancer (HI)

1.6 E-06 6.5 E-07 4.0 E-02 7.8 E-02 2.9 E-06 8.4 E-07 5.4 E-02 8.8 E-02

Bounding COPC estimate with wind 7.4 E-07 3.6 E-07 2.9 E-02 5.9 E-02 8.8 E-07 3.8 E-07 3.0 E-02 6.0 E-02 direction factor

8.2.3 Receptor grid spacing at far-field maximum impact locations The maximum impact locations for some of the atmospheric dispersion and deposition parameters, and the location of the maximum farming scenario risk levels are rather far from the facility. For example Figure 3.11 shows the maximum annual average concentration of volumeweighted particles at approximately 2200 meters north of the facility. At this distance the radial distance between receptor locations is 100 m, but because the modeling receptor locations were set up on a polar grid with an angular spacing of 10°, the angular spacing is approximately 400 meters. To verify that the maximum modeled impact does not fall between these receptor locations, an additional modeling run was performed using a refined receptor grid in the vicinity of the projected maximum impact location. A Cartesian grid spaced at 100 m intervals was centered about the location of highest projected health risk for the farming scenario (2.2 km due north of the facility). The grid was extended 500 m in each direction. Surface-weighted, particle deposition per unit emission rate was modeled because COPCs in this category (e.g., PCDD/Fs) lead to the highest estimated health risks for the farming scenario. Contours of modeled surfaceweighted particle deposition per unit emission rate are depicted in the Figure 8.2. As can be seen, the maximum values form a narrow band to the north, indicating that the northerly radial was sufficient to capture the maximum value (the baseline modeling is spaced at 100 m intervals along the radial). Although the peak value is not precisely at the center of the figure (it is a bit further north), this peak location was captured in the identification of the maximum risk estimates as it is centered around a receptor included in the risk assessment modeling. Importantly, higher modeled values at locations to the east and west of the northerly radial do not occur in the sensitivity modeling, indicating that refinement of the receptor grid in the maximum impact area would not substantially affect the risk estimates. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–15

2700

2600

2500

2400

2300

2200

2100

2000

1900

1800

1700 -500

Figure 8.2

-400

-300

-200

-100

0

100

200

300

400

500

Modeled surface-weighted particle deposition rates per unit emission rate (g/m2 per g/s) at locations near the maximum impact location (x and y coordinates are in units of meters from the Maine Energy stack).

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–16

8.3

Estimation of media concentrations uncertainties 8.3.1 Use of non-zero kse in watershed soil concentration calculations

In developing watershed calculations, the baseline risk estimates assume that COPC concentrations in soil decrease when soil erosion occurs. This assumption differs from the HHRAP’s recommendation to ignore soil erosion losses (and hence “double count” COPCs deposited to soil). The justification for the risk assessment’s use of a calculated value for COPC loss from soils in the watershed due to erosion, kse, is given in Section 5.4.1. Briefly, the reason cited in the HHRAP guidance for using a uniform value of zero for kse is that soil eroding off a site would be replaced by soil eroding onto the site. However, this assumption cannot be applied to the evaluation of soil concentrations over an entire watershed because by definition no soil erodes into the area being considered. The primary impact on predicted health risks caused by using a non-zero value for this parameter (rather than the value of zero as recommended by the HHRAP guidance), is that it decreases the mercury exposure of the recreational fisher adult and child by a factor of about 40 – 50% by reducing the predicted amount of mercury that enters the modeled waterbodies. The overall hazard index for recreational fishers at the Goosefare Brook site are (respectively) 0.17 using calculated (non-zero) kse values, compared with an index of 0.21 employing the default value of zero for kse. While the differences between these values is significant, even with the HHRAP’s default kse value of zero, the hazard indices are well below one.

8.3.2 Bounding estimates of COPC levels in fish in the Saco River As described in Chapter 4, and section 5.4.4, bounding calculations of COPC levels in the Saco River have been performed to conservatively estimate potential exposures of by way of ingestion of drinking water from the river. As a means of assessing the upper bounds of potential COPC exposures by way of fish ingestion, these bounding estimate concentrations have been used here to estimate the potential maximum COPC levels that might be present in fish in the Saco River. Table 8.9 shows the COPC-specific ELCR and HQ values for the recreational fisher scenario that as calculated using these upper bound COPC fish concentrations. Even using the upper bound of possible COPC concentrations in the Saco River, the overall ELCR and HI are still below the criteria levels of 10–5 and 1, respectively. Based on the simple analysis of section 8.2.2, the bounding calculations overestimate COPC levels in the river by at least a factor of 9. Therefore the calculations of COPC levels in Saco River fish based on the bounding calculations are also probably overestimated by at least this amount. Reducing the bounding estimates of COPC levels in the Saco River by this factor, and calculating the fish concentrations with these lower concentrations yields ELCR and HQ values that are smaller than those found using the full multi-pathway model for the unnamed pond on the Goosefare Brook. Therefore, the potential health risks and hazards that might result from exposure to emitted COPCs in the Saco River and its fish are less than those estimated in the baseline risk assessment. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–17

