Document 4QJEypoM8mb00xgVQgwrQd9EV

R&S 113538 bzo-hedical. research document description form Duplicate.in all cards: --^ year as--1961-- File number [Right justify [Numeric only] 77 78 I_L2 Sub-Index Code Author (s), as Last Name FS (No Punctuation) and coden for journal as JAMA preceeded by one blank space r..'. r. ^ -j, y>-' // ______ 61 62 11 J______________ 12 13 Title of Report: end with space-hyphen-hyphen-space- Follow with Index Terms. separated from each other with comma-space. Avoid other punctuation; do not abbreviate. i----------- :------------------------ ! / _ ^, ;--r C i J - -- 't J- 7, 'r vi.C.' -- e'J c /V- c-ir7~>'L Tts f'/.r? 21 22 23 24 Source (Journal, Vol., Number, Pages, Date ) 12 , /' ' .i /, / ' - rfsf' I v' --------- trx ?/, /ff# ____ - L > i. 61 62 "31* 32 Brief Summary 12 10 SUMMARY: 61 62 61 62 63 64 R&S 113539 on the Proposed Rules of the Environmental Protection Agency to Establish Policy and Procedures for Identifying, Assessing, and Regulating Airborne Substances Posing A Risk of Cancer IOAQPS 79-14, 44 Fed. Reg. 58642 (Oct. 10, 1979)] and on the Advanced Notice Proposed Rulemaking for Proposed Generic 'Standards CA-79-13, 44 Fed. Reg. 58662 (Oct.-10, 1979)] Submitted by Environmental Defense Fund Robert J. Rauch Joseph Kagoner Natural Resources Defense Council David Doniger February 21, 1980 TyL--- 0 -1- INTRODUCTION The Natural Resources Defense Council and the Environmental Defense Fund submit these comments on the Environmental Protection Agency's proposed policy for regu lating airborne carcinogens under Section 112 of the .Clean Air Act. We strongly support the adoption and implementa tion of a policy that will obtain effective control of cancercausing air pollutants. One American in four is expected to contract cancer. and one in five to die of it. The vast majority, of cancer cases are preventable. Chemical and radioactive substances released into the ambient air from stationary sources axe the sole or contributing causes of -a subs-han-hiAl fraction of .cases*-- Controlling these carcinogenic air' pollutants is a necessary step in the fight against this disease. Although Congress gave EPA an urgent mandate to con trol these chemicals as "hazardous air pollutants'* under Section 112 of the Clean Air Act, in ten years the Agency has ' set only four hazardous air pollutant standards. Yet there are as many as several hundred chemicals emitted from stationary sources that are known or strongly suspected to cause cancer. A quantum leap in ERA action under Section 112 is necessary if the Agency is to protect the public health, as the Clean Air Act requires. R&S 113540 As we detail in these comments, NRDC and EOF strongly support certain conclusions and determinations in the proposed policy. Specifically, we support:. - EPA's conclusion that the public is expo many--perhaps several hundred--air pollu---------released from stationary sources which may rea sonably be anticipated to cause cancer in humans. R&S 113541 - EPA's conclusion that even low levels of exposure to these pollutants increase the number of persons who may reasonably be anticipated to contract cancer and that no safe level of exposure to a carcinogen can be identified. - EPA's conclusion that airborne carcinogens pose a major public health problem requiring strongly precautionary control of emissions to prevent deaths and serious illnesses. - EPA's conclusion that positive results in human epidemiological studies and animal toxicological studies each are sufficient to establish that a substance is a human carcinogen. - EPA' s commitment to list and regulate as hazardous air pollutants many more carcinogens than the. Agency has to date. '* - EPA's determination to develop a set of quick and inexpensive "generic" control measures for certain industrial categories likely to release multiple carcinogens, to be imposed swiftly as a first step after listing, whil* mare protective_ standards are developed. EPA's commitment to examine whether there are stitute products or processes that- allow a complete elimination of carcinogenic emissions. - EPA's recognition that because new plants have greater pollution control opportunities than existing one?, and because expansion of reliance on carcinogens should not be encouraged, new plants should be subject to greater control re quirements than existing ones. *.* * * < In critical areas, however, the proposed policy mast he radically improved if it is to meet the mandate of the* law and the need to protect public health. Our comments are directed to six key areas: I. II. The Sole of Airborne Pollutants in the Induction of Cancer ---* *" * The Meed for a legally Binding, "Action Forcing* -Process for Screening, Listing, and Controlling a Meaningful Number of Hazardous Air Pollutants TIT- ' The Need to Resolve and Foreclose Generic * Scientific Issues ** IV. Determining the Appropriate Control Level for Airborne Carcinogens j ** V.The Need to Eliminate Arbitrary Dependence on Unreliable Quantitative Risk Assessments a VI. The Need for Improved Generic Controls as , an Interim Measure , At this time NRDC and EDF are submitting a summary of our position on these issues. Before the close of the period for supplementary comments, we shall submit thorough analysis and documentation in support of the positions conveyed in this summary. R&S 113542 r &S 113543 1-1. I* THE ROLE OF AIRBORNE POLLUTANTS IN THE INDUCTION OF CANCER The World Health Organization as well as respected members of the scientific community have estimated that 0 - 90 percent of human cancers are associated with environmental factors. Support for this conclusion derives in part from the great variation in the incidence of most cancers f both from place-to-place and time-to-time ? from the experience of migrants, among whom the risk of cancer changes to that of the adopted country in the course of one or two generations; and k ( ^ from the. increasingly frequent demonstration of excessive cancer among various industrial groups. Whereas some have suggested that lifestyle factors (diet, alcohol and cigarette smoking) are the major sources of environmental cancer, a recent NCI, NXEHS and NIOSH Report--'^ concluded that those estimates were based on the erroneous assumption that each cancer should be assigned to a single specific cause, were limited to the small number of known examples and were largely unsupported by documentation or critical review of data. In contrast,this NCI, NIEHS and NIOSH Report, the conclusions of which were_conaldered-bV-Anwarj..an Trrdnstci*i fjfp*1 Council consultants to fae__reasonable. estimated that in coming decades up to 20 percent and perhaps as much as 40 percent of cancer in the U.5. might be associated with occupational factors.^/ While the possi bility of overestimation has been raised regarding this report, it should be noted that the report only dealt with six substances instead of at least 18 substances known to be confirmed human occupational carcinogens. In arriving at these estimates, the NCI, NIEHS and NIOSH 1/ 45 Fed. Reg. 5031, 5033 (January 22, 1980). 2/ Ibid. R&S 113544 Report1-'/noted: 1-2 The initiation and development of cancer is a multi- phased, multi-causal process in which both external and internal factors act# probably at each of several stages# before frank# clinical cancer appears. It is likely that, many# if not most# cancers are influenced by two or more . different external factors acting simultaneously or sequen tially. Thus# alcohol by itself appears to be a minor cause of cancer -- but alcohol combined with cigarette smoking leads to risks 15 times higher than those experienced by nan-smoking non-drinkers. If a drinker smoker develops cancer of the oral cavity# to which 'cause* should it . be attributed? Drinking or smoking? If we could correctly identify the proportions of cancer incidence 'attributable to' each of the classes of environmental factors considered by Wynder and Gori # the sum of these percentages would be considerably higher than 100. One ' of the best-studied examples of interaction between exogenous agents is that between asbestos and cigarette . smoke in inducing lung cancer. Most lung cancers 'attributable to' asbestos are probably simultaneously 'attributable to'smoking. If current theories of a multi-causal process are correct# it seems likely that a large fraction of cancers which at first appear to be 'attributable to* smoking should also be 'attributable to' asbestos# radiation# and/or other occupational factors. ' Of the six substances considered in the NCI # NIEHS and NIOSH Report# several including asbestos# arsenic and vinyl chloride have been emitted into or are present in the ambient air in such a way that significant human exposure results:. Numerous | studies have reported an excess of cancer among individuals- ggsidin /. in close proximity to industrial sources emitting each of these same three substances. 2/ With regard to asbestos # Wagner' s-- study in South Africa # New- house ' s--^studies in the United Kingdom and Bohlig * s--^study in Hamburg t 1/ NCI# NIEHS and NIOSH Institutes. ' Report. 1978, p. 3. OSHA Hearing Exhibit 224b. 2/ Wagner# J. C., Steggs# C. A. and Marchand# P. Diffuse pleural mesothelioma and asbestos exposure in the North-Western Cape Province. Br. J. Ind. Med. 17: 260-271# 1960. 3/ Newhouse,-M. L., and Thomps n# H.. Mesothelioma of pleura and perit neum following exp sure to asbestos in the London * area. Br. J. Ind, -Med. 22: 261-269# 1966. 4/ Bohlig# H. and Hain. E. Cancer in relation to (cont'd on p. 1-3). 1-3 reported that the risk of mesothelioma has spread beyond the factory, mine or asbestos mill gate, in 1964 Newhouse and Thompson--Reported numerous cases of mesothelioma among individuals whose only identified asb stos exposure was associated with living one-half mile from an .asbestos factory. This form of cancer only rarely is found in the absence of exposure to asbestos. With -regard to arsenic, three studies have shown hn excess of lung cancer among inhabitants of communities with smelters emitting arsenic. Blot and Fraumeni--^demonstrated that the 36 counties in the U'^L UjS. with copper, lead and zinc smelting and refining industries had a significantly higher lung cancer mortality among males (p<0.001) and females (p<0.05) than did counties in the rest of the U.S. Those 35 unties with industries processing nonferrous ores other than copper, ead and zinc showed no such excess of lung cancer mortality. Accor ding to the authors, the most, likely explanation for the elevated lung cancer mortality was neighborhood air pollution from industrial sources of inorganic arsenic. In like manner, Newman, et al. ,3/ demonstrated that the rate of lung' cancer mortality** among females resident in Anaconda, Montana, the site of a copper smelter known to be the point source for arsenic air pollution was significantly ele- . vated (2.9/10,000 vs. 1.4/10,000; p<0.05) . (continued from preceeding page) environmental exposure, type of fibre, dose, occupation and duration of exposure, in Bogovski, P. Gilson, Jr., Timbrell, V. Wagner, J.C. (eds) : Proceedings of the Conference on Biological Effects of Asbestos, Lyon. 1/ Newhouse and Thompson, 0. cit. 2/ Blot, W. J. and Fraumeni, J. F. Arsenical air pollution and " lung cancer. Lancet 2; 142-144, 1975. 3/ Newman, J. A., Archer, V. A., Saccamano, G., Kuschner, M., Auerbach, 0., Grondauhl, R. D., and Wilson, J. C. Histologic types of bronchogenic carcinoma among members of copper mining and smelting communities. Ann. N-Y. Acad. Sci. 271: 260-268, 1976. R&S 113545 1-4 More recently# Matanoski--^reported that the lung cancer mortal^J^ for males residing in the census tract in which an arsenic? l-p=cUu<ri^=g insecticide plant .is located was significantly higher than for males in control census tracts. Exclusion of lung, cancer deaths of persons who had been employed in the'plant did not alter the statistical significance of the excess risk. The female rate of lung cancer mortality in r\ that. same census tract was higher than most control tracts , but the. numbers were too small to attach any statistical-significance to it* 2y 3 / j Regarding vinyl chloride, Infante--'and Iturra --'each have, reported u *r an excess of central nervous system (primarily brain) mortality among residents of communities having vinyl chloride monomer or polymerisa tion plants. These studies # while based upon relatively small numbers # are consistent with results showing a statistically significant excess of brain cancer among individuals occupationally exposed to vinyl chloride.