I.1. BACKGROUND

In May of 1991, the Environmental Protection Agency (EPA) announced a scientific reassessment of the human health and exposure issues concerning dioxin and dioxin-like compounds (56 FR 50903). This reassessment has resulted in two reports: a health reassessment document (EPA, 1994), and Estimating Exposure to Dioxin-Like Compounds [this three-volume report], which expands upon a 1988 draft exposure report titled, Estimating Exposure to 2,3,7,8-TCDD (EPA, 1988). The health and exposure reassessment documents can be used together to assess potential health risks from exposure to dioxin-like compounds. In a related area, EPA has also discussed the data and methods for evaluating risks to aquatic life from 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) (EPA, 1993).

The purpose of the exposure portion of the dioxin reassessment is to describe the causes and magnitude of background exposures, and provide site-specific procedures for evaluating the incremental exposures due to specific sources of dioxin-like compounds.

In September of 1992, EPA convened workshops to review the first public drafts of the health (EPA, 1992a) and the exposure documents (EPA, 1992b). The current draft of the exposure document incorporates changes as a result of that workshop as well as other review comments.

The exposure document is presented in three volumes. Following is a summary of the material contained in each of the three volumes:

This volume includes summaries of findings from Volumes II and III. It also includes a unique section on research needs and recommendations for dioxin-like compounds.

This volume presents and evaluates information on the physical-chemical properties, environmental fate, sources, environmental levels, and background human exposures to dioxin-like compounds. It summarizes and evaluates relevant information obtained from published literature searches, EPA program offices and other Federal agencies, and published literature provided by peer reviewers of previous versions of this document. The data contained in this volume is current through 1993 with some new information published in early 1994.

This volume presents procedures for evaluating the incremental impact from sources of dioxin released into the environment. The sources covered include contaminated soils, stack emissions, and point discharges into surface water. This volume includes sections on: exposure parameters and exposure scenario development; stack emissions and atmospheric transport modeling; aquatic and terrestrial soil, sediment, and food chain modeling; demonstration of methodologies; and uncertainty evaluations including exercises on sensitivity analysis and model validation, review of Monte Carlo assessments conducted for dioxin-like compounds, and other discussions. The data contained in this volume is current through 1993 with some new information published in early 1994.

I.2. TOXICITY EQUIVALENCY FACTORS

Dioxin-like compounds are defined to include those compounds with nonzero Toxicity Equivalency Factor (TEF) values as defined in a 1989 international scheme, I-TEFs/89. This procedure was developed under the auspices of the North Atlantic Treaty Organization's Committee on Challenges of Modern Society (NATO-CCMS, 1988a; 1988b) to promote international consistency in addressing contamination involving CDDs and CDFs. EPA has adopted the I-TEFs/89 as an interim procedure for assessing the risks associated with exposures to complex mixtures of CDDs and CDFs (EPA, 1989). As shown in Table I-1, this TEF scheme assigns nonzero values to all chlorinated dibenzo-p-dioxins (CDDs) and chlorinated dibenzofurans (CDFs) with chlorine substituted in the 2,3,7,8 positions. Additionally, the analogous brominated compounds (BDDs and BDFs) and certain polychlorinated biphenyls (PCBs, see Table I-2) have recently been identified as having dioxin-like toxicity (EPA, 1994) and thus are also included in the definition of dioxin-like compounds. However, EPA has not assigned TEF values for BDDs, BDFs, and PCBs. In the case of PCBs, research on the applicability of the TEF approach is ongoing but there is not yet any formal EPA policy. The nomenclature adopted here for purposes of describing these compounds is summarized in Table I-3.

Table I-1. Toxicity Equivalency Factors (TEF) for CDDs and CDFs.

Compound TEF

Mono-, Di-, and Tri-CDDs 0

2,3,7,8-TCDD 1

Other TCDDs 0

2,3,7,8-PeCDD 0.5

Other PeCDDs 0

2,3,7,8-HxCDD 0.1

Other HxCDDs 0

2,3,7,8-HpCDD 0.01

Other HpCDDs 0

OCDD 0.001

Mono-, Di-, and Tri-CDFs 0

2,3,7,8-TCDF 0.1

Other TCDFs 0

1,2,3,7,8-PeCDF 0.05

2,3,4,7,8-PeCDF 0.5

Other PeCDFs 0

2,3,7,8-HxCDF 0.1

Other HxCDFs 0

2,3,7,8-HpCDF 0.01

Other HpCDFs 0

OCDF 0.001

Source: EPA, 1989.

Table I-2. Dioxin-Like PCBs.

IUPAC No. Congener

77 3,3',4,4'-tetra PCB

81 3,4,4',5-tetra PCB

105 2,3,3',4,4'-penta PCB

114 2,3,4,4',5-penta PCB

118 2,3',4,4',5-penta PCB

126 3,3',4,4',5-penta PCB

156 2,3,3',4,4',5-hexa PCB

157 2,3,3',4,4',5'-hexa PCB

167 2,3',4,4',5,5'-hexa PCB

169 3,3',4,4',5,5'-hexa PCB

189 2,3,3',4,4',5,5'-hepta PCB

Source: EPA, 1992a.

Table I-3. Nomenclature for dioxin-like compounds.

Term/Symbol Definition

Congener Any one particular member of the same chemical family; e.g., there are 75 congeners of chlorinated dibenzo-p-dioxins.

Homologue Group of structurally related chemicals that have the same degree of chlorination. For example, there are eight homologues of CDDs, monochlorinated through octochlorinated.