Table 8.9

COPC

Recreational fisher scenario using bounding estimate of COPCs in the Saco River, COPC-specific, potential chronic health risk indices based on direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.— ELCR ELCR HQ HQ

Arsenic Beryllium Cadmium Chromium (total) Chromium (hexavalent) Copper Lead Mercury (elemental) Mercuric chloride Methyl mercury Nickel Selenium Silver Tin Vanadium Zinc Hydrogen chloride Acetone Benzene Benzoic acid Benzyl alcohol Bis(2-ethylhexyl)phthalate Bromomethane Butanol, nButanone, 2- methyl ethyl ketone Carbon disulfide Chloroform Chloromethane Cyclohexane Di-n-butylphthalate Dichlorobenzene, 1,2Dichlorobenzene, 1,3-

adult 1.6E-08 2.1E-08 2.1E-08 — 5.0E-08 — 1.5E-08 — — — 6.7E-10 — — — — — — — 1.6E-07 — — 1.1E-08 — — — — 5.1E-08 1.7E-08 — — — —

child 5.8E-09 1.3E-08 5.7E-09 — 1.8E-08 — 9.4E-09 — — — 2.4E-10 — — — — — — — 4.2E-08 — — 1.6E-09 — — — — 1.8E-08 5.9E-09 — — — —

adult 1.1E-04 9.6E-06 1.5E-04 3.0E-06 1.4E-06 6.4E-04 — 7.7E-07 1.7E-03 2.7E-03 1.1E-04 9.9E-06 5.6E-06 2.9E-06 1.7E-05 1.3E-04 3.0E-04 7.5E-05 1.7E-03 1.6E-07 1.3E-07 9.4E-05 1.3E-03 2.9E-02 3.4E-07 5.8E-05 1.1E-04 1.5E-04 1.2E-06 4.0E-07 1.9E-04 9.9E-03

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–18

child 2.1E-04 2.8E-05 2.1E-04 1.0E-05 2.8E-06 1.5E-03 — 1.7E-06 5.6E-03 2.1E-03 2.4E-04 1.1E-05 5.8E-06 8.2E-06 5.7E-05 1.3E-04 6.6E-04 1.4E-04 2.2E-03 2.2E-07 1.9E-07 6.6E-05 2.5E-03 5.4E-02 5.9E-07 8.8E-05 1.9E-04 2.9E-04 2.6E-06 2.8E-07 2.3E-04 1.0E-02

Table 8.9

COPC

Recreational fisher scenario using bounding estimate of COPCs in the Saco River, COPC-specific, potential chronic health risk indices based on direct and indirect long-term exposures. ELCR values are the excess lifetime cancer risk estimates for exposure periods shown. HQ values are the hazard quotients for each COPC; values below 1 indicate that no adverse health effects are expected to occur due to the exposure. The total of the HQ values is the hazard index (HI) for the scenario.— ELCR ELCR HQ HQ

Dichlorobenzene, 1,4Diethyl phthalate Ethylbenzene Freon 11 (trichlorofluoromethane) Freon 12 (dichlorodifluoromethane) Hexane Methanol Methylene chloride Methylnaphthalene, 2Methyl phenol, 2Methyl phenol, 3Methyl phenol, 4Naphthalene Phenol Propanol, 2- (isopropyl alcohol) Styrene Tetrachloroethene Toluene Trichloroethane, 1,1,1Trimethylbenzene, 1,2,4Vinyl chloride Xylene, mXylene, oXylene, pTotal PCDD/PCDF PCB Aroclor 1248 TOTAL

adult 2.2E-07 — 5.9E-09 — — — — 1.3E-08 — — — — — — — — 1.6E-06 — — — 3.7E-07 — — — 3.2E-07 1.8E-10 2.9E-06

child 4.8E-08 — 2.1E-09 — — — — 4.3E-09 — — — — — — — — 3.4E-07 — — — 1.3E-07 — — — 2.0E-07 4.0E-11 8.4E-07

adult 5.2E-04 7.2E-07 1.1E-04 3.1E-05 7.4E-05 3.0E-04 4.5E-04 6.0E-05 1.4E-05 1.7E-06 5.3E-07 4.8E-06 4.9E-06 1.3E-06 9.6E-06 6.1E-05 7.2E-04 3.7E-04 2.9E-05 2.0E-03 2.1E-04 2.8E-05 1.8E-05 2.8E-05 — 1.0E-05 5.4E-02

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–19

child 4.6E-04 5.1E-07 1.1E-04 4.8E-05 1.4E-04 2.5E-04 8.6E-04 9.7E-05 1.1E-05 2.3E-06 7.0E-07 6.4E-06 4.9E-06 2.0E-06 2.1E-05 7.0E-05 8.0E-04 3.8E-04 3.6E-05 4.0E-03 3.6E-04 5.2E-05 3.4E-05 5.2E-05 — 1.2E-05 8.8E-02