-More recently# Brady, et al.-^reported the results of a study of 26 confirmed cases of liver angiosarcoma and a similar number * of controls with an internal malignant tumor other than primary liver R&S 113546 1/ Matanoski, G. M. Progress in a community arsenic study. Proceedings of Toxic Substances in the Air .Environment Specialty Conference. Pittsburg, 1976# pp. 31-41. 2/ Infante# P. F. Oncogenic and mutagenic risks in' communities with polyvinyl chloride production facilities. Ann. N. Y. Acad. Sci. 271: 49-57, 1976. 3/ Iturra, H. A community vinyl chloride mortality analysis study. Proceedings of Toxic Substances in The Air Environment Speciality Conference. Pittsburgh, 1976, pp. 96-113. </ Waxweiler, R. J. Stringer# W., Wagoner# J. K. # Jones# J. Falk, H. # and Carter# C. Neoplastic risk among workers exposed to vinyl chloride. Ann. N. Y. Acad. Sci. 271: 40-48# 1976. 5/ Brady, J., Liberatore# F., Harper, P., Greenwald# P. , Burnett W., Davis, J.# Bishop, M., Polan, A., and Vianna# N. Angio sarcoma of the liver: an epidemiologic survey. J. N.C.I. 59: 1383-1385, 1977. ------------------- I-S cancer. These controls were matched with the cases on the basis of age of diagnosis* race* sex and general place of residence. The authors noted that of 10 female cases of liver angiosarcoma (no direct occupa tional or therapeutic exposure to arsenic* vinyl chloride or thorium dioxide) five lived for prolonged periods of time within one mile of a VC polymerization plant or a PVC fabrication plant whereas none of their matched controls did so. This observation* according to the ^authors* lent support to the hypothesis that indirect modes of exposure R&S 113547 VC-PVC might be important in the etiology of liver angiosarcoma. While the existence of cancer hazards in the extraplant environ ment' has been demonstrated* it must be recognized that these community- based studies involving asbestos, arsenic and vinyl chloride were succ ssfully completed only because of the relative of -the form of cancer under study* i.e. * mesothelioma; liver angiosarcoma or * brain cancer* all forms of cancer previously having been shown to be excessive in studies of occupationally exposed groups* or because of the emission of a single carcinogenic substance from predominantly * isolated industrial facilities. Therefore* given the high probability of multiple substance emissions associated with modem industrial processes and the limita5- tions of sensitivity inherent in current epidemiological methods* it . is unlikely that the true impact of cancer resulting from specific air pollutants can ever be fully documented. Nevertheless* epidemiolog ical as well as experimental observations would indicate that the potential impact of air pollutants on cancer indication could be even more substantial. Studies of individuals occupationally exposed to asbestos or irradiation have indicated that in the presence of another 549 II, THE HEED FOR A LEGALLY BINDING, "ACTION FORCING" PROCESS FOR SCREENING, LISTING, AND CONTROLLING A MEANINGFUL NUMBER OF HAZARDOUS AIR POLLUTANTS A. EPA*s Ten-Year Failure To Protect She Public From Carcinogenic Air Pollutants * In 1970, Congress enacted special provisions of the Clean Air Act directing EPA to protect the American people from cancer and other especially serious illnesses related to air pollution. Section 112 of the Act requires- swift, precautionary, * and highly protective federal regulation of the air pollutants that pose such risks. The Administrator must maintain and con tinually update a list of ".hazardous air pollutants," substances emitted from stationary sources that he judges "cause, or con tribute to, air pollution which may reasonably be anticipated to result in an increase in mortality or an increase in serious irreversible, or incapacitating reversible, illness." Within 360 days of placing a substance on the Hazardous Pollution | ListEPA must set standards to protect the public health with "an ample margin of safety." [Section 112 (a} {1} , (b) Cl) (A) , CbJCD.CEU In ten years EPA hs> failed to carry out this mandate or even to make a substantial beginning. The Agency has the legal responsibility to protect the. public from cancer and similar illnesses related to air pollution. But EPA has failed to screen, list, and regulate any meaningful fraction of the hazardous air pollutants. -Considering cancer alone, there are as many as II-2 V R&S 113550 several hundred substances that are reported to cause cancer in human or studies and that may be emitted to the air -* from stationary sources, yet to date EPA has never undertaken any organized screening of these substances, has listed only six hazardous air pollutants, and has set standards for only - four. And in every instance EPA has missed the 360-day statu- 1/ tory deadline for setting standards. B.- ' The Elements Of An "Action Forcing" Processs . A Candidate List, A Legally Binding Commitment . To List And Control At Least Twenty Hazardous Air Pollutants Each Year r And A Testing List . The airborne carcinogens policy will not remedy EPA*s ten-year failure to carry out its duties under Section 112 unless EPA simultaneously establishes a legally binding pr ce^P that will assure control of all hazardous air pollutants within a reasonable time, not-just a token handful. The process must include two elements. First, at the same time that it promul gates the policy, EPA must adopt a "Candidate List of Potential Hazardous Air Pollutants." The Candidate List must consist of the substances which an initial screening of the scientific literature, reports, and other information sources reveal have been reported to to cause cancer in human or animal studies and which may be emitted from stationary sources. Initially 1/ Standards for emissions of benzene have not been set even though the chemical was added to the hazardous air pollutant list nearly three years ago. The 360-day period for radio nuclides is still running. IX-3 R&S 113551 the Candidate List probably will consist of several hundred substances, and it must be updated periodically to reflect results of studies of additional substances. Second, also upon promulgation EPA must adopt a legally binding schedule for adding a number of candidate substances to the Hazardous Air Pollutants List each year, and for setting standards for them,until the public is protected from all ! .hazardous air pollutants. * The Candidate List would serve several important func tions. Pirst, it would notify the public, industry. Congress, and EPA itself of the true scope of the potential airborne carcinogens problem. This perspective is essential to EPA for rational internal planning and priority-setting. Undoubtedly it will also help the Agency make the case ^within the Adminis tration and to Congress for additional resources to control hazardous air pollutants. Second, the Candidate List would stimulate industries and others to develop and come forward with data on a candidate substance's toxicity and on whether and to what extent it is emitted and the public exposed. Third, the list would encourage voluntary action by industries to curb emissions of .suspect substances. The binding schedule for listing and regulating a minimum number of hazardous air pollutants each year is necessary to guarantee that actual progress accompanies the airborne R&S 113552 .-II-4 carcinogens policy. Without this commitment, EPA could con* -tinge its current failure to act despite the promises of the ' policy# or could slip back into inaction in the future. NHDC and EDF believe EPA must list as hazardous air pollutants and set standards for a minimum of 20 candidate substances each year. Anything less would fail -to guarantee j,' 0 minimally adequate progress toward the goal of protecting the public from the panoply of hazardous air pollutants, and \ thus would fail to meet the mandate of Section 112 of the Act. f The commitment to a Candidate List and to listing and regulating a minimum number of hazardous air pollutants per year would provide a rational basis for EPA's proposed priorilj^p setting process, which as contemplated now is largely a sham. EPA cannot rationally set priorities unless it has defined the number of potential hazardous air pollutants and established the rate at which it will control them. Without identifying * the potential problems, EPA can have no assurance that the pollutants it selects for regulation are the most serious threats to the public health. And without a binding minimum number to control each year, priority setting can easily degenerate into rationalization for inaction -- an interminable scholastic discussion over which hazards are- most important, never leading to control of more than a handful. With a binding schedule, this cannot happen. r. i II-5 * Adoption of a Candidate List and a minimum number of hazardous air pollutants listed and controlled annually would ease our misgivings concerning EPA*s proposed criteria for listing a hazardous air pollutant. Under Section 111(A). of the proposed rules, EPA would list only pollutants judged, to present a * significant" risk of cancer- This is defined as pollutants found to cause cancer in humans or animals for which there is evidence of "significant" public exposure* (44 Fed. Keg. at 58654) In the proposal notice EPA promises that 'the "significance" requirement is intended just to avoid listing substances for which there is no evidence of public exposure through air pollution, such as "laboratory curious!- ties." (44 Fed. Reg. at 58647] But unless EPA is bound to controlling a minimum number of hazardous air pollutants each *m year, there will he an unacceptable danger that SPA officials, now or in the future, will use the "significance" requirement to frustrate control of any meaningful number of pollutants. 1/ With such a commitment, continued inaction cannot occur. One more element is essential to a rational airborne carcinogens policy and priority-setting process: an "Air Pollutants Testing List." The Testing List would include the pollutants for which EPA, in screening substances for the Candidate List, found insufficient health effects data. Like 17 We have further comments on priority-setting in Section V of these comments. R&S 113553 t* * II-6 the Candidate List, the Testing List must be established when the rules are promulgated and updated periodically. The Testing List would help guide EPA and other agencies in determining priorities for testing programs. Moreover, by publicizing the need for more health effects data on these pollutants, the List will stimulate industry and the scientific community to perform and report such research voluntarily- C- Substances For The Candidate List Before the close of the period for supplemental comments, JJBDC and EPF will submit a list of substances we believe must be placed on the Candidate List. These ar substances: (1) that are reported to increase the risk of cancer either in at least one human epidemiological study or in at least one mammalian bieassay, and * (2) that may be released as air pollutants from stationary sources. Of course our list is only a start; EPA must undertake a more comprehensive screening of the literature and other information sources available to it. 9 D. Substances For Immediate Listing And Control As Hazardous Air Pollutants 11SZC and EDF also will submit a list of pollutants that we believe should be the first hazardous air pollutants listed and controlled after promulgation of airborne carcino gen rules. The evidence that these substances "cause, or R&S 113554 R&S 113555 I II--7 contribute to. air pollution** and that .they* "may reasonably be anticipated" to cause cancer in humans is so compelling and has been available to EPA for so long that the Agency would be arbitrary and capricious to delay listing them any* longer. The only valid reason for not immediately listing one or more of these, in our view, would be that EPA judges another substance to be a higher priority and lists it instead. The precautionary nature of the hazardous air pollutant section of the Clean Air Act deserves special empha sis. Tn 1977, Congress amended definition of a hazardous air pollutant to emphasize that EPA must take preventive action on the basis not only of well-established incontrovertible fact, but on suggestive, probative, though not completely certain, information as well. Congress replaced the earlier defini tion of hazrdous air pollutants -- substances which "cause, or