Isomer Substances that belong to the same homologous class. For example, there are 22 isomers that constitute the homologues of TCDDs.

Specific Denoted by unique chemical notation. For example, 2,4,8,9-

congener tetrachlorodibenzofuran is referred to as 2,4,8,9-TCDF.

D Symbol for homologous class: dibenzo-p-dioxin

F Symbol for homologous class: dibenzofuran

M Symbol for mono, i.e., one halogen substitution

D Symbol for di, i.e., two halogen substitution

Tr Symbol for tri, i.e., three halogen substitution

T Symbol for tetra, i.e., four halogen substitution

Pe Symbol for penta, i.e., five halogen substitution

Hx Symbol for hexa, i.e., six halogen substitution

Hp Symbol for hepta, i.e., seven halogen substitution

O Symbol for octa, i.e., eight halogen substitution

CDD Chlorinated dibenzo-p-dioxins, halogens substituted in any position

CDF Chlorinated dibenzofurans, halogens substituted in any position

PCB Polychlorinated biphenyls

2378 Halogen substitutions in the 2,3,7,8 positions

The procedure relates the toxicity of 210 structurally related individual CDD and CDF congeners and is based on a limited data base of in vivo and in vitro toxicity testing. By relating the toxicity of the 209 CDDs and CDFs to the highly-studied 2,3,7,8-TCDD, the approach simplifies the assessment of risks involving exposures to mixtures of CDDs and CDFs (EPA, 1989).

In general, the assessment of the human health risk to a mixture of CDDs and CDFs, using the TEF procedure, involves the following steps (EPA, 1989):

1. Analytical determination of the CDDs and CDFs in the sample.

2. Multiplication of congener concentrations in the sample by the TEFs in Table I-1 to express the concentration in terms of 2,3,7,8-TCDD equivalents (TEQs).

3. Summation of the products in Step 2 to obtain the total TEQs in the sample.

4. Determination of human exposure to the mixture in question, expressed in terms of TEQs.

5. Combination of exposure from step 4 with toxicity information on 2,3,7,8-TCDD to estimate risks associated with the mixture.

Samples of this calculation for several environmental mixtures are provided in EPA (1989). Also, this procedure is demonstrated in Volume III of this assessment in the context of the demonstration of the stack emission source category. The seventeen dioxin-like congeners are individually modeled from stack to exposure site. TEQ concentrations are estimated given predictions of individual congener concentrations using Steps 2 and 3 above.

I.3. OVERALL COMMENTS ON THE USE OF THE DIOXIN EXPOSURE DOCUMENT

Users of the dioxin exposure document should recognize the following:

1. This document does not present detailed procedures for evaluating multiple sources of release. However, it can be used in two ways to address this issue. Incremental impacts estimated with procedures in Volume III can be compared to background exposure estimates which are presented in Volume II. This would be a way of comparing the incremental impact of a specific source to an individual's total exposure. If the releases from multiple sources behave independently, it is possible it model them individually and then add the impacts. For example, if several stack emission sources are identified and their emissions quantified, and it is desired to evaluate the impact of all sources simultaneously, then it may be possible to model each stack emission source individually and then sum the concentrations and depositions at points of interest in the surrounding area.

2. The procedures and estimates presented in this three-volume exposure document best serve as an information source for evaluating exposures to dioxin-like compounds. This document was not generated for purposes of supporting any specific regulation. Rather, it is intended to be a general information source which Agency programs can adopt or modify as needed for their individual purposes. For example, the demonstration scenarios of Volume III were not crafted as Agency policy on "high end" or "central tendency" scenarios for evaluating land contamination, stack emissions, or effluent discharges. Rather, they were designed to illustrate the site-specific methodologies in Volume III.

3. The understanding of the exposure to dioxin-like compounds continues to expand. Despite being one of the most studied groups of organic environmental contaminants, new information is generated almost daily about dioxin-like compounds. This document is considered to be current through 1993, with some information published early in 1994 included as well. Section IV of Volume I, Executive Summary, discusses research needs for dioxin exposure evaluation.

REFERENCES FOR INTRODUCTION

NATO/CCMS (North Atlantic Treaty Organization, Committee on the Challenges of Modern Society). (1988a) International toxicity equivalency factor (I-TEF) method of risk assessment for complex mixtures of dioxins and related compounds. Report No. 176.

NATO/CCMS (North Atlantic Treaty Organization, Committee on the Challenges of Modern Society). (1988b) Scientific basis for the development of international toxicity equivalency (I-TEF) factor method of risk assessment for complex mixtures of dioxins and related compounds. Report No. 178.

U.S. Environmental Protection Agency. (1988) Estimating exposure to 2,3,7,8-TCDD. U.S. Environmental Protection Agency, Office of Health and Environmental Assessment, Washington, DC; EPA/600/6-88/005A.

U.S. Environmental Protection Agency. (1989) Interim procedures for estimating risks associated with exposures to mixtures of chlorinated dibenzo-p-dioxins and -dibenzofurans (CDDs and CDFs) and 1989 update. U.S. Environmental Protection Agency, Risk Assessment Forum, Washington, DC; EPA/625/3-89/016.

U.S. Environmental Protection Agency. (1992a) Health reassessment of dioxin-like compounds, Chapters 1-8. U.S. Environmental Protection Agency, Office of Health and Environmental Assessment, Washington, DC. EPA/600/AP-92/001a through EPA/600/AP-92/001h. August 1992 Workshop Review Draft.

