Tobacco Documents Online

~;sK, 14 CONNECTICUT'S DIOXIN AMBIENT AIR QUALITY STANDARD Hari V. Rao, Ph.D. and David R. Brown, Sc.D. Connecticut Department of HeaLth Services Division of Environmental Epidemiology and Occupational Health 150 Washington Street Hartford, CT t ABSTRACT Connecticut is the first state in the crm nLry to have adopted au ambi- ent air quality standard for dioxins at 1 pg/m3, 2,3,7,8-TCDD equivalents, as annual average. This paper describes the scientific basis and the meth- odology used by the State Department of Health Services (the risk assessment agency) in assisting the Department of Environmental Protection (the risk management agency) to establish a health-based riioxin standard. This stan- dard protects the public health from the aggregate effect of all sources of dioxin emissions in the vapor and particuLate phases. The risk assessment methodology included: ;,a limit on total daily dioxin exposure from all media and sources based on reproductive effects, a multi-media non-source specific exposure assessment, an apportionment by media of the health-based limit (including background dosing rate), an evaluation of inhalation bioavaila- bility and cancer risk based on a calculation oC a range of upperbound cancer risk estimates using different potency, bioavailability, and particle phase assumptions. : Key Words. Reproductive Effects,'sAfulti-media Exposure, Dose Apportton- ment, Bioavailability, Carcinogenic Potency. INTRODUCTION This report describes the scientific basis for Connecticut's primary Ambient Air Quality Standard (AAQS) for dioxins established by the State Department of Environmental Protection (DEP) at I. picogram per cubic meter (1 pg/m3) 2,3,7,8-TCDD equivalents as annual average (a picogram is one trillionth of a,gram). Dioxins and 2,3,7,8-'fCUD are used interchangeably in this report. The dioxin AAQS is based on the State Department of Health Services' (DHS) analysis of a Dioxin Ifealth Risk Assessment prepared by Hart Associates. (1) The Health Risk Assessment was prepared in response to a legislative mandate, and designed to be consistent with DEP's Air Toxics Program requirements. (2) Thus, the standard setting process was a bilat- eral effort, which utilized the expertise ot the DIIS in risk assessment, and of the DEP in risk management. A Scientific i'anei served as an advisory body.

The details of the Dioxin Risk AnaLysiS, Llie key assumptions used, and their rationale are presented below: r DIOXIN DAILY DOSE AND BODY BURDEN The risk analysis developed a rationale for consideration of daily dose and body burden for dioxins which are independent of specific sources. This was considered appropriate for the following reasons: First, dioxin exposure is multimedia in nature. Dioxins have been de- tected in air, soil, sediments, suspended sediments, water, fish, meat, and milk as well as in human adipose tissue, breast milk, and blood.~3+4) The total body burden represents the sum of potentiatl.y significant individuat contributions from the various media, sources and routes of human exposure. By comparing the relative contributions from each medium with the health- based limit, significant exposures can be identified and necessary measures taken to reduce such exposures. Second, the evidence suggests that there exists a background body burden of dioxins in the general population of industriatized nations.0) The body burden measurements provide an indlcnLion of past exposure, and can be used to calculate the total daily dioxin intake from the background expo- sure. The health impact from this daily background exposure can be as- sessed, and thus incorporated into the standard setting process. Ilowever, the ways in which all of the various media, sources and routes contribute to this background exposure remain unknown. Although there are uncertainties in assessing exposures, an advantage of this approach is that it places the various media of exposure, including the overall background mutti-media exposure, in perspective. REPRODUCTIVE EFFECTS AND DOSE LIMIT Reproductive, developmental, and carcinogenic effects were determined to be health impacts of concern and were evaluated. Animal studies on 2,3,7,8,-TCDD toxicity have clearly demonstrated that it is a developmental and reproductive toxin in a variety of species at relatively low doses. A review of the pertinent developmental and reproductive studies can be found in EPA's Health Assessment Document for PCDDs.