8.3.3 Site-specific, BAFfish values for mercury A second departure from the HHRAP default guidance with regard to modeling mercury fate and transport is in the area of the biotransfer of mercury from surface waters to fish. Rather than employing the HHRAP default value and method for estimating the mercury bioaccumulation, the baseline mercury fate and transport models used in this risk assessment use a bioaccumulation factor (BAF) more appropriate for evaluating the potential impact of mercury emitted from the Maine Energy on the local lakes and ponds near the facility. The reasons for using a non-default BAF value are described in Section 5.5.1. The BAF value used in the baseline mercury fate and transport calculations is based on a recent study of mercury in water, sediment, and biota of small lakes in Vermont and New Hampshire (Kamman, et al., 2004). This study provides a very good dataset and methodology for deriving such a BAF that is appropriate for evaluating the potential impact of mercury emitted from the Maine Energy on the local lakes and ponds near the facility. The study itself gives four different values for the mean log BAF, one for each combination of measured total mercury and methyl mercury concentrations and for measurements taken in both the hypolimnion and the epilimnion layers of the study lakes. The BAF used for the baseline calculations in this risk assessment was calculated as the overall mean log value of the total mercury results. As a means of assessing the sensitivity of the overall risk levels to the selected BAF value and basis (i.e., total mercury or methyl mercury), additional noncancer mercury hazard quotients and total exposure hazard indices for the recreational fisher scenario at the Goosefare Brook site have been calculated using all four of the BAF values given in the study. The HQ and HI values for these five BAF values and bases are shown in Table 8.10. As described in Section 5.5.1, the selection of a BAF based on total mercury levels chosen for the baseline calculations because the estimation of total mercury concentrations in surface waters is less subject to modeling uncertainties than the estimation of methyl mercury concentrations. The use of a BAF based on a combination of the hypolimnion and epilimnion data was chosen for the baseline calculations because recreational anglers are likely to catch fish from both levels depending on the fish species and time of year. The HQs calculated using methyl mercury-based BAF values and those calculated using total mercury-based BAF values are not proportional to the BAF values because the methyl mercury BAFs are applied to only the fraction of the total mercury in the water that is estimated to be present as methyl mercury. It should be noted that all mercury in fish is assumed to be present as methyl mercury under all modeling conditions. Although there is considerable variation among the results, even the maximum HI value of the sensitivity calculations is below the target noncancer risk criterion of 1.0.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–20

Table 8.10

Noncancer hazard quotients for exposure to mercury in fish, and overall total hazard indices for the recreational adult fisher scenario using the baseline mercury BAF value and basis, and the four BAF values and bases from Kamman, et al., (2004).

Mercury species used as basis for BAF Water depth/temperature for which BAF was calculated

total mercury

methyl mercury

baseline (weighted average)

hypolimnion

epilimnion

hypolimnion

epilimnion

BAF value

82,000

23,000

180,000

440,000

870,000

HQ (methyl mercury)

0.13

0.036

0.28

0.25

0.50

HI (all COPCs)

0.16

0.064

0.31

0.27

0.52

8.4

Uncertainties in quantifying exposure

The final step in estimating the public’s exposure levels to compounds emitted from the Maine Energy facility involves the use of various adult and child ingestion (and inhalation) rates for the environmental media being considered. The ingestion rates for the recreational fisher and farmer scenarios are based on USDA food consumption rates as cited in the U.S. EPA Exposure Factors Handbook (U.S. EPA 1997b). Because site-specific ingestion rates for recreational fishers and farmers near the Maine Energy facility might differ from these national-based values, there is some uncertainty in the use of these parameters in the model. While, the ingestion rates are not necessarily high for overall consumption of the modeled foods, it is difficult to quantitatively assess the degree to which these ingestion rates might over or under-estimate actual conditions, and, as applied in the risk assessment, it is assumed that all of the food ingested by these individuals is from the area impacted by the COPCs emitted from the Maine Energy facility. For example, all of the fish ingested by the recreational fisher is assumed to be from the waterbody at the location where the greatest impact is predicted from the Maine Energy facility. Likewise, for the recreational farmer scenario, all of the ingested vegetable, meat, and dairy products are assumed to have been grown or raised at the location with the greatest impact from the facility. Table 8.11 shows the assumed ingestion rates for homegrown foods and locally caught fish. The rates themselves may not be significant overestimates of local consumption rates. However, in assessing the values in the table, it should be considered that it is assumed that the foods are all of local origin, and that these rates are assumed to apply for 6 years for the child scenarios, 40 Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–21

years (total) for the adult farmer scenario, and 30 year (total) for the adult fisher scenario. It should also be noted that for the recreational farmer scenario, it is assumed that all of these foods are homegrown (including the feed used to raise livestock). As used in the exposure model, these ingestion rates are normalized to body weight and typically expressed in units such as mg/kg-day; they are expressed in Table 8.11 in more familiar units. Table 8.11

Assumed consumption rates of homegrown produce, meat, and dairy products for the recreational farmer scenario, and locally caught fish for the recreational fisher scenario. Child ingestion rate Adult ingestion rate (ounces per week, except milk) (ounces per week except milk)