U.S. Environmental Protection Agency. (1992b) Estimating Exposure to Dioxin-Like Compounds. U.S. Environmental Protection Agency, Office of Health and Environmental Assessment, Washington, DC. EPA/600/6-88/005B. August 1992 Workshop Review Draft.

U.S. Environmental Protection Agency. (1993) Interim Report on Data and Methods for Assessment of 2,3,7,8-Tetrachlorodibenzo-p-dioxin Risks to Aquatic Life and Associated Wildlife. Environmental Research Laboratory, Duluth, MN, Office of Research and Development, U.S. Environmental Protection Agency. EPA/600/R-93/055. March, 1993.

U.S. Environmental Protection Agency. (1994) Health Assessment for 2,3,7,8-TCDD and Related Compounds. Public Review Draft. EPA/600/EP-92/001.

II.1. CHEMICAL STRUCTURES AND PROPERTIES

Polychlorinated dibenzodioxins (CDDs), polychlorinated dibenzofurans (CDFs), and polychlorinated biphenyls (PCBs) are chemically classified as halogenated aromatic hydrocarbons. The chlorinated and brominated dibenzodioxins and dibenzofurans are tricyclic aromatic compounds with similar physical and chemical properties, and both classes are quite similar structurally. There are 75 possible different positional congeners of CDDs and 135 different CDF congeners. Only 7 of the 75 possible CDD congeners, and 10 of the 135 possible CDF congeners, those with chlorine substitution in the 2,3,7,8 positions, are thought to have dioxin-like toxicity. Likewise, there are 75 possible different positional congeners of BDDs and 135 different congeners of BDFs (see Table II-1). The basic structure and numbering of each chemical class is shown in Figure II-1.

There are 209 possible PCB congeners, only 11 of which are thought to have dioxin-like toxicity. These dioxin-like congeners have four or more chlorine atoms with

Figure II-1. Structure of Dioxins and Furans.

Table II-1. Possible number of positional CDD (or BDD) and CDF (or BDF) congeners

Number of Congeners

Halogen substitution CDDs (or BDDs) CDFs (or BDFs) PCBs

Mono 2 4 3

Di 10 16 12

Tri 14 28 24

Tetra 22 38 42

Penta 14 28 46

Hexa 10 16 42

Hepta 2 4 24

Octa 1 1 12

Nona 0 0 3

Deca 0 0 1

no more than one substitution in the ortho positions (positions designated 2, 2', 6 or 6' in Figure II-2). Dioxin-like PCBs are listed in Table I-2. These compounds are sometimes referred to as coplanar PCBs, since the rings can rotate into the same plane if not

blocked from rotation by ortho-substituted chlorine atoms. The physical/chemical

properties of each congener vary according to the degree and position of chlorine substitution. The basic structure and numbering of each chemical class is shown in Figure II-2.

In general, these compounds have very low water solubility, high octanol-water partition coefficients, low vapor pressure and tend to bioaccumulate. Volume II presents congener-specific values for water solubility, vapor pressure, partition coefficients and photo quantum yields.

Despite a growing body of literature from laboratory, field, and monitoring studies

Figure II-2. Structure of dioxin-like PCBs.

X = 1 to 5, Y = 1 to 5, X + Y > 1

examining the environmental fate and environmental distribution of CDDs and CDFs, the fate of these environmentally ubiquitous compounds is not yet well understood. In soil, sediment, and the water column, CDDs/CDFs are primarily associated with particulate and organic matter because of their high lipophilicity and low water solubility. In a detailed evaluation of ambient air monitoring studies in which researchers evaluated the partitioning of dioxin-like compounds between the vapor and particle phases, a principal conclusion was that the higher chlorinated congeners, the hexa through hepta congeners, were principally sorbed to airborne particulates, whereas the tetra and penta congeners significantly, if not predominantly, partition to the vapor phase. This finding is consistent with vapor/particle partitioning as theoretically modeled in Bidleman (1988). Dioxin-like compounds exhibit little potential for significant leaching or volatilization once sorbed to particulate matter. The available evidence indicates that CDDs and CDFs, particularly the tetra- and higher chlorinated congeners, are extremely stable compounds under most environmental conditions. The only environmentally significant transformation process for these congeners is believed to be photodegradation of nonsorbed species in the gaseous phase, at the soil-air or water-air interface, or in association with organic cosolvents. CDDs/CDFs entering the atmosphere are removed either by photodegradation or by deposition. Burial in-place, resuspension back into the air, or erosion of soil to water bodies appears to be the predominant fate of CDDs/CDFs sorbed to soil. CDDs/CDFs entering the water column primarily undergo sedimentation and burial. The ultimate environmental sink of CDDs/CDFs is believed to be aquatic sediments.

Little specific information exists on the environmental transport and fate of the 11 coplanar PCBs. However, the available information on the physical/chemical properties of coplanar PCBs coupled with the body of information available on the widespread occurrence and persistence of PCBs in the environment indicates that these coplanar PCBs are likely to be associated primarily with soils and sediments, and to be thermally and chemically stable. PCBs volatilize from the surfaces of soils and water bodies and are dispersed via air movement. Subsequently they can be deposited back into soil or water. In water bodies, they can be spread via sediment transport. Though not rapid processes, these mechanisms account for the widespread environmental occurrence of PCBs. Photodegradation to less chlorinated congeners followed by slow anaerobic and/or aerobic biodegradation is believed to be the principal path for destruction of PCBs.