(5) The reported adverse %outcomes include reduced fertility,,~litter size and survival, offspring body weight changes, as well as cleft palate and kidney abnormalities. Among the available studies, Murray et al's (6) 1979 study on Sprague Dawley rats, and Allen et al's (7) 1979 study on rhesus monkc,ys were considered appro- priate for quantitative assessment. Murray et al's study examined the effect, of dietary exposure to 2,3,7,8-TCDD on reproduction in Sprague Dnwiry rats over three generations. The rats were given 0, 0.001, 0.01, and 0.1 ul;/kf;/day. No significant toxic effects were observed in the F0 generation during 90 days treatment prior to mating. The study showed that the lowest dose, 0.001 ug/kg/day had no effect on fertility, litter size or fetal survivnL. The authors concluded that the doses 0.01 and 0.1 ug/kg/day procfucvci significant effects on the reproductive capacity through three gener:rLions, F0, FL, and F2. 'tlie study indicated that the 0.001 ug/kg/day couLd he considered as a nn effect level. A reanalysis of the Murray et al. d.;t:r rrni.nl; a different statistical approach concluded the lowest dose was an rCfecL level and that a no effect level could not be determined. (8) Since there was a question relative to the no effect level in the rat study, DIIS considered the data on 2,3,7,8- TCDD's effects on reproduction in rhesus munkrry .. The doses administer.ed in

the diet were 0, 1.8 (50 ppt) and 18 ng/kg/day (500 ppt) up to 9 months. Following 7 months of treatment, the femnlr•rc rorr m:rted with untrentad males. At the higher dose (18 ng/kg/day), Llrr-r(' wns a decrease in serum es- tradiol and progesterone. The menstrual cyclc w:r, however not affected. Only three animals conceived, after whicl, two alcrrCed and one had a nr>rmal birth. At the 1.8 ng/kg/day dose serum estr:rcliol and progesterone levels were normal. Eight treated females were mated with untreated males; there were six pregnancies, four abortions and two normal births. DHS judged that the results of the sAnsitive rhesus monkey studies could potentially support a more conservatire Fcfect level than that reported in the rat studies. Ac- cordingly, the 1.8 ng/kg/day lowest effects level was used to calculate a health-based limit on the total dioxin daily intake from multimedia exposure. Besides reproductive effects, tumorogenic effects have been observed in rodent experiments following chronic exposure to low levels of 2,3,7,8-TCDD. For example, Kociba et al's 1978 two year study tested cancer response in rats at doses of 1, 10, and 100 ng/kg/day. The l0 ng/kg/day dose caused a statistically significant increase in liver tumors in experimental animals versus controls. (9) A comparison of the rodent dose-respons.- rl:rL:l from the cancer study by Kociba (9) and reproductive study by Murray ((') showed that the two ex- perimental adverse outcomes observed in separate bioassays, reduced fer- tility and liver tumors, appear to be the rerorit of exposure to equi-toxic doses of 2,3,7,8-TCDD, i.e., 10 ng/kg/day. T'lif, experimental exposure dura- tions were different, 3 months for first reproductive effects, and 24 months for tumor effects (time to fatal tumor data are not available). For a con- gener like 2,3,7,8-TCDD, the cumulative close is more critical than the dose rate. (10) The data from reproductive and cancor bioassays support ttiis view. By factoring the exposure duration ancl c:rlculating the cumulative dose for each outcome, it can be shown thlt 17. t:rr 25 percent of the cumula- tive cancer dose causes fertility effects in tiw r;nme species. The cumula- tive dose analysis suggests that the potential :iflverse reproductive effects from dioxin exposure present a substantial immedinte concern and cancer is a chronic concern at the same levels of exposure. 