contribute to, an increase in mortality or an increase in serious, irreversible, or incapacitating reversible, illness" --- with the current definition -- substances which "cause, or contribute to, air pollution which may reasonably be antici pated to result in an increase in mortality or an increase in serious irreversible, or incapacitating reversible, illness." [Section 112(a)CD 3 The legislative history of this change makes very clear that Congress intended EPA to take precautionary action # II-8 R&S 113556 to prevent harm before it occurred. Congress perceived that the subtlety of eause-and-effect relationships usually pre cludes precise and certain delineation of pollution-induced injuries for years after they are first' recognized. In the ,, interim/ however, since the public's lives and health are at stake, the Clean Air Act directs the Administrator to take J * preventive action based on suggestive, developing evidence, as *i well-as on well established fact. The House Report specifically- endorsed the opinion of the District of Columbia Circuit Court of Appeals in Ethyl Corp. v. SPA, 541 F.2d 1 (D.C. Cir. 1976) (en banc), cert, denied 426 U.S. 941 (19761, in which this precautionary approach to public health regulation was most * fully expounded. The Souse Committee added that this approac to the evidence is mandatory, not just authorized: (A) substantial element of judgment, including making comparative assessment of risks, projections of future possibilities, establishing margins of' safety and margins of error, extrapolating from limited data, etc., are both necessary and permissible under the Act in order to protect public health... H.R. Rep. No. 294, 95th Cong., 1st Sess. 50-51 (1977) (emphasis ** * addedl (See, generally pages 43-51) ' --- By emphasizing that EPA must protect the public against risks of harm as well as actually demonstrated harm. Congress redoubled the urgency of listing a meaningful number of airborne carcinogens as hazardous air pollutants and setting standards to protect people from them. R&S 113557 4v II-9 E. SPA*s Obligation To Seek Additional Resources EPA may not complain that it lacks the resources to undertake these commitments. The Agency now faces so large * a backlog of hazardous air pollutants largely because it evaded with protecting the public from exposure to airborne carcinogens. The Agency also must take steps to protect ^the public health from hazardous air pollutants causing other killing and disabling diseases , e.a., non-neoplastic respiratory disease, reproductive disorders, mutations, central nervous system diseases, etc. As an interim measure, in selecting the 20 carcinogens to list- and control' as hazardous air pollutants each year, EPA must consider whether certain candidate substances may reasonably be anticipated to cause death or illness from other diseases as well as cancer. R&S 113558 11-1 * After promulgating the airborne carcinogen rules, EPA will have a responsibility to develop similar rules for identifying non-carcinogenic hazardous air pollutants. NRDC ed? are considering proposing criteria for additional Candidate Lists for' such substances. R&S 113559 Figure II-l * An "Action Forcing" System For Screening* Listing* and Regulating Hazardous Air Pollutants Universe o chemical and radioactive substances Screening for potential hazardous air pollutants v Tasting List Air pollutants for which liealth effects data is inadequate Considered for testing and for new source per formance stand ards Further screening and priority setting Candidate List ^ Air Pollutant Potential Hazardous List Air Pollutants At least 20 substances Substances which are added each year wliich reported to cause are air pollutants the cancer in hunans or ` Administrator judges animals and which "cause, or contribute may be emitted from to, air pollution vriiich stationary sources may reasonably be anti cipated to result in an Also must be extended! increase in mortality to Include substances! or an increase in serious which cause other killing irreversible, or incapaci and disabling diseases. tating reversible illness.* Standard setting . .. - i--... -------------- .-------------------------------------------------------------------------------------------- -- Hazardous Air Pollutant Standards At least 20 hazardous air pollutant standards each year, which provide "an ample margin of safety" to protect public health .111-1 R&S 113560 III- TES NEED TO RESOLVE AND FORECLOSE GENERIC SCIENTIFIC ISSUES i Experience regulating carcinogens in the 1970*s . has shewn that there are many scientific issues common to every proceeding that transcend the particulars of each sub stance. They concern either common properties of all carcino gens, or questions that are equally "on the frontier of scientific knowledge" for all. Given the state of scientific understanding of carcinogenesis now and for the foreseeable future, and given the precautionary mandate of the Clean Air Act and other health and environmental protection laws, these generic issues necessarily must be resolved in the same way, on the same central body of evidence, in each proceeding. In past proceedings, EPA and sister federal agencies have developed a consistent set of principles based both on the best scientific knowledge available today and on the pre cautionary public health protection policies embodied in their statutes. For example, that animal studies are a valid means . of identifying human carcinogens, and that no safe dose of a carcinogen can be identified with current scientific knowledge. We expect that EPA will hold to these principles in future proceedings. Yet in the current regulatory process, these issues are reconsidered from scratch in each regulatory proceeding. R&S 113561 * * III-2 despite the fact that generally no significant new information is presented that clarifies the issues on the frontiers of scientific knowledge or calls for a different resolution of them. This is enormously wasteful of the resources of the Agency, the scientific experts repeatedly called to testify or comment, and all parties to the regulatory proceedings. The only parties who gain are industries, who by relitigating these issues before EPA and the courts, delay the day when they must reduce the public's exposure to carcinogens. To adopt and implement effective airborne carcinogen rules, EPA must foreclose the opportunity for this wasteful and dilatory rehashing of the generic scientific' issues. In this generic proceeding EPA must adopt a set of principles, based on the best available scientific know0 ledge and the precautionary mandate of the Clean Air Act, that will not be proper subjects for relitigation in subsequent proceedings to set standards for individual carcinogens^ To be sure, these issues may be reconsidered as science progresses, but only a 1 proceeding- to amend the generic airborne carcinogen rules. The model for such action is the cancer policy recently adopted as a binding standard by the Occupational Safety and Health Administration.-. [45 Fed. Peg. 5001-5296 * (Jan.22, 1980)3 After the most exhaustive and thorough rulemaking proceedings imaginable, OSHA adopted binding scientific ** 'III-3 and policy conclusions that may- be reconsidered only- when significant evidence of a scientific breakthrough, 'is- presented t , and only as an amendment to the generic cancer policy* The central generic principles- the policy roust adopt and foreclose from reconsideration in substance-specific standard-setting proceedings are:' That positive results either in at least one human/ epidemiological study or xn at least one laboratory study in any mammalian species- are sufficient for regulatory purposes to establish that a substance ' is a human carcinogen* That no safe level of exposure to a carcinogen can be identified, , jrhat positive studies-roust override studies that 30 purport to show* no effect/ since the probability (A Co CO of failing to detect- real effects in epidemiologicaJ^ and animal studies is very high. - That in interpretinc the results*o animal tests: , . ppsitiye results are Valid even if the tumors are found only- at doses higher than those to which people are exposed through air pollution. Such doses are necessary to detect carcinogens in small groups of animals, . tumors found in either sex and at any- organ site are sufficient to establish that substance as a human, carcinogen, , malignant and purportedly "benign." tumors must _be^given equal' weight. . , an increase in the rate of spontaneous tumors , or appearance of such tumors earlier than in unexposed animals, is sufficient for a positive finding. R&S 113563 XXI--4 . positive findings in tests by inhalation and by other routes of exposure are entitled to equal weight. . so-called "initiators" and "promoters" should not be distinguished and regulated differently. scientific understanding of metabolic and pharmo- -nrrfc yafe ariegrla-fce trn rpQni: hman g&rcinogenicrty or a suastance snown to cause cancer in animals. . results at the traditional level of "statistical significance" (95 percent confidence) are suffi cient, but not necessary, for a positive finding - of carcinogenicity. For example, the appearanc of rare tumors can warrant a positive finding even if the incidence is not "statistically significant., - That regarding short-term or in vitro tests for mutagenesis: positive results in a battery of such tests indicate a substantial risk of human carcino-- genicity. ^ ___ positive results support positive findings in human or animal studies. / positive results can support the conclusion that a substance is carcinogenic when the animal or human evidence alone would not. positive results should give substances a high priority for animal or human studies, if these have not been performed. negative results should be given no weight unless the substance is tested in a battery of short-term tests known to produce positive results for similar substances. * when a complex mixture of chemicals yields positive n results in'a battery of such tests, and when it is reasonably anticipated that multiple constituents of the mixture are independent mutagens, such results ^are sufficient alone to establish that the m? is a'human carcinogen. - III-5' That regarding structural relationships among, chemical substances: structural similarity to known carcinogens, supported by positive results inshort-term * mutagenicity tests, or by human or animal evidence not alone sufficient, can establish that a substance is a human carcinogen. . chemicals structurally similar to known carcina gens should have a high priority for further testing. HKDC and EDF will document the support for these principles in further submissions before the close of the period for supplemental comment. R&S 113564 IV-1 IV. DETERMINATION OP APPROPRIATE CONTROL LEVELS FOR AIRBORNE CARCINOGENS 1. The Failure of the EPA Proposal to Address the Statutory Criteria Contained in 5112. Once a pollutant is listed under 5112# the most important determination remaining is the appropriate degree of control to satisfy the statutory requirement. Section 112 of the Clean Air Act requires EPA to establish a standard which will provide an "ample mm-rain nf to protect the public health. ' This requirement of an ample margin of safety stands in sharp contrast to the require ments contained in 5108 and 5109 which require only an "adequate margin of safety" for non life-threatening pollutants. Regrettably, the approach suggested by the Agency for determining emission standards which will protect public health with an ample margin of safety is at serious variance with the statutory requirements. .The Agency's proposal for meeting the statutory requirement relies on a two-step process. First, the Agency determines "best available technology" for both new and existing_sources after considering the economic, energy and environmental impacts of various contr 1 - technology options. For new sources, the Agency states that "for practical purposes this level of control-^.- . will, as a minimum, be quivalent to that which would be selected as the basis for a new source performance standard (NSPS} tinder 5111. The requirement of 'best available technology' for new sources would consider 'economic feasibility' and would not preclude new construction.1 1. 