II.2. SOURCES

Ancient human tissue sampling shows much lower CDD/F levels than found today (Ligon et al., 1989). Studies of sediment cores in lakes near industrial centers of the United States have shown that dioxins and furans were quite low until about 1920 (Czuczwa, et al., 1984; Czuczwa and Hites, 1985; Smith, et al., 1992). These studies show increases in CDD/F concentrations beginning in the 1920s and continuing until about 1970. Declining concentrations have been measured since this time. These trends cannot be explained by changes in natural processes and have been shown to correspond to chlorophenol production trends (Czuczwa and Hites, 1984). On this basis, it appears that the presence of dioxin-like compounds in the environment occurs primarily as a result of anthropogenic practices. This section will review the theories of formation and emission of these compounds, and then discuss the possible sources which can release them to the environment.

II.2.1. Theories of Formation During Combustion

The emission of CDDs and CDFs into the environment from combustion processes can be explained by three principal theories, which should not be regarded as being mutually exclusive: (1) contaminated feedstock, (2) formation from precursors, and (3) formation de novo. In general, the primary theories can be summarized as follows:

(1) The feed material to the combustor contains CDDs and CDFs and some portion survives the thermal stress imposed by the heat of the incineration or combustion process, and is subsequently emitted from the stack. While this explanation is not thought to be the principal explanation for dioxin and furan emissions from combustor sources (explanations 2 and 3 below are thought to be the predominant cause of these emissions), in fact it is the single theory best thought to explain the release of the dioxin-like, coplanar PCBs.

(2) CDDs/CDFs are ultimately formed from the thermal breakdown and molecular rearrangement of precursor compounds. Precursor compounds are chlorinated aromatic hydrocarbons having a structural resemblance to the CDD/CDF molecule. Among the precursors that have been identified are polychlorinated biphenyls (PCBs), chlorinated phenols (CPs), and chlorinated benzenes (CBs). The formation of CDDs/CDFs is believed to occur after the precursor has condensed and adsorbed onto the binding sites on the surface of fly ash particles. The active sites of the surface of fly ash particles promote the chemical reactions forming CDDs/CDFs. These reactions have been observed to be catalyzed by the presence of inorganic chlorides sorbed to the particulate. Temperature in a range of 250-450° C has been identified as a necessary condition for these reactions to occur, with either lower or higher temperatures inhibiting the process. Therefore, the precursor theory focuses on the region of the combustor that is downstream and away from the high temperature zone of the furnace or combustion chamber. This is a location where the gases and smoke derived from combustion of the organic materials have cooled during conduction through flue ducts, heat exchanger and boiler tubes, air pollution control equipment or the stack.

(3) CDDs/CDFs are synthesized de novo in the same region of the combustion process as described in (2), e.g. the so-called cool zone. In this theory, CDDs/CDFs are formed from moieties bearing little resemblance to the molecular structure of CDDs and CDFs. In broad terms, these are non-precursors and include such diverse substances as petroleum products, chlorinated plastics (PVC), non-chlorinated plastics (polystyrene), cellulose, lignin, coke, coal, particulate carbon, and hydrogen chloride gas. Formation of CDDs/CDFs requires the presence of a chlorine donor (a molecule that provides a chlorine atom to the pre-dioxin molecule) and the formation and chlorination of a chemical intermediate that is a precursor. The primary distinction between theories (2) and (3) is that theory (2) requires the presence of precursor compounds in the feed material whereas theory (3) begins with the combustion of diverse substances that are not defined as precursors, which eventually react to form precursors and eventually, dioxin-like molecules.

II.2.2. Estimates of Annual Releases of Dioxin-Like Compounds

PCBs were produced in relatively large quantities for use in such commercial products as dielectrics, hydraulic fluids, plastics and paints. They are no longer produced, but continue to be released to the environment through the use and disposal of products manufactured years ago. The chlorinated and brominated dioxins and furans, on the other hand, have never been intentionally produced other than on a laboratory scale basis for use in chemical analyses. They are, however, generated as byproducts from various combustion and chemical processes. Dioxin-like compounds are released to the environment in a variety of ways and in varying quantities depending upon the source. The dioxin like compounds have been found in all media and all parts of the world. This ubiquitous nature of these compounds suggests that multiple sources exist and that long range transport can occur. An unresolved issue is how the relative impacts from local versus distant sources compare at a particular location. Presumably in industrial areas local sources will dominate and in rural areas distant sources will dominate. However, site specific considerations such as stack height, wind patterns, magnitude of local sources, etc. could influence these comparisons.

The major identified sources of environmental release have been grouped into four major types for the purposes of this report:

• Industrial/Municipal Processes: Dioxin-like compounds can be formed through the chlorination of naturally occurring phenolic compounds such as those present in wood pulp. The formation of CDDs and CDFs resulting from the use of chlorine bleaching processes in the manufacture of bleached pulp and paper has in the past resulted in the presence of CDDs and CDFs in paper products as well as in liquid and solid wastes from this industry, although more recently this industry has made process changes to minimize CDD/CDF formation. Occasionally, municipal sewage sludge has been found to contain CDDs and CDFs.

• Chemical Manufacturing/Processing Sources: Dioxin-like compounds can be formed as by-products from the manufacture of chlorine and such chlorinated compounds as chlorinated phenols, PCBs, phenoxy herbicides, chlorinated benzenes, chlorinated aliphatic compounds, chlorinated catalysts, and halogenated diphenyl ethers. Although the manufacture of many chlorinated phenolic intermediates and products, as well as PCBs, was terminated in the late 1970s in the United States, the continued limited use and disposal of these compounds can result in releases of CDDs, CDFs, and PCBs to the environment.