'Cfre data from the rhesus monkey studies by Allen et al (7) provide furtlor support to the argument that the reproductive response is a very sensitivrl !:esponse. The experi- mental evidence points to a lower Lowest Obsorvorl I;f.fect Level (1.OEL), 1.8 ng/kg/day, compared with a LOEL of 10 ng/kg/ci:ry, identified in the rat studies. Factoring these values, and the exposure duration of six months in %the non-human primate studies, and thNee months in the rat reproductive bto- assays, even a smaller fraction, approximately 40 percent, of the cumulative administered dose (rats) can be estimated to elicit adverse reproductive effects in rhesus monkeys, i.e., the latter slTncies exhibits 2-3 fold great- er sensitivity than the rodent species. Tfre c:rrmul:ttive dose response anal- ysis places the sequence of health concerns - rr•productive, developmental and carcinogenic in perspective. A further concern arose from the experimr•nt.:rl observations of Moore et al which showed that high levels of the unmetabolized dioxin congener, 2,3,7,8-TCDD, were excreted in milk and that o:cch rat pup actually received a higher dose during the first week after birth than was administered ini- tially to the mother.(11) This study revealed tirnt while TCUD crosses the placenta in the rat, exposure of the offspring occurs mainly through nurs- ing. Thus the maternal milk pathway is a significant pathway affecting neo- natal development in rats. Dioxins have been clotected in human breast milk but there is no evidence to link dioxin exposure through nursing to human neonatal developmental toxicity. It was concluded that the animal data on reproductive effects can be used to derive a tnt.nl dose limit to human ex- posure in 2,3,7,8 - TCDD equivalents. Thus, Ltn: reported LOEL 1.8 ng/kg/day

in the rhesus monkey study was reduced by nn ilnrrrtainty Factor of 1000 to estimate the dose limit at 1.8 pg/kg/day. DOSE APPORTIONMENT BY MEDIUM Since cumulative dose is relevant to dioxin effects and since there are multiple media and sources of dioxins, it was necessary to consider an ap- portionment approach. Assumption. Considering the multi-media nature of dioxin exposure, the health-based limit of 1.8 pg/kg/day (based on reproductive effects as most sensitive response) should be apportioned by medium of exposure and an al- lowable level established for each medium. 11ris apportionment considers po- tential background exposures to be significant. Rationale. Although the background dioxi.n levets in the environment contribute to the total human body burden, this information was not factored in the calculation of the health-based total dose limit of 1.8 pg/kg/day. Thus the limit reflects the total theoretical pf~rmissible daily dose of di- oxins from all media of exposure, including background exposure. It repre- sents the maximum daily dose that should not be exceeded to assure that no adverse health effects occur over a lifetime o€ exposure to dioxins. There- fore when assessing only one of the several possible exposure media, it is necessary to apportion the health-based Limit to account for other potentLzL exposures. to di- oxins was to estimate the background contribution to total dioxin exposure. The average daily intake of dioxin can be estimated using a linear, one com- partment model: (3) The first step in apportioning multi-media exposure of humans Background Dose Rate = Body liurdon x Ln 2 / hnLf-life (ng/kg/day) (ng/ki;) (days) Assuming that a human weighs 60 kg, has 20 perc,,nt fat, and has 7 ppt dioxin in fat (ng/kg),(12) then the body burden of dioxin is 84.0 ng. The half life is assumed to be 5.8 years (2120 days)(L3) and the dose rate is esti- mated to be approximately 0.45 pg/kg/day. Travis and Hattemer-Frey esti- mated through half-life modeling and 70 kg assumption that human exposure to 2,3,7,8-TCDD is about 0.4 pg/kg/day. (~) An EPA calculation showed a range of estimates of daily intake of 2,3,7,8 - TCDD between 0.04 to 0.51 pg/kg/day. The daily dioxin intake value used by UfiS is in reasonable agreement with those reported above. This backsround exposure at 0.45 pg/kg/day (direct and indirect) represents 25 percent of the health-based limit of 1.8 pg/kg/day. Estimates of the relative contributions Err,m air, food, water, and soil to the daily human exposure to dioxins were calcuLated from literature data. The