Above ground produce

1.6

5.2

Protected produce

2.9

9.9

Below ground produce

0.8

2.4

Beef

1.9

19.7

Milk

1.8 quarts per week

3.9 quarts per week

Poultry

1.6

10.6

Eggs

1.6

10.7

Pork

1.5

9.2

Fish

2.8

20.3

8.5

Uncertainties in toxicologic data

Perhaps the greatest uncertainty in the risk estimates lies in the models used to predict the toxicologic potencies (especially the carcinogenic potencies) of the contaminants of interest. It is also the most difficult uncertainty to quantify and evaluate, and as such is usually treated in a manner that will overestimate potential risks. Consider the case of predicting incremental cancer risk caused by a given level of exposure to a particular compound that a person may encounter in the environment. In order to gauge whether a compound is a human carcinogen, groups of laboratory rodents are exposed, typically for most or all of their lifetimes, to very large doses of the compound — much higher doses than people typically (if ever) experience. If the doses induce an increased incidence of any type of cancer, compared to the rate observed in unexposed control animals, then the compound is deemed a carcinogen.1 Two or more of such tests with positive results suffice to label the compound a “probable human carcinogen,” even if no actual or useful data from exposed humans are available. 1

The tests must be suitable in other respects, and generally performed in accord with "good laboratory practice." Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–22

This qualitative designation of carcinogenicity is, in many cases, entirely appropriate. Rats, mice, and humans are all mammals that develop cancer from a variety of exposures, and while there are abundant differences among the three species, these differences are not so large as to suggest that compounds carcinogenic to one species will not be carcinogenic to others.2 But while the qualitative extrapolation from rodents to humans may be reasonably straightforward, the quantitative extrapolation required for risk assessment is highly uncertain. This is because the doses at which the rodents are tested are typically many thousands of times larger than doses experienced by humans. The central question is, are carcinogenic responses always proportional to dose, such that even at extremely low levels of exposure there is some risk of cancer, and that risk becomes zero only at zero dose? The answer is largely unknown. Knowledge of how specific compounds cause cancer may by helpful on a case-by-case basis, but such information is in most cases still too rudimentary to drive regulatory decision-making. The model assumed by the U.S. EPA in deriving carcinogenic potencies (as used in this risk assessment) assumes that there is a risk of cancer — however small — at any level of exposure, i.e., that a single molecule, if encountered in the critical (but not understood) manner, can cause cancer. The U.S. EPA has considered alternative types of models in which a threshold level of exposure is assumed to be necessary to cause or promote tumors, and in the case of chloroform in drinking water, has recently determined that the body of empirical evidence supports the threshold model. Similar determinations may also be likely for other compounds, casting doubt on the validity of low dose extrapolations. If other compounds are determined to behave in a manner similar to chloroform, many of the incremental cancer risk estimates within the risk assessment may be found to be zero (i.e., for exposures below the threshold level). With regard to cancer, it is assumed here that all rodent carcinogens are also human carcinogens, and that all compounds carcinogenic at high doses are also carcinogenic at vanishingly small doses. Even if these assumptions are valid, however, incremental cancer risk levels are still likely overestimated on average because the cancer slope factors that are derived by the U.S. EPA are intentionally designed to overestimate the true potency. The U.S. EPA quantifies the carcinogenic potency slope factors as upper confidence limits on mean values, meaning that the values are purposely biased on the high side to account for uncertainty in the empirical data. For similar reasons, non-cancer risks are also more likely to be overestimated than they are to be underestimated. The reference doses and concentrations developed by the U.S. EPA are designed to be levels that are likely to not cause adverse health effects. Typically, they are based on the weighted evidence of multiple studies in laboratory animals and (sometimes) humans, and 2

This discussion is necessarily simplified and greatly condensed; and there are exceptions to the rules outlined in this portion of the text. Some highly inbred strains of mice, for example, are uniquely susceptible to certain carcinogens — and other species, indeed even outbred strains of the same species, will not develop cancer at all under the same scenario of exposures. Some cancers in rats are found to occur in organs that are not present in humans. Decades of research and analysis have gone into the design, interpretation, and extrapolation of results from chronic rodent bioassays, and there are still improvements to be made. Regardless, the simplifications presented in the text are essentially accurate. Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–23

are based on levels that are observed to be free of health effects or on the lowest levels observed to cause health effects. The data from the toxicologic studies are rarely used directly, but rather are reduced in magnitude through the application of one or more safety factors that are designed to ensure that the affects-free level observed in laboratory studies also reflects a safe level of exposure for the general population. In deriving reference doses of concentrations from animal study data, safety factors are typically applied to: • • •

account for the fact that people may be more sensitive to a compound than are animals; protect individuals who might be more sensitive to the compound than the animals that were used in the study; and provide an extra degree of protection when the body of toxicologic data on a particular compound is limited.