44 Fed. Reg. 686S1 (Oct. 10, 1979). R&S 113565 IV--2 Tor existing sources, the Agency states: Caoj The selection of BAT for existing sources may require consideration of the technological problems associated -- with retrofit and related differences in the economic, __ energy, and environmental impacts. In practice BAT for cn existing sources would consider economic feasibilty a? and would not exceed the most advanced level of technology that at least most members of an industry could afford without plant closures. 1/ k Needless to say, neither of these.specifications for the initial deter ruination of appropriate control levels tracks the statutory languageJ Bather these are determinations which rely heavily on the considera tion of economic and other factors unrelated to the basic health issues.. As the Agency candidly noted, these procedures in fact be come more akin to determinations under 110 and 111 than 112. By themselves these determinations are obviously inadequate to satisfy the statutory criteria. Zt is only in the second phase of the control determination . process that EPA even attempts to address the statutory health criteria. In this second phase, the Agency proposes to use the -' / quantitative risk assessment developed for each chemical to determine _ whether the risk remaining after the application^ of^best available technology to the source category is unreasonable. While this '' / - . approach might be justified under other" statutes which require EPA to regulate only if it ean establish-an unreasonable risk, this approach simply does not track 112 of the Clean Air Act. Section 112 requires the Agency to- establish standards which .provide an ample margin of safety against the risk identified -- in g case of suffering incapacitating or irreversible health effects1 1. Id. r &S 113567 IV-3 from specific pollutants. By definition, any margin of safety is intended to make certain that no one suffers the threatened conse quences. On the other hand, the EPA approach not only provides no margin of safety, but rather explicitly accepts a certain amount of residual risk to certain individuals. Needless to say, for these people*there is no margin of safety, ample or otherwise. Even more disturbing, EPA admits that it cannot define the term "unreasonable risk" in advance. The Agency is able only to state that among the factors to be considered are " (1) the range f total expected cancer incidence and other health effects.in existing and future exposed populations to the anticipated operating life of existing sources; (2) the range of health risks to the most exposed individuals; (3) readily identifiable benefits of the substance or activity; (4) the economic impact of requiring additional control measures; (5) the distribution of the benefits of the activity versus the risk it causes; and (6)- other possible health and environmental effects resulting from increased use of substitutes.Needless to sav."the factors listed above are not / mentioned in 112, nor are they mentioned in the - legrlsla-H v** history. Indeed, the whole concept of unreasonable "residual risk is alien to t S112. In effect, EPA has attempted to rewrite the statute to suit its own conception of an adequate regulatory scheme. Nowhere has the Agency squarely addressed the issue of what constitutes an ample margin of safety for airborne carcinogens. To address this issue, it is necessary to return to the statutory language itself as well as the pertinent legislative history. IV-4 2. Legislative History of 5112 of the Clean Air Act- Because EPA has strayed so far from the original congressional intent, it may be useful to review the basic legislative history f S112, particularly the legislative history on the'ample margin of safety requirement. Section 112 of the Clean Air Act originated with S115 of the Senate Bill, S.4358. This bill required the Secretary of the Department of Health/ Education and Welfare (later the Administrator of EPA) to publish a "list of those air pollution agents or comb in a- tion of such agents which available material evidence indicated were hazardous to the health of persons and which should be subject, to a pro hibition or emission standard established under this section." Like 112 in the present lawr 5115(b) of the Senate Bill defined a hazardous * * * pollutant as one whose "presence chronically or ^intermittently, .in _trace_concentrations in the ambient air, either alone or in combination with other agents/ causes or will cause/ or contributes, to / an increase in mortality or an increase^in serious irreversible or incapacitating reversible^damacre to health."--*1^ * * To emphasize its concern over the hazards presented by these * * pollutants, the Senate Bill' created a presumption that emissions of hazardous air pollutants shouldJje'prohibited altogether unless "the Secretary finds under paragraph 2(B) of this subsection that a departure from such prohibition for any stationary source will not be hazardous to the health of the persons. . . -in short, the Senate Bill created a clear presumption- that a prohibition on emissions would be issued -- that is, a zero emission standard. A departure from this flat prohibition, that is the promulgation R&S 113568 1. S.4358, 115, 91st Cong., 2d Sess. (1970), reprinted in "A Legislativ History of the Clean Air Act Amendments of 1970," Vol. 1, Senate Committee on Public Works, 93rd Cong., 2d Sess. (1974), at 565--66. > Td . R&S 113569 IV--5 an emission standard which allowed some emissions, was authorized only if "the S cretary finds, has d on a preponderance of the evidence on the record compiled at such hearing, . . . that the agent is hazardous to health but a departure from the proposed prohibition will not be hazardous to health, and he shall publish an emission standard for such agent."--1/ a EDF and NRDC submit that the only situation where the "agent itself is hazardous to health but a departure from the proposed prohibition will not be hazardous to health" is one where the agent is hazardous to h alth only at certain levels and does not present a hazard to health at levels below that concentration. Clearly, airborne car cinogens do not fit into this category. The Agency itself has stated on numerous occasions that there is no threshold level for carcinogens below which no health hazard .exists. Indeed, there is evidence that the Senate was especially con cerned with pollutants which produced adverse health effects at even very low levels. For example. Senator Muskie stated:_____ The bill provides the Secretary with the. authority to prohibit the emission of hazardous substances. The Committee was presented with strong evidence that any level of emission of certain pollutants may produce adverse health effects that cannot be tolerated. 2/ In short,- it' is apparent that the Senate Committee clearly understood the difference between hazardous air pollutants which have definite health effects thresholds and those for which "any level of emission ... may produce adverse health effects. ..." In the latter case, the Senate clearly expected EPA to prohibit emissions altogether1 2 1. S. Rep. No. 91-1196 at 58; Legislative History, supra, Vol. 1 at 458. 2. 115 Cong. Rec. 32902 (Sept. 21, 1970). R&S 113570 . IV-6 * unless industry could make a convincing case that a threshold did in fact exist. . The House Bill contained similar language. The House, however, focused primarily on hazardous emissions from new sources. Section 112(B) of the House Bill set up two categories of new emission - s urces -- those that emit extremely hazardous pollutants and those which emit'other pollutants which imposed a less serious threat, to health. . Under the House Bill, EPA was required to set emission standards which would provide that: (1) If such emissions are extremely hazardous to health no new source of such emissions shall be constructed or operated, except where (and subject to such conditions as he.deems necessary and appropriate) the Secretary makes a specific exemption with respect to construction and operation. (2) In the case of other emissions', any new source of such emissions shall be designed and equipped to pre vent and control such emissions to the fullest extent compatible with the available technology and economic feasibility, as determined by the Secretary. 1/ In short, the House approach was similar to that of the Senate. The House created an explicit presumption that new sources of ex tremely hazardous emissions were to be prohibited altogether, unless the Secretary made a specific finding that their emiss ionswould not endanger health. Ironically, the approach^suggestecfby EPA in the proposed policy to regulate emissions of hazardous pollutants is much closer to that suggested by the House for the regulation of "other emissions," emissions which the House recognized were clearly much less hazardous and did not threaten irreparable injury. Significantly, it was only in the case of these other emissions 1. B.R.17255, 91st Cong., 2d Sess. 5 (1970); Legislative History Vol. 2, at 920-22. that the House was willing to have EPA consider the economic feasi bility Of availabl control techn logy. * The section of the Conference Bill which ultimately became 5112 tracked both the House and Senate thinking. The Conference substitute specifically adopted the Senate provision for emission * standards and recognized that such standards must provide "an ample t margin of safety" to assure public health protection. If there was t any remaining confusion as to whether an emission standard under the Conference Bill could require zero emissions, Senator Muskie dispelled that confusion. In an analysis of the Conference Committee Report prepared by Senator Muskie entitled "Discussion of Key Provisi ns the Senator stated: The standards must be set to provide an ample margin of safety to protect the public health. This could mean effectively that a plant would be required to close because of the absence of control techniques. Zt could include emission standards which allowed for no measurable emissions. 1/ In short, 5112 as finally adopted by the Congress created a r,__ v presumption of a zero emission standard for those hazardous pollutants 'v which had no identifiable health threshold. The presumption could be overcome in the course of rulemaking, but the burden of "pro was placed on those who would alter that presumption. In any event, the final emission standard was required to protect public health with an ample margin of safety.1 R&S 113571 1. "Discussion of Key Provisions," reprinted in Legislative History, supra, Vol. 1 at 133. XV-8 R&S 113572 3. Application of the "Ample Margin of Safety" Requirement to Airborne Carcinogens Having reviewed the legislative history on the ample margin of safety requirement, the next step is to apply this requirement to HPA's proposed policy on airborne carcinogens. As noted in the intro- dnction to this section, EPA has yet to come to grips squarely with the issue of what constitutes an ample margin of safety for an. airborne carcinogen. Rather than face this issue directly, the Agency has chosen to adopt a totally different standard -- that of unreasonable residual risk -- to determine whether controls beyond the equivalent of NSPS or PACT in the case of existing sources are called for by the statute. The. use of the unreasonable residual risk criterion explicitly accepts the fact that some people will contract cancer, even after BAT controls have been applied to new and existing sources of airborne carcinogens. This willingness to accept a cer-fr^iTM number of cancer deaths stands in sharp contrast to the statutory command that a standard protect affected individuals with an ample margin of safety. ) Ironically, the approach proposed by EPA to regulate airborne carcinogens may actually afford the population relatively less health protection than the approach taken to regulate pollutants covered by 5109 standards. Under 5109, EPA is required to adopt ambient air quality standards which provide an "adequate margin of safety." Although the safety factors which have been incorporated into the existing ambient air quality standards are quite modest, it can be argued that at least these standards do provide some margin of safety, that is, the standard is set sufficiently below the threshold at which health effects have been identified that even susceptible' members f the population are protected from the r &S 113573 IV-.9 observed effects. ^ In contrast.t the foregoing approach, EPA is explicitly V proposing to reject its responsibility to protect all members of the population with some margin of safety from the threat posed by airborne carcinogens. Because the Agency has been unable to identify a