available data on dioxin exposures were reviewed, in particular the Federal Ontario dioxin exposure assessment document. (14) This document assumed that dioxins in the Ontario environment are principally from incineration processes. Based on concentrations and contact rate tile relative contributions were estimated to be: air (60%), water (5%), soil (5%), and food (30%). DHS adjusted this apportionment to account for (i) potential beef and milk exposure, and (ii) sensitive sub-groups (infants and children - milk pathway). Thus DIIS estimated tlrr! relative contribution to be: air (40%), water (5%), soil (5%), and food (50%) in tiie Connecticut environment. The relative source contribution of 40 percent from the air med2um was

derived from the worst case exposure assessmPnt oC the Ontario environ- ment.(14) The Ontario data represent thr mon:cnr•omrnts of stack air (there was no ambient air data) and the leve l.s Cc,urnl i n s,tmpies of f ish, pork, poultry products, drinking water, human fat. and the soil in the vicinity of an incinerator. The Canadian assessment used (i) zn estimated maximum an- nual average ambient air concentration of 8.tr t>1;/m3 TCDD equivalents (60% apportioned intake); (ii) for water, a concentr:+tion of 0.002 ng/L TCDD equivalents (5% intake); (iii) for soil, a leveL of 81.1 pg/g TCDD equiva- lents (5% intake); and (iv) for food consisting of fish, poultry, pork and eggs, 29.6 pg/g (30% intake). No meat, milk and fruit analyses were pro- vided in the Ontario analysis, consequently, the Connecticut food apportion- ment was adjusted to 50%, and air to 40%. For the air medium (40% apportionment), the matrix was considered to include both vapor and particulate phases (the ambient air quality standard takes into account both phases). Dioxins and furans released from a variety of combustion sources have been shown to exi:;t. irr vapor and particulate phases (15). The vapor phase, as well as the pnrticulate phase (assumed to be 100 percent in the respirable range) represent an inhalation hazard. Moreover, volatilization from the background :;nd atmospheric transport of these semivolatile organics can potentially nd,l to inhalation exposure. Based on sampling data and modeling, 2,3,7,8-TC:DI) in the urban air has been reported to exist in the particulate phase between 40 and 80 percent (16) whereas the octaisomer is 100 percent particle bound. The vapor phase half-life through photol.ysis has been reported to be under six hours, and for the particulate phase the half-life is several hun- dred hours. At locations close to the spectrum of combustion sources expo- sure to vapor phase dioxins via inhalation r.nn c,ccur, in addition to direct inhalation of the respirable particulate phase. The background levels of dioxins in the vicinities of resource recovery £.nciLities in Connecticut have been measured. The values (48 hr nv(_,rnl;r,) :rr(•: Mean = 0.045 pg/m3 •F 0.77, Maximum = 0.719 pg/m3, Range = 0.004 to 0.719 pg/m3 dioxin equivl- lents, and N= 130. Fish samples (background mnnitoring) showed that the levels ranged from 0.23 to 8.95 pg/g for TCDF, and from a method detection limit of 0.05 to 6.15 pg/g for 2,3,7,8=fCDD. 1'Ire monitored background data for Connecticut, although limited, indicate that both the atmospheric and food chain exposures.are potentially significant human exposure pathways. The settling velocity is an important factor in determining the signif- `cance of exposure pathways. WhereaSe the irrlrnl:rtion exposure pathway con- tributes to a constant absorbed dose (1 pg/m3 x 20 m3/day = 20 pg/day) from the vapor and particulate phases (this inlraLed dose is independent of settling velocity), the dose estimate for the indirect food chain pathway ts dependent on the settling velocity assumption. For example the Iiart analy- sis (1) showed that the food chain contribution increased with increasing settling velocity from 56, 80, to 98 percent for settling velocities of 0.0003, 0.001, and 0.01 m/sec respectively, Cor the same ambient concentra- tion. The question arises as to the approprilte settling velocity to use in dose calculation. Travis et al's 1987 analysis used a settling velocity of 0.0023 m/sec and 100 percent particle-phase distribution to estimate the food pathway's contribution to total daily intake (98 percent). (3) On the other hand, a settling velocity of 0.001 m/sec wns used in the I{art document to estimate the food chain contribution (80 percent). (1) Additionally, if the particle phase distribution were to be £:rctcrred into the calculation (40 to 80 percent for 2,3,7,8-TCDD) then the food chain pathway percent contri- bution would be in the 56 to 72 percent r.urf;e. r:onnecticut's 40 percent re- lative source contribution from air, and 50 percent from food to the daily