Although there is no reason to assume a particular direction or bias with respect to any of these uncertainties, all are resolved in the direction of safety. In each case, an adjustment is made to reduce the empirical toxicity data to derive a “safer” level. In the end, the actual safe level for humans may be substantially higher than the derived reference dose or reference concentration, a factor that influences the interpretation of predicted hazard quotients in excess of one. The other factor that clouds the interpretation of data from animal studies is the fact that testing typically occurs at high doses that, in some cases, may affect or compromise bodily systems in ways that do not occur at the low doses that are characteristic of environmental exposure. For example, making animals ill through the feeding of large quantities of a compound may reduce their ability to fight off diseases and infections unrelated to the compound’s toxicity at normal, everyday levels of exposure. In general, the biases used to derive toxicologic data are likely to overestimate actual risk levels. Although it is not possible to quantify the degree of bias, it is by convention a prudent measure designed to compensate for other potential uncertainties, and overall, produce bottom-line estimates of risk such that they err on the high side of actual levels.

8.5.1 Toxicity of coplanar PCB congeners A specific factor that may add to the uncertainty of assessing the toxicological effects of exposures to various compounds is the lack of a full, detailed identification of the compound’s makeup. In this risk assessment, this condition occurs for PCBs. As noted in Section 2.2, PCB emissions from the Maine Energy facility were measured only as Aroclor 1248; the concentrations of specific congeners were not measured. Because 13PCB congeners which are referred to as either co-planar or dioxin-like, have greater carcinogenic potencies than the other congeners, the health effects of these congeners are often evaluated independently from the total PCB effects. Although the lack of congener-specific concentration data precludes the detailed evaluation of these congeners, an estimate of their potential health effects is possible. The typical composition of Aroclor 1248 has been tabulated by the U.S. Agency for Toxic Substances and Disease Registry (ATSDR, 2000). Using this profile for Aroclor 1248, the TEFs Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–24

for dioxin-like congeners from the HHRAP, and the cancer slope factor for 2,3,7,8-TCDD, an estimated cancer slope factor of 6.87 kg-d/mg can be estimated for the dioxin-like congeners of Aroclor 1248. This factor is approximately 3.4 times the slope factor of 2 kg-d/mg for total PCBs. However, because the ELCRs estimated in the risk assessment for PCB exposures are all below 10–10, the effect of this risk on the overall results is negligible.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

8–25

9

Conclusions

The multi-pathway risk assessment for the Maine Energy facility examines plausible ways that people might be exposed to potentially hazardous chemicals it releases into the environment. The analysis is developed in a conservative (health protective) manner according to regulatory guidance that considers activities and habits that might lead to elevated (high-end) exposure levels. To evaluate whether these high-end exposures might result in significant risks to human health, the exposure levels are evaluated with respect to compound-specific toxicological data. Two types of health-based evaluations are made within the risk assessment. First, the potential for each compound to increase an exposed individual’s lifetime cancer risk is assessed. Second, the likelihood that each compound might cause adverse health effects other than cancer is evaluated. The incremental, or excess, lifetime cancer risk for an exposed individual is calculated by multiplying each compound’s predicted exposure rate with its estimated potency to cause cancer in humans. The resulting cancer risk estimate is the exposed individual’s additional risk of getting cancer in his or her lifetime, above and beyond the background level that people get cancer from all causes, which is 1 in 2 for men and 1 in 3 for women. This excess risk is compared with regulatory benchmark levels to evaluate whether the estimated risk is acceptable. Historically, the Maine Department of Human Services has established an acceptable incremental cancer risk level of 1 in 100,000 (or 10 in 1,000,000). This risk level may be expressed in scientific notation as 10–5 or 1 E-5, and it represents an increase in cancer risk above the background level of 0.003% for a woman and 0.002% for a man. The potential for emitted compounds to cause noncancerous health effects is evaluated by comparing the predicted level of exposure for each compound with a level of exposure that is believed to be safe, i.e., a level that can be tolerated without risk to health (unlike incremental cancer risk, where a risk is assumed for any level of exposure). The ratio of the estimated exposure to the safe, or reference, exposure level is referred to as the compound’s hazard quotient (HQ). If a compound’s HQ is less than 1, the exposure level is less that the reference exposure level, and no adverse health effects are expected to occur. For any given scenario, the sum of all the HQs is referred to as the hazard index (HI). If the HI is less than 1, then, overall, no adverse effects are expected. Although the health effects evaluated using the hazard index include diseases that affect different organs which differ among compounds, these broad categories of potential health effects are grouped because they are evaluated in a similar manner. If the hazard ratio is greater than one, the level of exposure exceeds the level thought to be potentially harmful, and the possibility of adverse health effects might exist. However, since the reference doses and concentrations used to characterize safe values frequently embody safety factors, it is incorrect to conclude that hazard ratios greater than one will in fact correspond to Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