threshold for carcinogens, it has simply decided that it will accept a certain number of deaths and ignore the ample margin of safety requirement. . In short, the Agency feels it has no responsibility to protect each, exposed member of the population. Rather, it assumes that it has fulfilled its statutory responsibility if it protects most members of the population, even if the remaining residual members contract cancer. For these residual members of the .. population, there is obviously no margin of safety, ample or therwise. The foregoing analysis is also supported by applicable case law on the ample margin of safety requirement. Although there has been no explicit holding on section 112, there is case law interpreting the "ample margin of safety requirement" in 307 of \ the Federal Water Pollution'Control Act, which is directly applicable. . In Environmental Defense Fund v. Environmental Protection Agency, 598 F.2d 62 (D.C. Cir. Nov. 3, 1978), the D.C. Circuit ruled directly * on the ample margin of safety requirement contained in 5307. Even * more important, the court referred directly to the use of that term in the Clean Air Act. To quote the court's opinion:1 1. EFA's recent revision to the ozone standard makes it arguable whether the present standard provides any margin of safety for susceptible populations. Nevertheless, the Agency has at least accepted its responsibility to set a standard with an adequate mar gin of safety in theory. XV-10 R&S 113574 The "ample margin of safety" `standard was taken originally ' *` from technical jargon for use in section 112 of the Clean Air Act Amendments of 1570, the hazardous pollutants provision. See note 28 supra. One academic commentator has traced . the history of the use of this term: The term [margin of safety] seems to havebeen borrowed in the first instance from the field of engineering. It is a common engineer ing practice to incorporate a factor of safety info the design of a structure to "ensure that under full working loads the stress will nowhere ekceed a safe working limit." The safety factor is meant to compensate for uncertainties and variabilities in design, materials, workmanship, and so forth; the greater the variability, the larger the factor of safety. ^ That the use of the term "margin of safety" in section 307 (a) was similarly meant by Congress to take into account and compensate for uncertainties and lack of precise predictions in the area of forecasting the effects of. toxic pollutants is confirmed by reference to the Clean Air Act. FWPCA section 307 (a) is very similar to section 112 of th Clean Air Act. Among the similarities'is'the requirement in both that any standard set should provide "an ample margin of safety." 1/ In short, the court makes it clear that_its interpretation of the ample margin of safety requirement in 5307 of the Federal Water Pollution Control Act is taken from its understanding of the comparable phrase in 5112 of the Clean Air Act. To, place the ample margin of safety requirement in context, the court also refers to the adequate margin of safety requirement in 5109 dis cussed above: Congress . . . provided in section 307(a) of the FWPCA, as in section 112 of the Clean Air Act, that the "margin of safety" be "ample." This is in contrast to the requirement of section 109 of the Clean Air Act that primary air standards assure an "adequate margin of safety"; the require ment of the Safe Drinking Water Act that con taminant levels for public water supplies be set1 1. EDF V. -SPA, 598 F.2d at 80-81. K* * ' IV-11 at a level to assure "an adequate margin of safety"; or the requirement of section 303(d)(1)(C) of the.[1972 Act] that waste load allocations 0 be established at simply "a margin of safety." "Ample" is defined as "abundant; plentiful; more than adequate." Clearly Congress intended that in dealing with toxic pollutants that pose a threat to human health, margins of safety should be generous to ensure protection of human health and aquatic ecosystems to the greatest extent possible. Hall, The Control of Toxic Pollutants Under the Federal Water Pollution Control Act Amendments of 1972, 63 Iowa L. Rev. 609, 629, 630 (1978) (footnotes omitted) (emphasis in original). 1/ Of particular importance in the case of airborne carcinogens is the court's statement that "[t]he foregoing analysis demonstrates that the term 'margin of safety! was intended to provide protection 'against hazards which research has not yet identified.'"--^ it is precisely because of the difficulties in quantifying the risks posed by airborne carcinogens that the court concluded that a margin of safety was necessary in the setting of standards to control these substances. ------ The extent of government authority to protect public from carcinogens is demonstrated by the court's discussion of a series of cases regarding OSHA's attempts to regulate carcinogens / , in the workplace. In commenting on these cases, the coart noted: These cases demonstrate the inevitable tension attending regulation of carcinogens. Frequently, such regulations have severe economic impact. Indeed, sometimes, as alleged by industry petitioners in this case, such regulations may jeopardize plants or whole industries, and the jobs depending on them. In such circumstances, the temptation to demand that the agency furnish conclusive proof of carcinogenicity as support for the regulations is great. However, the decision to delegate authority to an agency to control suspected R&S 113575 1. Id. at 81. .2 Id. IV-12 carcinogens is a legislative judgment that is not open to question in this court. Congress's direction to EPA to protect against incompletely understood dangers could not be carried out if we were to adopt the proof requirements advocated by industry petitioners. ... If regulation were withheld until the danger was demonstrated conclusively, untold injury to public health could result. Accordingly, we find that Congress has allowed EPA to support a prohibition on the basis of strongly contested and merely suggestive proof.1/ Perhaps most significant, the D.C. Circuit has explicitly re jected the consideration of economic and technical feasibility in setting standards for carcinogens. As noted above, EPA proposes to rely very heavily on technical and economic feasibility judgments in defining best available technology and in determining what additional controls it might consider if it determines that an unreasonable r sidual risk remains after application of BAT. In Hercules, Inc, v. Environmental Protection Agency, 568 F.2d 91 (D.C. dr. 1978) dec!* the same day as EPF v. EPA, the court explicitly held that the anple margin of safety requirement of 5307 of the FWECA does not authorize EPA to consider economic and technical feasibility factors /^ in setting standards. In reaching this conclusion, the court went through a lengthy analysis comparing the technology-based provisions of the FWPCA with those provisions^designed to control toxic substances Because a similar distinction exists in the Clean Air Act, it may be useful to review the court's reasoning. Indeed, the court in its analysis, once again, relies on the distinction in the Clean Air Act between 55111 and 112. 03 <71 1. EPF v. EPA, supra, at 88-89 -4 R&S R&S 113577 "IV-13 The legislative background explains why Congress focused on public and environmental protection, rather than discharge control technology, in the setting of toxic standards. The regulatory scheme is similar to that of the Clean Air Act Amendments of 1970, Pub.L.No. 91-604, 84 Stat. 1676, which distinguished between pollutants subject to technology-based regulation under section 111, and hazardous substances, subject to health-based regulation tinder section 112. Recognizing that "certain pollutants" required special treatment because of risk to health, Congress enacted section 112, dealing with hazardous pollutants, without provision for considerations of feasibility. 1 Senate Comm, on Public Works, 93d Cong., 2d Sess., A Legislative History of the Clean Air Act Amendments of 1970 (hereafter. Clean Air Legislative History), at 133 (Comm. Print 1974) (remarks of Sen. Huskie) (The Senate subcommittee on Air and Water Pollution "was presented with strong evidence that any level of emissions of certain pollutants may produce adverse health effects that cannot be tolerated." Id. at 227.). 1/ After reviewing similar language in the legislative history of the 1972 Federal Water Pollution Control Act Amendments, the court concluded: In light of this clear statutory wording and legislative history, it is not surprising that numerous commentators agree that considerations of technological and economic feasibility do not play a part in standard--setting for toxic substances. 2/ _____ In reaching this conclusion, the court explicitly recognized that " 'n the technology may not presently exist which will reduce levels of carcinogens to levels low enough to provide an ample margin of safety. The court noted that the Supreme' Court in Train v. NHDC, 421 U.S. 60, 91 (1975), determined that: ; . . certain sections of the Clean Air Act imposed require ments of a "technology-forcing character." In Union Electric Co. v, EPA, 427 U.S. 246, 257, 96 S.Ct. 2518, 49 L. Ed.2d 474 (1976), the Court elaborated that such statutes "are expressly designed to force regulated sources to develop pollution control devices that might at the time appear to be economically or technologically infeasible." 3/1 1. Hercules, Inc. v. EPA, supra, at 112. 2. Id. See also discussion at 111, "...nor is there present any ^erm commonly used to den te a feasibility consideration..." 3. Id., n.41. 17-14 In response to Velsicol's argument that, if feasibility is not considered in setting standards for carcinogens, the resulting re lations will be promulgated at irrationally strict levels, the court noted that: However, the congressional selection of factors is a legislative determination that the need of the public and the environment for protection from toxic chemicals is more important than the problems of stringent regulation. This congressional determination is a national response to the dangers presented by toxic substances. The meaning of the statute being clear, it is not this court's prerogative to impose considerations of feasibility. See Tennessee Valiev Authority v. Hill. 437 U.S. 153, 98 S.Ct. 2279, 57 L':Ed.2d 117 (1978); Union Electric Co. v. EPA, 427 U.S. 246, 96 S.Ct. 2518, 49 L.Ed.2d 474 (1976). 1/ In short, it is now well settled that EPA may not consider questions of technical and economic feasibility in setting emission standards under S112 of the Clean Air Act. The Federal Water Pollution Control Act cases are directly applicable since the court's analys^^ in those cases proceeds from its understanding of Congress' intent regarding similar provisions in 5112 of the Clean Air Act. Finally, EPA is in no position to argue that the ample margin of safety requirement in 5112 has implicitly been overruled by the *v passage of the Toxic Substances Control Act and the adoption f /' unreasonable risk criteria for-regulation in. that Act. The D.C. Circuit in EPF v, EPA dealt explicitly with this issue. In that case industry had argued that the adoption of the Toxic Substances Control Act effectively "preempted" EPA's authority under 5307 of the 1972 Federal Water Pollution Control Act Amendments. They 1. Hercules, Inc. v. EPA, supra, at 112, R&S 113578 asserted that EPA's ban on the discharge of polychlorinated biphenyls (PCBs) into waterways made manufacturing and processing f PCBs impossible and that TSCA's phaseout timetable and exemption authority showed that Congress did not intend PCBs to be completely and imme diately phased out. The court rejected the industry argument and agreed with EPA. To quote the court: Strictly speaking, the claim made by industry petitioners is one of "repeal by implication," the overriding of one statute by a later enactment sub silentio, rather than . a claim of "preemption." "Zt is, of course, a cardinal principle of statutory construction that repeals by implication are not favored." United States v. United * Continental Tuna Corp., 425 U.S. 164, 168-69, 96 S.Ct. 1319, 1323, 47 L.Ed.2d 653 (1976) (collecting cases). "Implied repeals occur if two Acts are in irreconcilable conflict." Runyon v. McCrary, 427 U.S. 160, 172--73 n.10, 96 S.Ct. 2586, 2595, 49 L.Ed.2d 415 (1976) (emphasis added). When confronted with statutes that are "capable of co existence, it is the duty of the courts, absent a clearly ; expressed congressional intention to the contrary, to re gard each as effective." Regional Rail Reorganization . Act Cases. 