dioxin intake are consistent with an averai;e sottLing velocity of 0.001 m/sec, and 60/40 particle-vapor phase rlist.riirrrt ion :insumpttons. It should be emphasized that the exposure assessmenL nnd dosc apportionmeut for the Connecticut assessment are based on the assumption that direct inhalation intake is from vapor and particulate phases, nrrrl the indirect intake is pri- marily from the particulate phase. The apportioned daily dosing rate asso- ciated with air exposure is 0.72 pg/kg/day (40 percent of 1.8 pg/kg/day). The equivalent dioxin concentration in ambient air is 2.2 pg/m3 (0.72 pg/kg/day x 60 kg/20 m3 per day). The calculntions and conversions are based on the do-,e limit of 1.8 pg/kg/day. DEP RISK MANAGEMENT The DEP considered the DHS assessment, the Hart assessment and their factors in the risk management phase. Connecticut DEP reviewed a range of health-based estimates for a dioxin equivalent AAQS - 0.1 to 2.2 pg/m3. The lower bound value (0.1 pg/m3) comes from the initial Hart analysis and the upper bound, from the DIIS anal- ysis. DEP decided to reduce by a factor of 2.2 the highest concentration in the range and derived a level of 1 pg/m3. This dLoxin equivalent concen- tration of 1 pg/m3 was proposed and adopted as the AAQS. The DEP management decision to apply nn adrlitLonai safety factor of 2.2 was based on the desire for an added margLn o[ protection against potential carcinogenic and immunotoxic effects and ou:opc•rrtting considerations. Tit[s safety factor assured that no exceedences of Lite health-based limit would occur through indirect and background exposures. According to DEP, the de- cision considered other management inputs, such :is monitoring and enforce- ment as well as analytical and statistical considerations. The following analysis explains the heal.Lh rationale for the 2.2 factor and shows how the standard of 1 pg/m3 is protcctive of human health: At the maximum ambient dioxin concentration of 1 pg/m3 from all combustion sources, the daily inhaled dose can be estimnte<1 to be 0.33 pg/kg/day (60 kg human body weight and 20 m3 air breathed in n iiny). This calculated dose represents about 18 percent of the health-based limit of 1.8 pg/kg/day. The Hart document provided an estimate of the indirect contribution from 1 pg/m3 air dioxin concentration to be about L.0 pg/kg/day (100 percent par- Iticle phase assumption and 0.001 m/sea settli.rr,n, velocity). This intake Is 55 percent of the limit. Adjusting for 40 to 80 percent particle phase, tite indirect dose can be estimated to be 0.4 to 0.8 pg/kg/day (22 to 44 percent of the limit). Additionally the background c:rn potentially contribute to an estimated 0.45 pg/kg/day (25 percent of the limit). Thus, the direct (inhaled), indirect, and background intakes can contribute up to 65 to 98 percent of the health-based limit. If the 2.2 ::aEfaty factor is not applied to the 2.2 pg/m3 estimate, a potential doubliny; ,F Lhe (lose would occur and the target limit would be exceeded. A liil;itr•r snCety factor (1U) was considered not necessary since DEP proposed t:r> nol;ulnte individual sources through an emission standard that ensures r,tr•h !u,urrr2 will have an insign[f- icant impact on ambient air. The dioxin AAQS I pr/m3 is thus considered to be protective of public health. DIOXIN EXPOSURE AND CANCER RISK The potential cancer risks from chronic rxpoa;ure to dioxins via inges- tion (indirect pathway) and inhalation (clirect patlrwny) were evaluated. The calculation used the standard approach thnt: tlw prurhtct of the exposure dose (pg/kg/day) and the potency value represent> Lho puLential upperbound risk.