9–1

the actual incidence of health effects. Rather, hazard ratios exceeding one are indicators of potential concern over the possibility of adverse health effects. Two types of hazard quotients are assessed to reflect different types of exposures to compounds emitted from the Maine Energy facility. Chronic hazard quotients are calculated to assess health effects that might be associated with exposure to compounds that could occur over extended periods of time; acute hazard quotients are evaluated to gauge the nature of exposure to elevated concentrations of compounds in air that are predicted to possibly occur on an occasional basis. The overall results of the risk assessment of the Maine Energy facility are summarized in Tables 9-1 and 9-2. The total estimated lifetime incremental risks of cancer are listed in Table 9-1. These values reflect the sum of the estimates for all known or potentially carcinogenic compounds found in the Maine Energy facility emissions. The compounds and exposure pathways that contribute principally to each cancer risk estimate are also provided in Table 9-1. The incremental risk levels due to Maine Energy facility emissions are larger for the recreational farmer and fisher scenarios, reflecting the conservative nature of the risk assessment and additional indirect exposures included in these scenarios. Objectively, the lifetime incremental cancer risk estimates are quite small, especially when compared with the background (overall) risk of getting cancer. As can be seen from the values in Table 9-1, the highest excess lifetime cancer risks associated with emissions from the Maine Energy facility total an incremental risk of 4 in 1,000,000 for the recreational farmer. This estimated risk level is more than a factor of two smaller than the regulatory benchmark of 10 in 1,000,000, and it represents an increase of about only 0.001% above background cancer incidence levels. Table 9-2 presents risk estimates for compounds that, at sufficient levels of exposure, could cause adverse health effects other than cancer. The highest overall hazard index is well below 1 for both chronic (long-term) and short-term risks. Potential short-term risks have been evaluated based on both the facility’s emission levels under normal and upset operating conditions. The greatest HI is 0.2 for the fishing scenario as evaluated in the unnamed pond on the Goosefare Brook. This value is far below a level at which adverse effects might occur. Additionally, these values represent the sum of all of the hazard ratios for the individual compounds, and, strictly, hazard ratios should be separated into categories of specific health effects. More detailed information on these risk estimates, including risk estimates for each COPC under each exposure scenario, is presented in Chapter 7 of the risk assessment report. Most of the risk estimates presented in Tables 9-1 to 9-2 correspond to the estimates of emissions from the Maine Energy facility when it is operating under normal operational conditions, at full capacity, continuously throughout the year. Since the facility does not always operate at full capacity (e.g., it is shut down for periods of maintenance each year), the emission rates, and hence risk estimates, are overestimated, even accounting for potential upset conditions when emissions might be higher over short periods. Even so, a series of risk estimates is presented in the uncertainty section of the risk assessment report (see Chapter 8) based upon the highest emission rates measured during facility testing. These risk estimates tend to be about twice as large as the best-estimate values (at full operational loading) summarized in Tables 9-1 and 9-2. This factor of two does not alter conclusions relative to typical regulatory risk criteria, Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

9–2

as incremental cancer risks would remain well below 10 in 1,000,000, and hazard indices well below one. Thus, basing risk estimates on the highest measured emission rates would not lead to risk estimates of concern. Chapter 8 also contains risk estimates that have been calculated using somewhat different modeling assumptions than have been applied in the baseline estimates. Some of these sensitivity and uncertainty analyses result in slightly higher potential risk estimates, but none of them produce estimated risk indices that exceed the health-based criteria levels.

Table 9-1

Receptor

Resident Recreational Farmer

Summary of Incremental Cancer Risk Estimates a Incremental cancer risk estimate (Target limit = 10 in 1,000,000) b

Principal exposure pathways

Principal COPCs and fraction of total risk

2 in 1,000,000

drinking water homegrown produce

tetrachloroethene 34% vinyl chloride 22% PCDD/Fs 20%

4 in 1,000,000

homegrown animal products/produce

PCDD/Fs 63% tetrachloroethene 16% vinyl chloride 10%

locally caught fish PCDD/Fs 39% 2 in 1,000,000 homegrown produce tetrachloroethene 25% drinking water vinyl chloride 16% a The risk estimates shown here include risks due to both direct (Table 7.1) and indirect exposures (Tables 7.3, 7.4, and 7.6). The estimates are based on continuous operation of the facility, using compound emission rates measured under stressed operating conditions, and for the exposure pathways shown in Table ES-2. b Incremental cancer risks shown here are reported in the body of the report in scientific notation; a risk of 8 in 100,000,000 may be also shown as 8 × 10–8 or 8 E-8. Recreational Fisher

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

9–3

Table 9-2

Summary of Hazard Indices and Total Ratios to Biddeford 24–hour Ambient Air Limits Used to Evaluate Risks of Non-Cancer Health Effectsa

Receptor

Hazard Index (Acceptable limit = 1)

Principal exposure pathways

Principal COPCs and fraction of total risk

Chronic (Long-Term) Exposure Scenarios Residentb

0.08

inhalation drinking water soil ingestion

n-butanol 68% mercuric chloride 7% 1,3 dichlorobenzene 6%

Recreational Farmerb

0.06

inhalation drinking water soil ingestion

n-butanol 72% 1,3 dichlorobenzene 7% 1,2,4 trimethylbenzene 6%

Recreational Fisher

0.2

locally caught fish drinking water

methyl mercury 76% n-butanol 17% dichlorobenzene 2%

Short-Term Exposure Scenarios 1-Hour Basis Hazard Index normal operation

0.003

Inhalation

chloroform 20% methanol 17% propanol, 2- (isopropyl alcohol) 16%

1-Hour Basis Hazard Index upset conditions

0.01

Inhalation

arsenic 23% lead 16% hydrogen chloride 13%

Total of Ratios to 24-Hour AALs normal operation

0.02

Inhalation

benzene 37% methanol 18% hydrogen chloride 17%

Total of Ratios to 24-Hour AALs upset conditions

0.03

Inhalation

benzene 33% lead 21% hydrogen chloride 17%

a

The risk estimates shown here include risks due to both direct (Table 7.1) and indirect exposures (Tables 7.3, 7.4, and 7.6). The estimates are based on continuous operation of the facility, using compound emission rates measured under stressed operating conditions, and for the exposure pathways shown in Table ES-2. b The maximum Hazard Indices for these scenarios are for the child receptors