• Combustion and Incineration Sources: Dioxin-like compounds can be generated and released to the environment from various combustion processes when chlorine donor compounds are present. These processes can include incineration of wastes such as municipal solid waste, sewage sludge, hospital and hazardous wastes; metallurgical processes such as high temperature steel production, smelting operations, and scrap metal recovery furnaces; and the burning of coal, wood, petroleum products, and used tires for power/energy generation.

• Reservoir Sources: The persistent and hydrophobic nature of these compounds cause them to accumulate in soils, sediments and organic matter and to persist in waste disposal sites. The dioxin-like compounds in these "reservoirs" can be redistributed by dust or sediment resuspension and transport. Such releases are not original sources in a global sense, but can be on a local scale. For example, releases may occur naturally from sediments via volatilization or via operations which disturb them such as dredging. Aerial deposition and accumulation on leaves may lead to releases during forest fires or leaf composting operations.

As awareness of these possible sources has grown in recent years, a number of changes have occurred which should reduce the release rates (Rappe, 1992). For example, releases of dioxin-like compounds have been reduced due to the switch to unleaded automobile fuels (and associated use of catalytic converters and reduction in halogenated scavenger fuel additives), process changes at pulp and paper mills, new emission standards and upgraded emission controls for incinerators, and reductions in the manufacture of chlorinated phenolic intermediates and products.

Table II-2 presents CDD and CDF air emission estimates for Germany, Austria, the United Kingdom, the Netherlands, Switzerland and the U.S. All the countries except Austria estimate that municipal waste incinerators are an important source (new emission standards in Germany indicate that the emissions from this source are now nearer the lower end of the range listed in Table II-2). Medical waste incinerators, wood burning and metal smelters/refiners also appear to be generally important sources. Rappe (1992) and Lexen et al. (1992) have identified emissions from ferrous and non-ferrous metals smelting and refining facilities as potentially the largest current source in Sweden. Rappe (1992) reported that changes in various industrial practices have lead to reductions in dioxin emissions in Sweden from 400 - 600 g of TEQ/yr in 1985 to 100 - 200 g TEQ/yr in 1991.

Table II-2. CDD and CDF air emission estimates for West Germany, Austria, United Kingdom, Netherlands,

Switzerland, and the United States.

a CR = Confidence rating. First letter is rating assigned to "production" estimate; second letter is rating assigned to "emission factor" : H = High Confidence, M = Medium, Confidence, L = Low Confidence.

b See Kraft black liquor boilers below. ^c See Sewage sludge incineration below.

d Leaded fuel production in the United States and the manufacture of motor vehicle engines requiring leaded fuel have been prohibited in the United States.

e TOTAL reflects only the total of the estimates made in this report. There are many unknowns as reflected by the number of blank cells.

f It is not known what fraction, if any, of the estimated emissions from forest fires represents a "reservoir" source. The estimated emissions may be solely the result of combustion.

statistically derive estimates of the range of uncertainty surrounding the central emission estimates. Instead, a judgement-based approach was used that assigned a factor of 10 from the low to high end of the range for the low confidence class, a factor of 5 for the medium confidence class and a factor of 2 for the high confidence class. It is emphasized that these ranges should be interpreted as judgements which are symbolic of the relative uncertainty among sources, and not statistical derivations of uncertainty. The emission factors and production values used to generate air emission estimates are illustrated in Figure II-3. Key source categories are discussed below:

• Hospital Waste Incinerators: Collectively, this may be the largest source in the United States. This is due to the facts that most of these incinerators do not rely on highly sophisticated control technologies, are high in number (over 6000 facilities) and burn high chlorine content waste. Although the dioxin emissions from these facilities are collectively large, individually they are relatively small. Therefore, local impacts may also be relatively small. However, the area of impact is an uncertain issue in general for combustors. Germany recognized the importance of these facilities several years ago and instituted emission limits which required facilities to upgrade their technology or ship waste to hazardous waste incinerators.

• Municipal Waste Incinerators: The current emissions from this category appear relatively high, but upgrading is occurring that should substantially reduce these emissions in the near future. Dioxin is also present in the ash generated from these facilities. The amount estimated to be in municipal incinerator waste ash nationally is the largest among the few source categories where estimates could be made concerning solid residues.

• Cement Kilns: EPA is currently evaluating dioxin levels in the clinker dust and stack emissions from these facilities. The preliminary information suggests that collectively these facilities could be a moderate to large source. About 16% of the facilities burn hazardous waste as an auxiliary fuel; limited data suggests that the CDD/F levels in clinker dust and stack emissions of these kilns may be significantly higher than the kilns which do not burn hazardous waste.

• Wood Burning: A large quantity of wood is burned at industrial operations, but the practice has not been well characterized. The emission estimates presented here are based on stack tests at two facilities. A number of studies have found dioxins in the emissions and ash/soot from wood fires in nonindustrial situations. The emission estimates for residential wood burners were made on the basis of two recent European studies. CDD/Fs may also be emitted during forest fires, but very little direct emission data are available for evaluating this issue. The estimates shown here were derived from tests on wood stoves under conditions of uncontrolled draft. Considering the many differences between combustion in wood stoves and forest fires, these estimates must be considered highly uncertain. Only one test has been conducted that directly measured CDD/F in smoke of forest fires (Clement and Tashiro, 1991). Low levels were detected, but the authors caution that some portion of these emissions could represent resuspended material from aerial deposits rather than originally formed material. The theory that much of today's body burden could be due to natural sources (such as forest fires) has been largely discounted by testing of ancient tissues which show levels much lower than those found today (Ligon et al. 1989).