Ingestion Risk. The ingestion risks from tlie indirect exposure pathway were estimated for three deposition scenarinr;. 'llie calculation assumed that 100, 80, or 40 percent of the dioxin In Lhr amhiont air ia In the par- ticle phase and that the settling velocity Is U.UUL m/sec. The estimated doses, 1.0, 0.8, and 0.4 pg/kg/day were multiplied by the oral potency values to calculate the potential upperbound ri!clcs, as shown in Tzble 1. Inhalation Risks. Inhalation risk assessment for the direct exposure pathway focussed on the administered dose via inhalation, 2,3,7,8,-TCDD's potency via inhalation, and the effect of the matrix factor on dioxin's potency. Administered Dose. The concentration in the ambient air (pg/m3) and the contact rate (20 m3/day) for a human body weight assumption (60 kg), and 100 percent pulmonary absorption determined the administered dose (pg/kg/day) via inhalation. A second calcuLation considered the particle/ vapor nature of the airborne dioxins as wetl ns the matrix effect and used aa absorption factor of 50 percent to calcul,3te tfie administered dose. Inhalation Potency and Matrix Effect. Since 2,3,7,8-TCDD's potencies for the vapor and particulate phases via inhalation are not known, the oral potency values are used to predict the inhnlrtLinn risk. This is considered a conservative approach for the following reasvns: Dioxin's potency estimates are based on the administered dose. In the lifetime feeding study the rats were given a liet mixed with dioxin dis- solved in acetone matrix. (9) The bioavailabil.ity of dioxin in this matrix is reported to be 85 percent, compared with the values, 25 to 50 percent for the soil matrix, (17, 18) and 1 to 4 percent for the flyash matrix. (19) These bioavailability values are for oral/gastrointestinal absorption. The absorption values for the soil and flyash matrix indicate that considerably higher doses, 2 to 20 times higher than the do:;e in acetone matrix are re- quired to produce the same dioxin concentr.atiuri in the liver and elicit a tumor response. Such a shift in the dose/rrrpo~nno curve relative to the matrix effect would predict a lower potency ostimate for ingested dioxin. Thus the use of the oral potency values (ac.i,lnno mntrix) to predict the inhalation risk, particularly for inhaling p:rt.iclc bound dioxins, is con- sidered conservative. Table 1 presents the esLimoited inhalation risks for 100 and 50 percent absorption assumptions. i Table 1. Dioxin'Exposure and Upperbound Cancer Risk Estimates A Dose * (pg/kg/day) Direct 0.33 0.17 Indirect 1.00 0.80 0.40 10- Risk Specific Dose fg kg day) ** Assumption EPA 6 C'C JG FDA 60 EPA *** 100 100% Abs 5.5x10-5 9xlU-6 5.5x10-6 3.3x10-6 50% Abs 2.8x10-5 4.5x1U-(' 2.8x10-6 1.7x10-6 100% part 1.7x10-4 2.8x10-5 1.6x10-5 1.0x10-5 80% part 1.3x10-4 2.2xiU-5 L.3x10-5 8x10-6 40% part 7x10-5 1.1x1U-5 7x10-6 4x1U-6 * The dose was estimated for a dioxin concentration of 1 pg/m3, a breathing rate of 20 m3/day and a human body weight of 60 kg. ** Femtogram (one quadrillionth of a gram). *** Proposed potency estimate. ,