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

9–4

Table 9-2 also presents short-term risk estimates that account for occasional “upset” conditions when operations of the Maine Energy facility deviate outside of their normal ranges. As described in Chapter 2, the Maine Energy facility is designed and operated to minimize the effects of process upsets, and some “upset” conditions that occur in practice (such as facility shutdowns) actually lead to decreased long-term emissions. Consequently, the risk assessment evaluates potential acute risks associated with short-term increases in facility emissions. The upset scenarios summarized in Table 9-2 indicate a worst–case hazard index 0.01 over a 1-hour period, and a maximum sum of ratios of ambient concentrations to Biddeford 24-hour AALs of 0.03, indicating overall safety factors of 30 to 100 between (1) the ambient concentrations of COPCs that might result during a facility upset and (2) levels of potential concern. As another gauge of potential health risks due to emissions from the Maine Energy facility, the highest modeled concentrations of COPCs due to emissions from the Maine Energy facility were compared with applicable Ambient Air Limits (AALs) established by the City of Biddeford’s Air Toxics Ordinance. No predicted COPC concentrations exceed any 24-hour or annual-average AALs at any location. At the worst-case, the COPC nearest its AAL is 200 times smaller than the permissible level. In summary, • • • •

Emissions of a wide range of compounds from the Maine Energy facility have been measured; The highest expected personal exposures to these compounds by direct and indirect pathways have been modeled using methods that, in general, significantly over-predict actual exposure levels; The modeled exposures are estimated to produce less than a 0.001% increase in the risk of cancer and are well below the U.S. EPA’s reference dose and concentration levels for non-cancer effects; and The worst-case predicted concentrations of COPCs due to Maine Energy facility emissions are well below the Ambient Air Limits established by the City of Biddeford to protect public health..

Based on these findings, emissions of the Maine Energy facility present no significant risks to people living in its vicinity.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

9–5

10 References American Cancer Society (1996). Cancer Facts and Figures — 1996. Atlanta, GA: American Cancer Society. ATSDR (2000). Agency for Toxic Substances and Disease Registry , Toxicological Profile for Polychlorinated Biphenyls (PCBs) U.s. Department of Health and Human Services, Public Health Service, Atlanta, Georgia. Auer, H.A. Jr. (1978). Correlation of land use and cover with meteorological anomalies. Journal of Applied Meteorology 17:636-643. BIDDEFORD (2004). City of BIDDEFORD, Maine: Revised Code of Ordinances, Codified through Ord. No. 2004.102, adopted Nov. 3, 2004. Chapter 115 Regulated Toxic Air Pollutants Appendix I. Table I. Available at: http://library7.municode.com/gateway.dll/ME/maine/309?f=templates&fn=default.htm& npusername=10440&nppassword=MCC&npac_credentialspresent=true&vid=default, accessed October, 2005. Cambridge Environmental (1992). A Health Risk Assessment of the Waste-to-Energy Plant Proposed for Green Island, New York. April 15, 1992. Cambridge Environmental (1996). A Health Risk Assessment for the Maine Energy Recovery Company Facility, BIDDEFORD Maine. October 31, 1996. Cambridge Environmental (2002). Risk Assessment for the Evaluation of Kiln Stack Emissions and RCRA Fugitive Emissions from the Lone Star Alternative Fuels Facility, Greencastle, Indiana. Submitted to U.S. EPA Region 5, report dated April 2002. Cambridge Environmental (2004). Risk Assessment Protocol for the Evaluation of Multipathway Impacts of Emissions from the Maine Energy Recovery Company Facility in BIDDEFORD, Maine. November 2004. Earth Tech (1995). Air Quality Impact Analysis for use in Permit Application for Maine Energy Recovery Company, Biddeford Maine (Chapter6). Entropy (1992). Stationary Source Sampling Report , Maine Energy Recovery Company, Biddeford Maine. Stack testing conducted February/March, 1992, Report no.12690.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