• Metals Industry: Secondary smelters which recover metal from waste products such as scrap automobiles have the potential for dioxin formation due to chlorine in the plastic in the feed material. Processes in the primary metals industry, such as sintering of iron ore, have also been identified as potential sources. Germany (see Table II-2) has identified the metals industry as potentially one of the most important. Table II-3 estimates moderate emissions for secondary copper smelting (based on testing at only one facility) and relatively low emissions for secondary lead smelting (based on testing at three facilities). No data are available to estimate emissions from other secondary smelters or primary smelters. Accordingly, these facilities are a high priority for future emissions testing.

• Diesel Vehicles: The literature on dioxin emissions from diesel vehicles is quite limited and somewhat contradictory. The tunnel study by Oehme et al. (1991) suggests a relatively high level of emissions. This study is based on Norwegian fuels which may differ in composition from U.S. fuels and, although aggregate samples were collected representing hundreds of vehicles, the indirect method of analysis introduces uncertainty. Much lower emissions were measured by Marklund et al. (1990) on the basis of direct tailpipe tests involving diesel fuel in a heavy-duty Swedish vehicle (Marklund et al., 1990). This study reported no emissions at a detection limit of 100 pg/l or approximately 0.05 ng/km. This is a factor of 100 lower than the emission rate reported by Oehme et al. (1991). Because this study's results are based on only one vehicle using Swedish fuel, this emission factor is also quite uncertain. These two studies yield a very wide range of emission estimates and clearly suggests that further testing is needed.

• Coal-Fired Utilities: The importance of these facilities remains unknown. Only one U.S. facility has been tested and no detectable levels of dioxin were found. If dioxin were present at the detection limit, an emission factor can be calculated which suggests that, due to their number, these plants could collectively represent a moderately sized source. The potential importance of this source is enhanced by several factors. In addition to being numerous, they are large in size and their high stacks indicate that they could impact very large areas. Testing is currently underway to better characterize these emissions.

• Pulp and Paper Mills: These facilities can have dioxin releases to water, land and paper products. The paper industry has recently made process changes which they estimate have reduced dioxin emissions by 90% from 1988 to 1992 (NCASI, 1993). Extensive surveys encompassing virtually all mills have been conducted, making this industry one of the best characterized in terms of dioxin emissions.

The other combustors evaluated in this report appear to be relatively minor sources on a national scale (although their local impacts could be important to evaluate). These include sewage sludge incinerators, hazardous waste incinerators, Kraft liquor boilers, drum and barrel reclaimers, tire combustors, carbon reactivation furnaces and scrap electric wire recovery facilities. The releases associated with chemical manufacturing could not be quantified due to the lack of test data. Potentially such releases could occur via the product itself or as emissions to the air, land or water. Such releases have lead to the termination of production of PCBs and some phenoxy herbicides. Recently, some claims have been made that significant dioxin emissions may occur during the production of vinyl chloride monomer and associated products. These claims have been strongly disputed by the industry. Insufficient emission data are currently available to make an independent evaluation.

Several investigators have attempted to conduct "mass balance" checks on the estimates of national dioxin releases to the environment. Basically, this procedure involves comparing estimates of the emissions to estimates of aerial deposition. Such studies in Sweden (Rappe, 1991) and Great Britain (Harrad and Jones, 1992) have suggested that the estimated deposition exceeds the estimated emissions by about 10 fold. These studies are acknowledged to be quite speculative due to the strong potential for inaccuracies in emission and deposition estimates. In addition, the apparent discrepancies could be explained by long range transport from outside the country, resuspension and deposition of reservoir sources, atmospheric transformations or unidentified sources. Bearing these limitations in mind, this procedure has been used here to compare the estimated emissions and deposition in the United States.

Deposition measurements have been made at a number of locations in Europe (see Volume II) and two places in the United States (Koester and Hites, 1992). These limited data suggest that a deposition rate of 1 ng TEQ/m²-yr is typical of remote areas and that 2-6 ng TEQ/m²-yr is more typical of populated areas. Applying the values of 1 ng TEQ/m²-yr to Alaska and 2-6 ng TEQ/m²-yr to the continental United States, the total U.S. deposition can be estimated as 20,000 to 50,000 g TEQ/yr. This range can be compared to the range of emissions for the United States, 3,300 to 26,000 g TEQ/yr, as presented in Table II-3. It is not clear whether this type of mass balance can ever be refined to the point where definitive conclusions can be drawn. However, it remains one of the few methods of evaluating the existence of unknown sources.

II.3. OCCURRENCE AND BACKGROUND EXPOSURES

Polychlorinated dibenzo-p-dioxins (CDDs), polychlorinated dibenzofurans (CDFs), and polychlorinated biphenyls (PCBs) have been found throughout the world in practically all media including air, soil, water, sediment, fish and shellfish, and other food products such as meat and dairy products. The highest levels of these compounds are found in soils, sediments, and biota; very low levels are found in water and air. The widespread occurrence observed is not unexpected considering the numerous sources that emit these compounds into the atmosphere, and the overall resistance of these compounds to biotic and abiotic transformation.