419 U.S. 102, 133-34, 95 S.Ct. 335, 353, 42 , L.Ed.2d 320 (1974) (quoting Morton v. Mancari, 417 U.S. 535, 551, 94 S.Ct. 2474, 41 L.Ed.2d 290 (1974))- Not only is there no "clearly expressed congressional intention", that TSCA repeal EPA's-prior regulatory authority, and no "irreconcilable conflict" between TSCA and the . 1972 Act, but, in fact. Congress carefully and explicitly harmonized TSCA with prior enactments. Section 9(b) of TSCA, 15 U.S.C. 52608(b) (1976) (emphasis added), provides, in part: -----' R&S 113579 If the Administrator determines that a risk to health or the environment associated with a chemical substance or mixture could be eliminated or reduced to a sufficient extent~bv actions taken under the authorities contained in . . , other Federal laws, the Administrator shall use such authorities to "protect against such risk unless the Administrator determines, in the Administrator's discretion, that it is in the public interest to protect against such risk by actions taken under this chapter. This subsection shall not be con strued to relieve the Administrator of any require ment imposed on the Administrator by. such other Federal laws. R&S 113580 IV-16 Section 9 (b) leaves EPA the choice of regulating toxic * substanc s under TSCA, other statut s (such as the 1972 Act) , or both. Throughout debate over passage of TSCA, representatives of the chemical industry urged that it would be unnecessarily duplicative of existing authority, and that EPA should not be given a choice among multiple regulatory authorities. Congress rejected those arguments which petitioners, in effect, now seek to resurrect. Congress determined that a choice among regulatory authorities was necessary so that EPA could use the most effective means available to combat unknown and potentially extreme risks from toxic substances, and that judicial review of EPA's choice was inappropriate. Our interpret tation of section 9(b) is reinforced by section 6(e)(5), 15 U.S.C. 52605(e)(5) (1976), which states: "[t]his subsection [6(e) (governing PCBs) ] does not limit the authority of the Administrator, under any other provision of" this chapter or any other Federal law, to take action respecting any polychlorinated biphenyl." 1/ In the case of 5112, the argument is even more compelling. Con gress amended the Clean Air Act in 1977 following the passage of the 1976 Toxic Substances Control Act and decided to retain, the ample margin of safety language contained in 5112.-2/ Congress did emend 5112 but increased EPA's ability to regulate carcinogens by ai orizing work practice standards where emission standards were not feasible. In no way did the amendment to 5112 reduce EPA's re sponsibility to set standards to protect health with an ample margin of safety. Thus, there is no statutory basis in TSCA or any other subsequent-legislation to argue that the ample margin of safety requirement in 5112 has "been repealed by implication. Unless EPA ** find some firm statutory or other support for its unreasonable residual risk standard, the approach must fall.1 1. EDF v. EPA, supra, at 76-77. 2. In addition to explicitly retaining the "ample margin of safety requirement," Congress expanded EPA's authority by directing it to regulate not only substances which "cause" irreversible health effects, but also those which may "reasonably be anticiwj pated" to do so. This further emphasizes the precautionary % appr ach embodied in 5112. . IV-17 4. Proposed Approach for Determining Appropriate Control Levels for Carcinogens Regulated under S112 Having reviewed the deficiencies of the EPA proposal for deter mining control requirements in light of the legislative history of 5112 and applicable case law, the next step is to present a modified approach which will more closely meet the requirements of the law. DF and NRDC believe that a number of approaches axe feasible, but will describe them only briefly in this section of the comments. A more detailed discussion will be presented in the public hearings and in subsequent rebuttal comments. To clarify the analysis, we will divide our comments into requirements applicable to existing sources and those applicable to new sources. Furthermore, in this section, we will deal primarily with proposals designed to reduce emissions from discreet process sources and defer discussion of EPA's proposed housekeeping requirements to reduce fugitive emissions until the next section. Finally, much of the discussion which follows will focus on control requirements applicable to the synthetic organic chemical industry, since this industry is the source of many of the potential 'candidates for regulation under 5112. However, EDF and NRDC wish to' emphasize that control of inorganic carcinogens under \ S112 should receive equal priority. Because control requirements for these latter sources are'somewhat different, however, they will not be stressed in these comments and will be reviewed in more detail lat r at the public hearing. R&S 113581 A. Control Requirements Applicable to Process Bnissions From Existing Sources of Airborne Carcinogens From the foregoing review of the legislative history and applicable case law, it would be relatively easy to conclude that IV--18 Congress intended EFA to establish a zero emission standard for any airborne carcinogen for which evidence of a threshold was lacking. Needless -to say, this would cover virtually all of the known carcinogens# since thresholds have not yet been established for any . .of This# of course# does not mean that thresholds could not exist- It may be that as additional research is completed that' thresholds will be identified for at least some carcinogens,. In the meantime# EDF and NBDC suggest that the Agency adopt the approach recommended in the legislative history -- that is# that m presumption be created for a standard of no measured emissions unless the affected industry can establish that such a standard is . not v achievable through either product substitution or process change. To make certain that a decision on appropriate controls is made in a timely fashion# EDF and NKDC suggest that the proposed policy provide that a no measured emission standard 9t R&S 113582 ao into effect within one year of the listing of a chemical# unless an alternate standard has been agreed upon by the affected industry and the Agency. This will give industry a strong incentive to quickly review the opportunities for product substitution as well as process change. If the companies can demonstrate to EPA that neither approach is available to reduce emissions to essentially the zero level# the industry would have the burden of demonstrating that no control technology was available which could achieve the no measured emission level. " If control technology capable of achieving no measured emissions is not currently available, the Agency should establish a schedule of ..phased reductions leading to the zero emission goal. This schedule would'require immediate installation of the best available R&S 113583 -IV-19 technology and would mandat specific additional reductions along a phased schedule. This approach is consistent with the require ments of 5112 and. is further authorized by the Supreme Court1 s holding in Union Electric Co. v. EPA, 427 U.S. 246, 257. (1976) where the Court specifically found that certain sections of the Clean Air Act are expressly designed "to force regulated sources to develop pollution control devices that might at- the time appear to be economically or technologically infeasible." This approach will also reduce the resource burden on EPA considerably and allow the Agency to regulate many more carcinogens under 5112 than would otherwise be possible. If EPA adopts the approach outlined in the proposed policy, it is quite likely that a tremendous expenditure of time and resources would be required to establish even the baseline BAT standards. EPA has proposed to go through essentially the NSPS process for each pollutant sub ject to regulation under 5112. EDF and NRDC feel that this approach is b th unwise and ultimately unworkable. Effectively, it will mean that development of even proposed standards may take from one to two -years after a decision has been made to regulate a particular chemical. Furthermore, EPA itself is assuming a completely unnecessary burden of proof which is self-imposed. The legislative history makes clear that a prohibition on emissions is to be the presumption in cases of airborne carcinogens. If industry wishes to challenge that presumption, it should carry the burden of proof. EPA's only responsibility is to determine whether a listed substance meets the health criteria outlined in 5112. If it does, and if no thresh old has been identified, then the burden of proof as to appropriate control requirements shifts to industry. If industry chooses to not R&S 113584 * IV--20 not assume this burden, then a "no measured emissions" standard would automatically go into eff ct 360 days after listing. B. Control Requirements Applicable to New Process^/ Emission Sources _____ The approach for determining appropriate emission limits for new sources of airborne carcinogens is similar to that used for existing sources, except that the additional problem of expansion at existing sites must be addressed. For new "greenfield" sites, the presumption of a no measured emission standard applies, unless industry can conclusively demonstrate that substitutes for the product are . not available or that a combination of process change and control technology cannot achieve a no measured emission level. For expansions at existing sites, a different set of rul s must be adopted. EDF and NRDC suggest that the Agency extend its existing emission offset policy to these modifications at existing sources. Under this approach, a facility could expand only if it could demonstrate that the additional emissions generated by the expansion would be offset bv an additional reduction in_ emissions, of the same hazardous pollutant from the existing facility. This . offset approach will preclude any increase in ambient concentrations of hazardous pollutants at existing sites. Needless to say, a simple best available control technology standard for expansions at existing sites would not ensure this goal. EDF and NRDC submit that the prTy alternative to such an offsets approach is a flat prohibition on expansions of sources at existing sites. Adoption of an offset 1. "Process," as used herein, refers to emissions from discreet vents or "stacks" intended to convey emissions to.the outside air. R&S 113585 XV-21 policy, however, should make this unecessary if existing sources continue to move toward a zero or no measured emission goal. The proposals outlined above depart significantly from those .contained in the EPA airborne carcinogen policy. EOF and NRDC do not believe that the provisions applicable to new sources set out in the policy meet the requirements of S112. Although the EPA policy would establish presumptive emission standards based solely on health effects considerations, the Agency has provided for an elaborate exemptions process which will effectively ensure that most sources are regulated under one of the exceptions, rather than under the presumptive emission standard. Moreover, the risk avoidance criteria, by providing for a waiver from the presumptive emission standard, will actually encourage the - proliferation of facilities emitting carcinogens in presently lightly populated areas. Needless to.say, this is not an emissions control approach; it is an emission dispersion approach. Unfortunately, EPA has no authority to control the use of land around new sites. It cannot impose subsequent growth controls and prevent people . from moving into areas around new sources of carcinogens. If there is any doubt about this proposition, one only has to consider the experience of airport operators who have located facilities in previously "remote" areas. Dulles Airport in Washington, D.C. is an immediate example that comes to mind. In short, EPA simply cannot prevent people from "coming to the nuisance." The only solution is either a prohibition on new sources or an extremely tight set of standards. V-l R&S 113586 V. TEE NEED TO ELIMINATE ARBITRARY DEPENDENCE ON UNRELIABLE QUANTITATIVE RISE ASSESSMENTS . : '* # Under the.proposed roles EPA would continue to.mak unwarranted use of unreliable quantitative methods for assessing * , ** -human risk. ' This is neither required* lay the Clean Air . Act nor a rational exercise of EFA*s discretion^. .However desirable quantitative accuracy may`be in principle/ the scientific knowledge necessary for reasonably reliable precise estimates simply is not yet available. Current knowledge about chemical and radioactive carcinogenesis is primarily qualitative# not quantitative. Through epidemio-- .logical and animal studies, gnfamanm* *t -t a n tests, and studies of - - structural relationships, it is possible to identify reliably substances that pose a cancer risk for humans. It is not cur-. rehtly possible, however,, to predict the size, of that risk with anywhere near the precision and reliability needed for the esti-- ' mate to be useful in making regulatory decisions. We summ arize * .. ** ** i here the reasons why quantitative risk assessments may not- . rationally be used as EPA has proposed,, and we will file supple mentary comments with fnrther information, on "these.paints. A. The Unreliability of Quantitative Risk Assessments Derived from Animal ` Studies Most human cancer hazards axe identified through tests bn laboratory mice, rats, or other rodents. Though qualitatively reliable, these tests cannot provide usefully precise.and reliable V-2 R&S 113587 quanti.