10–1

Entropy (1994). Stationary Source Sampling Report , Maine Energy Recovery Company, Biddeford Maine. Stack testing conducted July 8–10, 1994, Report no.10912. HEAST (1997). Health Effects Summary Tables: FY-1997 Update. Washington, DC: Office of Research and Development and Office of Emergency and Remedial Response, U.S. EPA. July 1997. EPA 540/R-97-036. Howard, P.H., Boethling, R.S., Jarvis, W.F., Meylan, W.M., and Michalenko E.M. (1991). Handbook of Environmental Degradation Rates. Chelsea, MI: Lewis Publishers Inc. IRIS (2001). U.S. EPA, Integrated Risk Information System http://www.epa.gov/ngispgm3/iris/subst/index.html Jones, K.H. (2003). Harrisburg Screening Risk Assessment. Prepared for City of Harrisburg, August 22, 2003. Kamman, N., Driscoll, C.T., Estabrook, B., Ever, D.C., and Miller, E.K., (2004). Biogeochemistry of Mercury in Vermont and New Hampshire Lakes: An Assessment of Mercury in Water, Sdiment, and Biota of Vermont and New Hampshire Lakes. Comprehensive Final Project Report, May, 2004. Vermont Deportment of Environmental Conservation, Waterbury, VT. Maine Department of Inland Fisheries & Wildlife (2001). Depth map for Wilcox Pond, surveyed August 1959, revised - 2001. Available at http://www.state.me.us/ifw/pdf/depthmaps/Region%20A/L5620A.PDF. NCDC (1993). Solar and Meteorological Surface Observation Network, 1961–1990 (SAMSON). CD-ROM database, Version 1. National Climatic Data Center (U.S. Department of Commerce), September 1993. OEHHA (2000). Consolidated Table of OEHHA/ARB Approved Risk Assessment Health Values. California Environmental Protection Agency: Office of Environmental Health Hazard Assessment. October 2000. Persson A.., Dayton, D.C., and Nordin, A. (2000). Chromium speciation in combustion atmospheres. Presented at the EF Conference, Park City, Utah, May 2000. Available at http://www.chem.umu.se/dep/inorgchem/forskning/Combust/ONLINE%20Presentations_files/Chromium%20poster.pdf. Sandelin, K., Backman, R., and Nordin, A. (2001). Equilibrium distribution of arsenic, chromium, and copper in the burning of impregnated wood. The 6th International Conference on Technologies and Combustion for a Clean Environment “Clean Air.” Porto, Portugal, July 9-12 2001.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

10–2

Seinfeld, J.H. (1986). Atmospheric Chemistry and Physics of Air Pollution. New York: John Wiley & Sons. Soil Conservation Service (1982). Soil Survey of York County Maine. United States Department of Agriculture. U.S. EPA (1995). User's Guide for the Industrial Source Complex (ISC3) Dispersion Models. Volume II - Description of Model Algorithms. EPA-454/B-95-003b. U.S. EPA (1996). 40 CFR Part 51 Appendix W: Guideline on Air Quality Models, Section 8.2.3 U.S. EPA. (1997a). Mercury Study Report to Congress. Office of Air Quality Planning and Standards and Office of Research and Development. EPA-452/R-97-005. U.S. EPA (1997b). Exposure Factors Handbook. EPA/600/P-95/002Fa, August 1997. U.S. EPA (1998a). Human health risk assessment protocol for hazardous waste combustion facilities. (HHRAP) Peer review draft. Office of Solid Waste and Emergency Response. July 1998. EPA530-D-98-001A, B & C. U.S. EPA (1998b). User's Guide for the AERMOD Meteorological Preprocessor (AERMET). Revised Draft. U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, November 1998. U.S. EPA (1998c). AP 42, Fifth Edition. Compilation of Air Pollutant Emission Factors, Volume 1: Stationary Point and Area Source. Chapter 11.17 Lime Manufacturing. February 1998. Available at http://www.epa.gov/ttn/chief/ap42/ch11/final/c11s17.pdf. U.S. EPA (1999). Human health risk assessment protocol for hazardous waste combustion facilities. Peer review draft. Errata—August 2, 1999. Memorandum from Barnes Johnson, Director Economics, Methods, and Risk Analysis Division. U.S. EPA (2000). External Peer Review-Human Health Risk Assessment Protocol for Hazardous Waste Combustion Facilities—Peer Review Comments. TechLaw, Inc. Dallas. May, 2000. RCRA Docket No. F-1998-HHRA-FFFFF. U.S. EPA (2002a). User’s Guide for the AMS/EPA Regulatory Model - AERMOD. EPA-454/R-02-001. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina. U.S. EPA (2002b). User’s Guide for the AERMOD Meteorological Processor (AERMET). EPA-454/R-02-002. U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

10–3

U.S. EPA (2002c). Addendum to User's Guide for the AERMOD Meteorological Preprocessor (AERMET) (August 2002). U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards. EPA 454/R-02-002b. Wesely, M.L, P.V. Doskey, and J.D. Shannon (2002). Deposition Parameterizations for the Industrial Source Complex (ISC3) Model. Draft ANL report ANL/ER/TR–01/003, Argonne National Laboratory, Argonne, Illinois. Wischmeire, W.H., and Smith, D.D. (1978). Predicting Rainfall Erosion Losses—A Guide to Conservation Planning. Agricultural Handbook No. 537. U.S. Department of Agriculture Washington, D.C.

Cambridge Environmental Inc 58 Charles Street Cambridge, Massachusetts 02141 617-225-0810 FAX: 617-225-0813 www.CambridgeEnvironmental.com

10–4