II.3.1. United States Food Data

All available data on background levels in United States food are summarized in Table II-4. "Background" concentrations are defined here as those for which no source of dioxin-like compound contamination was identified to have impacted the concentrations reported. The background TEQ estimates are presented first assuming that nondetects equal half the detection limits and second assuming that nondetects equal zero. For food groups such as eggs, a wide range of TEQ estimates are seen indicating a high percent of nondetects among individual congeners. The higher of the two TEQ estimates, that calculated using half the detection limit for nondetects, are generally comparable to the TEQ estimates derived from studies conducted in Germany (Fürst et al. 1991) and Canada (Gilman and Newhook, 1991). The German and Canadian studies did not, however, report how nondetects were treated in deriving their TEQs, but did report many nondetects in some food groups. In summary, the limited number of United States food samples and the high incidence of nondetects make an uncertain basis for estimating national background levels, although they are reasonably consistent with food level estimates reported for Canada and Germany. It is clear that more data are needed to adequately characterize the levels of dioxin-like compounds in the United States food supply. Although a large scale survey could confirm residue levels of CDD/F, some attention also needs to be paid to sampling/analytical methodology. Since many of the detected values are only a few multiples above reported detection limits, significant uncertainty results in reported mean values when there are many nondetects in a food category.

II.3.2. Summary of Media Levels

The estimated levels of CDD/CDFs in environmental media and food are summarized in Table II-5 and shown graphically in Figure II-4. Except for the TEQ levels in European food which are based on data reported for German food by Fürst et al. (1990), all other TEQ levels presented in Figure II-4 are based on the data analyzed in this study. The background TEQ levels of CDD/CDFs in water and air were found to be lower than in any of the other environmental media evaluated and were not included in Figure II-4. For most

Table II-4. Summary of CDD/F levels in United States food (pg/g fresh weight)

ND = Nondetect; DL - Detection Limit

Footnotes:

NDA = No data available.

a Values are the arithmetic mean TEQs and standard deviations.

b Standard deviations could not be calculated because detection limits for most samples were not reported.

c Value was calculated from the raw data used in EPA (1991b) using half the detection limits for nondetects.

d Standard deviation could not be calculated because data were limited for the congener that contributed the most to the total TEQ.

e Soil, sediment, and air values based on data from a variety of European countries (see Tables B-17 to B-30); egg data based on Beck et al. (1989); and other food levels based on data from Germany (Fürst et al., 1990).

f TEQ calculated from Fürst et al. (1990) for fresh water fish by assuming 7% fat content (EPA, 1993).

g TEQ assumed to be the mean of the midpoints of the ranges reported in four European studies (Clayton et al., 1993; König et al., 1993a; Liebl et al., 1993; Wevers et al., 1993).

h TEQ calculated from Fürst et al. (1990) by assuming 4% fat content.

i TEQ calculated for cheese from Fürst et al. (1990) by assuming 8% fat content.

j TEQ for beef calculated from Fürst et al. (1990) by assuming 19% fat content.

k TEQ for veal calculated from Fürst et al. (1990) by assuming 19% fat content.

l TEQ calculated from Fürst et al. (1990) by assuming 15% fat content.new Table II-5

media, the average levels appear to be similar between North America and Europe. However, differences were noted in three areas:

• Sediment: The background levels in Europe were estimated to be higher than North America. It should be noted, however, that only the 2,3,7,8-TCDD/F and OCDD/F congeners were analyzed for background sediment sites in the United States and Europe. The sediment data are quite variable and can be very high in impacted areas (i.e., 2,3,7,8-TCDD levels over 1000 ppt have been measured in industrial areas). Also, it was difficult to interpret whether some of the European data truly represent unimpacted areas. Thus, these differences may be due more to the weakness of the data base and interpretation difficulties, rather than real differences.

• Dairy Products: The data on dairy products suggest that North America levels are higher than European. Dairy products include a wide variety of food items with varying amounts of fat. Thus, the CDD/F levels would vary correspondingly. Differences in the mix of dairy products used for the North America and European estimates could explain these differences.

• Pork: The pork data suggests that North America levels are higher than European levels. The low number of samples collected in both Europe and North America may mean this estimate is not representative.

In general, the differences noted above probably reflect the sparseness or inequalities in the data rather than real differences. The small number of samples available for analysis, particularly for food, should be considered when evaluating data from the United States and elsewhere. The human tissue data (see discussion below) suggest similar body burden levels in the North America, Europe and other industrial countries. Thus, it seems likely the media levels would also be similar. Large scale "market basket" type food surveys would be needed to confirm these levels.

II.3.3. Conclusions for Mechanisms of Impact to Food Chain

CDD/F can enter aquatic systems by either direct effluent discharges or atmospheric deposition. CDD/Fs in the atmosphere can deposit directly onto water bodies or onto watersheds and run off into the water system. The mechanism of impact which dominates in aquatic systems will depend on site specific conditions.

This assessment proposes the hypothesis that the primary mechanism by which dioxin-like compounds enter the terrestrial food chain is via atmospheric deposition. Deposition can occur directly onto plant surfaces or onto soil. Soil deposits can enter the food chain via direct ingestion (i.e. earth worms, fur preening by burrowing animals, incidental ingestion by grazing animals, etc). CDD/F in soil can become available to plants by volatilization and vapor absorption or particle resuspension and adherence to plant surfaces. In addition, CDD/F in soil can adsorb directly to underground portions of plants, but uptake from soil via the roots into above ground portions of plants is thought to be insignificant (McCrady, et al. 1990).

Support for this air-to-food hypothesis is provided by Hites (1991) who concluded that "background environmental levels of PCD/F are caused by PCD/F entering the environment through the atmospheric pathway." His conclusion was based on demonstrations that the congener profiles in lake sediments could be linked to congener profiles of combustion sources. Further argument supporting this hypothesis is offered below:

§ Numerous studies have shown that CDD/Fs are emitted into the air from a wide variety of sources (see Chapter 3 of Volume II).

§ Studies have shown that CDD/Fs can be measured in wet and dry deposition in most locations including remote areas (Koester and Hites, 1993; Rappe, 1991).