Depending on the scientific assumptions used and the estimate of tile carcinogenic potency of dioxin, the range o[ ii{perbounds on total risk from indirect and direct exposure to 1 pg/m3 ranges Crom 2 x 10-4-Z 5.5 x 10-5 + 1.7 x 10-4 ) to 6 x 10-6 (1.7 x 10-6 + 4 x 10-6). In the scientific judgement of DHS, the risk was estimated to be at the lower end of the range or 6 x 10-6. A potential risk in Ltie 10-6 range is con- sidered acceptable for an aggregate air standard for all sources of dioxin emissions. However, for individual Resource Recovery Facilities, the poten- tial cancer risk from ambient impact is even lower and is in the 1-0-7 ran e. The background estimate (0.45 pg/kg/day) represents a risk in the 10-~ - 10-6 range, depending upon the assumptions used. DISCUSSION AND CONCLUSION Connecticut is the first state in the country to adopt a dioxin AAQS that protects the public health from the comhined effects of all sources of dioxin emissions. The AAQS is an aggregate stnntllyd and is different from the "standards" of other states which are in fact only maximum allowable impacts for individual RRFs. According to Connecticut DEP, "no resource recovery facility in the state is predicted to hnve an ambient impact of more than 0.037 pg/m3 dioxin equivalents." (2) Tliis maximum predicted impact for each RRF represents about 4 percent o[ tile AAQS, and is consis- tent with the limits imposed by other states (Mnsslcliusetts 0.15, Pennsyl- vania 0.3, Rhode Island 0.02 to 0.2, and New tl:impsliire 0.09 to 0.27 pg/m3). The predicted maximum impact for each RRF in the state represents a potential upperbound carcinogenic risk in the LO'6 to 10-7 range. No adverse reproductive and immunological effects ar.e expected to occur at this impact level. The calculated dose for 0.037 pp/m3 i.mpact level (direct and indirect) is 0.049 pg/kg/day and it repre.<•nts about 3 percent of the health-based limit, compared with the backgronnd's 25 percent. CLearly, there is a need to identify the sources contribuLing to the considerable background intake, and minimize such exposure. The non-source specific approach used in Llica risk assessment is a departure from incinerator-specific risk asses:;monts. Appropriately, this approach takes into account the body burden and d,tity dose from all media and sources as well as the background exposures when estimating a total daily dose and comparing with the target dose limit. The target dose limit assumption facilitated apportioning the dose by media.' While developing a .rationale for uhe 40 percent air appor,tlonment, it became apparent that the risk assessment is sensitive to the settling velocity assumption used. Inhalation (44 percent) and meat/milk ingestion (56 percent) are significant exposure pathways for a minimum particle settLiiiG velocity.(0.0003 m/sec) assumption. (1) The relative significance of CEie Lnhalation exposure decreases as the settling velocity increases. At ttte settling velocity 0.U1 m/sec, the indirect food chain pathway dominates (98 percent) although tile concentration of dioxins in the air is tile svne. Therefore, risk assess- ments o.f this type should estimate exposures bascad upon several settling velocities as a means of estimating a range of potential exposures. This is particularly necessary since the location of tile pLant affects the settling velocity. This type of analysis will help tile risk manager to assess inha- lation exposure separate from indirect multiple patflway exposure,,lf desired. A further improvement in dioxin risk assessment would come from the knowledge of vapor/particulate phase distribution of dioxins in ambient air. This information would improve the analy5is of the indirect exposure pathways, since the deposition of dioxins from tfie ambient air fias been shown to be particle phase dominated.(16) i

REFERENCES 1. F. C. Hart Associates, ."Multiple PaLliway Iluman Rxposure and Ilealth Risk Assessment of Polychlorinated Dibenzo-p-Dioxins and Polychlorinated Dibenzofurans from Municipal Solid Wn,t~ incinerators," Prepared for State of Connecticut, Department of llcalLh Services, February, 1987. 2. Connecticut Department of Environmental Protection (DEP), "Basis for Standards and Procedures and Response to Comments on Proposed Resource Recovery Regulations (Air Poliution Provisions)," (1988). 3. C. C. Travis and H. A. Hattemer-Frey, "Human Exposure to 2,3,7,8-TCDD," Chemosphere 16, 2331-2342 (1987). 4. C. C. Travis and H. A. Hattemer-Frey, "A Perspective on Dioxin Emissions from Municipal Solid Waste Incinerators," Risk Analysis 9, 91-97 (1989). 5. Environmental Protection Agency, "Healtli Assessment Document for Polychiorinated Dibenzo-p-dioxins," Office of Health and Environmental Assessment, May 1984. 6. F. J. Murray, F. A. Smith, K. D. Nitschke, C. G. Humiston, R. J. Kociba, and B. A. Schwetz, "Three-Generation Reproduction Study of Rats given 2,3,7,8-Tetrachlorodibenzo-p-,[ioxin in the Diet," Toxicology and Applied Pharmacology 50, 24t-252 (1979). 7. J. R. Allen, D. A. Barsotti, L. K. LambrecfiL, and J. P. Van Miller, "Reproductive Effects of Halogenated Arumntic Hydrocarbons on Nonhuman Primates," Annals of New York Academy of Sciences 320, 419-425 (1979). 8. I. C. T. Nisbet and M. B. Paxton, "Statist.icril Aspects of three generation studies of the reproductive Loxicity of 2,3,7,8-TCDD and 2,4,5-T," The American Statistician 36, 290-298 (1982). 9. R. J. Kociba, D. G. Keyes, J. E. Beyer, R. M. Carreon, E. E. Wade, D. A. Dittenber, R. P. Kalnins, L. E. Frauson, C. N. Park, S. D. Barnard, R. A. Hummel, and C. G. Humiston, "Results of a Two-Year Chronic Toxicity and Oncogenicity Study of 2,3,7,8-TCDD," Toxicology ; and Applied Pharmacology 46, ;79-303 (i978). 10. T. H. Umbreit, E. J. Hesse, and M. S. Ga1lo, "Reproductive Toxicity in Female Mice of Dioxin-Contaminated Soils from a 2,4,5 - Trichlorophenoxyacetic Acid Manufacturing Site,' Archives of Environmental Contamination and Toxicology 16, 461-466 (1987). 11. J. A. Moore, M. W. Harris, and P. W. Albro, "Tissue Distribution of 14C Tetrachlorodibenzo-p-dioxin in Prern,int and Neonatal Rats," Toxicology and Applied Pharmacology 37, 146-147 (1976). 12. D. G. Patterson, J. S. Holler, S. J. Smith, J. A. Liddle, E. J. Sampson and L. L. Needham, "Human Adipose Dal•i for 2,3,7,8-TCDD in Certain U.S. Samples," Chemosphere 15, 2055-2060 (1986). 13. H. Poiger and C. Schlatter, "Pharmacokinet.ic:; of 2,3,7,8 - TCDD in Man," Chemosphere 15, 1489-1494 (1986).

14. Ontario Ministry of The Environment, "ScicntiEic Criteria Document for Standard Development No. 4-84 - Poly(•hit)rit):ited Dibenzo-p-Dioxins and Polychlorinated Dibenzofurans," i9iiiistry of The Environment, Toronto, March, 1986. 15. T. F. Bidleman, "Atmospheric Processes," Environ. Sci. Technol 22, 361-367 (1988). 16. T. F. Bidleman, "Gas-Particle Distribution and Atmospheric Deposition of Semi Volatile Organic r,umpounds," Presented at the EPA/ORNL Workshop on Risk Assessment for Municipal Waste Combustion, June 8-9, 1989. 17. T. H. Umbreit, E. J. Hesse, and M. S. Ga1lo, "Bioavailability of Dioxin in Soil from a 2,4,5-T Manufacturing Site," Science 232, 497-499 (1986a). 18. T. H. Umbreit, E. J. Hesse, and M. S. Ga11o, "Comparative Toxicity of TCDD Contaminated Soils from Times Beach, Missouri, and Newark, New Jersey," Chemosphere 15, 2121-2124 (1986b). 19. M. Van den Berg, K. Olie, and 0. liutzinger., "Uptake and Selective Retention in Rats of Orally Administered Chlorinated Dioxins and Dibenzofurans from Fly Ash and Fly Aski Extract," Chemosphere 12, 537-544 (1983). a