§ Numerous studies have shown that CDD/Fs are commonly found in soils throughout the world (see Chapter 4 of Volume II). Atmospheric transport and deposition is the only plausible mechanism that could lead to this widespread distribution.

§ Models of the air-to-plant-to-animal food chain have been constructed. Exercises with these models show that measured deposition rates and air concentrations can be used to predict measured food levels (Travis and Hattemer-Frey, 1991; also see Chapter 7 of Volume III).

Alternative mechanisms to the air-to-food hypothesis seem less likely:

- Uptake from water into food crops and livestock is minimal due to the hydrophobic nature of these compounds. Travis and Hattemer-Frey (1987, 1991) estimate water intake accounts for less than 0.01% of the total daily intake of 2,3,7,8-TCDD in cattle. Experiments by McCrady, et al. (1990) show very little uptake in plants from aqueous solutions.

- Relatively little uptake is expected in food from soil residues that originate from sources other than atmospheric dispersion, i.e. pesticides, sewage sludge, and waste disposal operations. Pesticides are discussed below. Sewage sludge application onto agricultural fields is not a widespread practice and the amount of CDD/F in this material is quite low compared to the amount emitted to the atmosphere (See Chapter 3 of Volume II). Waste disposal operations can be the dominant source of CDD/F in soils at isolated locations such as Times Beach, but are not sufficiently widespread to explain the ubiquitous nature of these compounds.

- The contribution of CDD/Fs to the environment via pesticides has been reduced in recent years but remains somewhat uncertain. In the past, CDD/Fs have been associated with certain phenoxy herbicides. Many of these compounds are no longer produced and EPA has sponsored data call-ins requiring certain pesticide manufacturers to test their products for dioxin content. The responses, so far, indicate that levels in these products are below or near the limit of quantitation (see Chapter 3 of Volume II).

- Uptake into food from paper products also appears to be minimal. In the early 1980s, testing showed that CDD/Fs could migrate from paper containers into food. Current levels in paper products are now much lower, and food testing in products such as milk and beef have shown detectable levels prior to packaging, suggesting packaging is not the major source (see Chapter 4 of Volume II).

A related issue is whether the CDD/F in food results more from current or past emissions. Sediment core sampling indicates that CDD/F levels in the environment began increasing around the beginning of the twentieth century and have been declining since about 1980 (Smith et al, 1992). Thus, CDD/Fs have been accumulating for many years and may have created a reservoir that continues to impact the food chain. As discussed in Chapter 3 of Volume II, researchers in several countries have attempted to compare known emissions with deposition rates. These studies may suggest that annual deposits exceed annual emissions. One explanation may be that the reservoir sources cause deposition through volatilization/atmospheric scavenging or particle resuspension. These mass balance studies are highly uncertain and it remains unknown how much of the food chain impact is due to current versus past emissions.

II.4. TEMPORAL TRENDS

Small amounts of dioxin-like compounds may be formed during natural fires suggesting that these compounds may have always been present in the environment. However, it is generally believed that much more of these compounds have been produced and released into the environment in association with man's industrial and combustion practices, and as a result, environmental levels are likely to be higher in modern times than they were in prior times. However, the trend may now be reversing (i.e., releases and environmental levels may be gradually decreasing) due to changes in industrial practices (Rappe, 1992). As discussed earlier, the potential for environmental releases of dioxin-like compounds have been reduced due to the switch to unleaded automobile fuels (and associated use of catalytic converters and reduction in halogenated scavenger fuel additives), process changes at pulp and paper mills, improved emission controls for incinerators, and reductions in the manufacture and use of chlorinated phenolic intermediates and products.

Studies that may be used to assess temporal trends in human exposure to dioxins and furans are extremely limited. Analysis of sediment core layers has shown increases in CDD/CDF concentrations beginning in the 1920's and continuing until the late 1970's (Smith et al, 1992). Another useful study for evaluating changes in human exposure over time is EPA's National Human Adipose Tissue Survey or NHATS. The purpose of NHATS is to monitor the human body burden of selected chemicals in the general U.S. population (EPA, 1991a). The results of this study indicate that exposure to certain dioxins and furan congeners may have decreased over this 5-year time period. However, further studies are needed to verify that these changes are not a result of protocol changes, but actual reductions in exposures. A recent study by Patterson et al. (1994) found decreases in PCB body burdens from 1982 to 1988/89 based on human tissue and blood testing.

II.5. BACKGROUND EXPOSURE LEVELS

Table II-6 illustrates the derivation of a background exposure level to CDD/F for the United States on the basis of diet. This estimate was derived using the upper-range background concentrations (i.e., those calculated using one-half the detection limit for the non-detects) and central estimates of ingestion rates. This approach yields a total background exposure estimate for CDD/Fs of 119 pg TEQ/d. The exposures by pathway are diagrammed in Figure II-5.

The background exposure estimates are intended to be representative of the general population. They do not account for individuals with higher consumption rates of a specific food group (e.g., subsistence fishermen, nursing infants, and subsistence farmers--these are discussed Section II-6). The fish concentration used to estimate background exposures, represents the average value found in fish from fresh and estuarine waters (see Section 4.5 of Volume II). Correspondingly, the ingestion rate used here reflects the per capita average ingestion rate of fresh/estuarine fish (EPA, 1989). Many individuals are likely to have higher ingestion rates of marine fish. However, the limited data on marine species indicates that the dioxin levels may be one to two orders of magnitude lower than fresh/estuarine water fish (also see Section 4.5 of V. II).