-tati.ve estimates of human risk.- The sources of uncertainty ace: ' I. ' -- Hlgh-to-Low Dose Extrapolations* For practical reasons, laboratory studies mast be conducted . with small groups of- animals, usually not more . than several hundred, usually at doses much higher than those to which the general public is exposed. -The first obstacle for quantitative risk assess ments is to extrapolate the results of the high- dose tests to the incidence likely in animals at lower doses. Reflecting scientific, uncej. Lainties over the, mechanisms of cancer induction, there .`are numerous mathematical models for extrapolating -to low dose risks. Although the.models all "fit" -the high-dose data reasonably. well, they yield - -- //widely divergent estimates, of; the risks associated with low doses. For example, the major models . .. differ by a factor of 100,000 on.the dose projected .to cause one extra cancer in a million animals.!/ -- Scaling Factors. Additional uncertainty is added by the need to extrapolate across species '.lines from animals to humans. Rodents are smaller than humans. They live approximately l/35th as - long. Their rates of metabolism and cell division . are much faster. These and other differences com plicate quantitative comparisions. There-is uncer tainty whether the proper way to scale doses is ' on the basis of relative body weights, relative . cell surface areas, relative lifespans, or on some other basis Jj - -* Recent attempts to quantify the risk from saccharin graphically illustrate the magnitude of - the uncertainties attributable just to differences among models and scaling assumptions. Using the results of recent animal sutdies, the various plausible models, and different w^ians about ' bow- to cross species lines, the Rational Academy of Sciences attempted to quantify the effects of 1/ Cornfield, "Carcinogenic Risk Assessment," 19S Science 593, 694 (1977). 2/ "Problems of low Doses of Carcinogens," 64 Wash. Acad. ofSci. 63 (1974). ----------------------- V--3 consuming one can of diet soda per day* The NAS concluded that in each 50 million people there could be as few as 0.0007 cancers per year or as many as 3,640. This is a range of error more than 5 million-fold.*/ See Table V--1, on the following page. A more recent _ study calculates an 'even higher risk for erpo^ sure "to saccharin, extending the'range to _ / .r^O-millioh-'fold.f Estimates for the human risk associated, with exposure to one. part per million of vinyl chloride extend over a one million-fold range as well. See Table V--2. * * -.Differences- in Species? Sensitivity to Carcinogens. There is ample evidence that different species ---- -even different strains within species -- have markedly different sensitivities to carcinogens. - Animal species can exhibit differences of as much as two or more orders of. magnitude. The inter species differences are neither regular nor pre dictable. One test species can be more sensitive than another to one carcinogen, and less .sensitive to a second carcinogen. Few eompaxisions are possible between test animals and humans, since there is no direct human evidence for most car- cinogens. In some instances, however, the human species has been shown to be far more sensitive than the animal species that* reacted most strongly in laboratory experiments. These differences . are not accounted for in the above estimates of uncertainties attributable to the mathematlca 1 models and sealing assumptions. They widen the potential error even more. * ,' ` . * -- Individual Variation in Human Sensitivity. Even beyond interspecies differences, there is greater . individual variation in sensitivity among' humans than among test animals. This is a result of . the fact that humans are genetically heterogeneous, while test animals are bred for maximum genetic homogeneity. Because purposeful experimentation, on humans is ethically unacceptable, and because 1/ National Academv of SciencesSaccharin s Technical Assessment of Risks and Benefits, p. 3-72, -table 3-8 {197SJ. 2/ Miller and Howe, Saccharin -- The Bisks and Benefits, Nature 8 (1578). R&S 113588 3589 TABLE V-I O-ta Iwfcwnd Hwwn Rbfc from Siuaiii Inguaw < g/By (Tiew IUB 11 mom. am * n fin Jmwrfina. ant. Iff*. , "w-itrj-`rwr--t*---------* -- *--*--r ' ***-^1^*--*^***^ tW*------ Ifflt. .uz. Source; National Academy of Sciences, Saccharin; Te^TTcal -Assessment~of Risks and Benefits at 3-72 (1978).' TABLE V-2 Gtttnama f urottn* ftfa* firm* town a* I |n nqiCMaUl iwn llanofa. Source: OSHA, "Identification, Classification, and Regulation f Potential Occupational Carcinogens," 45 Fed. Reg. 5001, 5200 (January 22, 1980). . ' V-5 R&S 113590 no method of measuring human sensitivity to carcinogens is available, the range of human, sensitivities is an important source of impre- . cision of uncertain dimensions in quantitative xisJc estimates* - Differences in Metabolism, Pharmokinetics, and Repair Capabilities. ' ...................* * ` * Very little information is available on the role of metabolic, pharaokinetic, -and repair mechanisms in carcinogenesis, or on differences in such media-- . --misms within and between species. Nonetheless,, while there is enough basic similarity that ani-- mal sutdies are reliable means of identifying -human carcinogens, there are differences that are likely to affect the magnitude of carcinogenic response in different species. These differences * add further unpredictable uncertainty to quanti tative risk assessments. - Interactions and Synergi'sms. Animals are usually' exposed to only one suspect substance in .carefully controlled settings. Humans,.however, are exposed, to a wide variety of other chemicals and radiation, and other carcinogenic co-factors in almost infi nite combinations. Because of interactions among these co-factors, exposure to a carcinogen, can result in a much -greater incidence in humans than in animals. In addition, some risk factors interact syner-. gistically. The risk from exposure to both is far * greater than the sum of the risks from exposure to each alone. There is good reason to expect that * two--or-more--factor synergism is common. Yet, most synereisms have not been identified, let alone quantified, because little human epidemiological data is available and animals are exposed only to single substances. - Differences Due to Sporadic, Uneven Exposure vs. Constant Exposure. Test animals typically are exposed to a constant dose for a predetermined prolonged period, usually a full lifetime. People, however, tend to receive uneven exposures from day to day- Some may receive only a single or sporadic high doses. Some are exposed only in middle age, while others are exposed as children. The relation of .constant exposure to sporadic-, uneven exposures equalling the same total does is unknown. T sts in animals have shown that a single, massive dose of a carcinogen in a young, animal can `cause cancer later in life, but the f * V-6 quantitative relationship between such doses and constant exposure is highly uncertain. These uncertainties preclude deriving usefully precise quantitative risk estimates from animal experiments. $ Some of the sources of uncertainty are "wild cards", on which it is impossible -even to place reasonable outer boundaries. Estimates can be wrong by thousand-fold or million-fold factors. * -B. The' Unreliability, of Quantitative Risk.' Estimates Derived from Human Studies Similar limitations prevent derivation of reliable quan- ' -titative estimates from epidemiological studies. The most serious . limitation is that there is epidemiological evidence for only ~a few substances identified as carcinogenic in animal experiments. Beyond this, epidemiological studies are relatively insensitive? 'in such studies there are - inherently high probabilities of failing to detect a true effect at all, and of misrepresenting the mag- * . nitude of an effect even if it is identified. The reasons for the insensitivity and quantitative, inaccuracy of epidemiological studies are: - - Confounding Factors.. Epidemiological studies are performed in an uncontrolled setting, intro ducing many unknown -factors that can obscure true relationships between cause and disease. Most epidemiological studies are conducted among occu pational groups, who are exposed to chemicals at higher doses and more regularly than the general population. Nonetheless, many well-designed studies would fail to detect cancer caused by a chemical or radioactive substance unless it is occurring at a rate five times higher than nor- . mal. Even the best large*^pI3emTological studies R&S 113591 V-7 usually cannot reliably detect effects oc-. : curring at less than 50 percent above the .. normal rate.' Their insensitivity for idea- tifyiag carcinogens at all extends to the . . quantification of risk. ' - The Effects of Latency.' Latency periods -- the time between the first exposure and diag-.. aosis of diseases -- of 20 years or* more are for carcinogens. Studies which do not follow exposed subjects for their full lifetines can fail to detect a carcinogenic effect -entirely, or can miss the. bulk of late--occurring .cases. - Failure to Follow All Members of' the Exposed Group, ________ ' ........................... A common failure or occupational studies is the failure to follow employees who change jobs. Since ill employees are the most likely to leave employ, ignoring this factor can be significant. Their state of health, or even if they are still alive, is not determined. This tends to under-- . state the true risk. Follow-up problems are even greater for many non-occupational studies. - * - Ignorance of True Exposure Level's and Durations. Dose information is often highly speculative, especially for periods two or more decades ago. problem is especially severe for epidemio-- . logical studies concerning environmental pollu tants. Even monitoring and modelling done today is subject to serious errors. Without accurate dose information, no quantitative conclusions . ' can be made. - Failure to Relate Common Cancers to Their Causes and Failure to Detect Rare Cancers. Some kinds of cancer, such as lung cancer, are relatively common and have many causes. Frequently, such cancers are assigned in toto to smoking (or to unknown causes), when another substance is also a causal factor. In the opposite extreme, a sub stance may cause a relatively rare cancer at a rate that is below the sensitivity of epidemio logical studies. Both factors lead to underesti- - mating human risk by an uncertain amount. R&S 113592 V-8 -- Errors in Reporting. Few states have accurate, reliable systems for recording cause of death.' Even where organized records are kept, mistakes in diagnosis can distort an epidemiological . study.- ' - Errors in Describing the Study and Control or Failures to Adjust for Age, Socioeconomic Status, and Other Variables. ................................ "....................... Err"^ in ripggyifajng- the grremp exposed and' the - comparison control can introduce serious errors into quantitative estimates. The same result can occur from failures to adjust for age,' income and race, and other variables. ,* m. - * **_,` , The result of these sources of uncertainty is that epi demiological studies, even when available, can misstate the true carcinogenic risk by several orders of magnitude or more. C. The Unreliability of Exposure Monitoring and and Mcdellinc Data ................................. ' ` `A necessary element of quantitative risk assessments is accurate data on the magnitude of human exposure to airborne, carcinogens. Such data is needed to determine exposure levels from current emissions and to predict the levels to be ejected .from new emissions. But for hazardous air pollutants EPA does n't generally have, either the historical monitoring data base _ or modelling methods of sufficient. accuracy for reliable-quanti tative risk assessments. Modelling methods for even the conven-- * tional pollutants, such as TSS and SO2 are not capable of making predictions consistently within even a factor of two of actually monitored levels. Many carcinogenic air pollutants are reactive, and modelling for reactive substances' (such as photochemical oxidants) is even less advanced- R&S 113593 V-9 * R&S 113594 * The lack of precise monitoring data and modelling methods farther undermines the reliability`of quantitative risk assessments. When all the sources of uncertainties are considered together, huge errors in risk estimates are nearly unaviodable on -a frequent basis. D. ERA Itself Recognizes the Uncertainties that hake Quantitative Risk Assessments - Unreliable and Imprecise ....................... _ . * * NRDC and EDF find EPA's proposed extensive reliance on quantitative risk assessments all' the more difficult to under-- stand because EPA itself appears to recognize, or at least pay lip service to, the irreducible uncertainties inherent in quan titative risk assessments. The proposal states: The assumptions and procedures ... for . extrapolation and for exposure estimates are subject to considerable uncertainty. . Where only animal data are available to assess the magnitude of cancer risk to human populations, the differences in sus- *- ceptibility between animal species and **' humans, and the need to extrapolate dose- response data in very low ambient concen-- trations, result in risk estimates that must be regarded only as rough indications of effect. Uncertainty in exposure estimates arises from the use of limiting monitoring, pollutant transport models, mobility of the exposed population and other factors. In combining these exposure estimates with dose extrapolations to provide estimates of cancer incidence, the total uncertainties are increased. (44 Ted. Beg. 58649.) Our belief that this represents only lip service is reinforced by the absence of any detailed discussion, of the V-10 R&S 113595 uncertainties in the Statement of Basis, by the essence of any reference to weaknesses of quantitative risk assessments in Part D of. the proposed rule (44 Fed. Reg. 58654), and by EPA's blithe commitment to use these techniques anyway. ; E. Nothing in the Clean Air Act looses Quantitative Requirements...................... < EPA's current fascination, with quantitative risk assess- *-- ** meat methods finds no support in the Clean Air Act. Nothing in'Section. 112 imposes the prerequisite of a quantitative risk * * assessment before EPA may protect the public from a carcinogenic *k air pollutant. Congress recongized that under the current state of scientific knowledge cancer hazards can be identified even though their precise magnitude cannot be quantified. But Congress did not require EPA to delay regulation until this precise quan tification -is possible or force EPA to undertake sham quantitative analyses. Rather, because the public's lives and health are at' stake. Congress directed EFA to take highly precautionary,, pro- tective action to prevent cancer based on qualitative evidence *, ` identifying carcinogens, because that is the'only reliable evi- dence available. 4 The Clean Air Act requires EPA to list as hazardous air pollutants each substance that "may reasonably be anticipated to result in an increase in mortality or an increase in serious, irreverisble, or incapacitating reversible, illness.** [Section 112(a) (1), (b) (1) (A)} The Agency must swiftly propose and R&S 113596 V-ll promulgate standards providing the public health with "an ample margin of safety," unless it is proved that the substance "clearly" does not have those effects. [Section 112(b) (1) (B)l * Nothing in these phrases or elsewhere in the law or legislative history even suggests that EPA must be able to demonstrate the ' magnitude of the cancer hazard before regulating, or in determining how much to reduce emissions. F. Even if Balancing Risks and Benefits were Legal Under Section 112, Use of Quantitative Risk Assessments to Determine- "Unreasonable Residual Risk" Would be Arbitrary................... EPA proposes to use quantitative risk assessment as the centerpiece of determining whether * the "residual risk" remaining from emissions allowed by a BAT standard are "unreasonable." . ' EPA states that "despite their lack of precision" and " considerab^ uncertainties," the Agency has the "belief" that the. quantitative risk assessment methods "can ... provide useful estimates" for use in making such unreasonable risk determinations. [44 Fed. Reg. 58649] We have already shown that such a balancing of risks and. benefits is not authorized under Section 112 of the Clean Air Act- Even if FA*s risk-benefit analysis proposal were legal, the use of the grossly unreliable and imprecise quantitative risk assessments would be an abuse of discretion. Although HRDC and EOF strongly support EPA's commit ment to considering whether emissions of a carcinogenic air pol lutant can be eliminated completely by substitutions and process i changes -- indeed we believe such measures are required -- we do not believe that the unreliable and imprecise methods for ... V-12 = " * * * - . . R&S 113597 # 0# quantitative risk assessments can play a rational part in a d gisin whether certain emissions of carcinogenic substances are reasonable. The likelihood of huge errors in the risk ' estimates means that EPA has adopted only an illnsicrr. of pre--. cisioa-.. While quantitative accuracy is desirable, false pre- * cision is arbitrary. EPA* s summary recitation of the limitations of these , methods has not in the past restrained decision-makers and others * from treating these numbers as though they were reliable indica tions of the true risk, and will not in the future. Since EPA cannot prevent these estimates from being abused, by itself as . well as by others, we believe quantitative risk assessment should play no role in determining the level of standards for v' * hazardous air pollutants. * Even if it were legal and rational to use quantitative risk assessments in determining whether residual risk is unrea sonable, the proposed rules fail to adequately insulate, the determination of the winiwwii BAT standard from this risk-benefit* balancing process. The proposal merely promises that quantitative risk assessments will not be used in determining BAT. But since the estimates will have been generated at an earlier stage, NEDC and EDF have no confidence that EPA will ignore the estimates in determining BAT. We fear that the careful division that EPA has proposed will in fact disappear, and that even the definition of the minimum best technology will be compromised by the use of unreliable quantitative estimates. V-13 R&S 113598 G. Quantitative Risk Assessments Methods Possibly may be Used Rationally for a Very Rough Form of Priority Setting, But not as SPA has Proposed .............. .. Considering the unreliability of the' quantitative risk assessment .methods, their use for making any. sort'of close dis tinctions among chemicals is not rational* Nonetheless, EPA proposes to use quantitative risk assessment as the major element in setting priorities in a manner that makes closer distinctions , 4t than the techniques can reliably allow. , Xt has sonetimes been suggested that even though quantitative risk assessments are not accurate enough to be used in any way for determining the level of standards, they could be used to help in priority setting. The fundamental assumption here is that even if the exact magnitude of the risk associated with the chemical cannot be known, .the relative seriousness of two hazards can be compared. That is, using a. single mathematical model and a constant set of assumptions about interpreting the data one could rank chemical hazards in order of their seriousness . There are important reasons why this is nowhere, near as easy as it sounds. First, many of the elements that 1 ad to uncertainties in the risk assessment operate as "wild- cards," that is, they do not necessarily tend in the same direction as with the same force with different chemicals. All other things, being equal, such rankings might be made. But if one chemical V-14 participates in synergistic reactions and another does not, 1, bets are off. Similarly, two chemicals may,appear to be comparable in risk, but if they are metabolized differently, the true risks may well be vastly different. So even the rationality of using quantitative risk' assessments as a means of ranking chemicals in order of seriousness is substantially doubtful. ` It might not be arbitrary, however, to use these methods for priority setting as long as only -the very grossest distinctions are made, between the most serious hazards and ones which are less pressing (though by no means should be ignored) . This might be a justifiable approach as long as the distinctions that are made are at least several orders of magnitude in the apparent risk. Closer distinctions are not reliable because of the Pwild card** elements described above- l. t Ultimately we do not believe that a priority setting system based on even these sorts of gross distinctions will be much more useful to the Agency-than one based solely an such factors as the number of pounds of a substance emitted per'year, * or the number of persons living in the surrounding area and potentially exposed. A very serious problem with the use of quantitative * risk assessments even for priority setting purposes is that the unreliable numbers generated are then used, inadvertently or not, in an arbitrary manner in the .subsequent- proceedings' to R&S 113599 V--15 determine .the .level of a standard. Thus, NRDC and EDF believe that the use of quantitative risk assessment for even gross priority setting will be valid only if the Agency makes much more clear the. range of error that precludes the use- of these estimates for standard setting, and commits itself not to use 'these results , of the priority setting exercises in such subsequent standard setting" . It is very important that the es timates created for \ priority s'etting purposes be given in the 'form of the broad ranges within which the risks may lie. . To put forward a single' point estimate, or even a single point estimate and an upper bound, is unacceptable (particularly if the term "best estimate" is used) , because of its misleadingly precise character- / w O) o R&S R&S 113601 * VI-1 VI. THE NEED FOR IMPROVED GENERIC INTERIM CONTROLS FOR ORGANIC CARCINOGENIC CHEMICALS m 1. Introduction In a separate rulemaking notice entitled "National Emission Standards for Hazardous Air Pollutants: Advance Notice of Proposed . Generic Standards," EPA has proposed a series of "housekeeping requirements" as a first step in controlling carcinogenic emissions from rganic chemical facilities. As pointed out in the preamble, these generic standards are focused primarily on fugitive emissions (as opposed to process emissions) and would require little or no *V capital investment on the part of the affected industries. While EDF and NRDC strongly endorse the concept of requirements which can be applied quickly after the listing of'*a 112 pollutant, we . feel that the proposed standards are deficient in several respects. These deficiencies fall into two areas: first, procedural deficiencies in the way the Agency intends to implement the proposed standards, and second, substantive deficiencies which involve the requirements themselves. Each of these will be addressed separately. 2. Procedural Improvements According to the preamble to the proposed generic standards, EPA intends to separately propose the standards each time a new chemical is added to the list under 5112. .This approach is likely to prove both time consuming and possibly ineffective. First, EPA should substantially reduce its workload by proposing in one proceeding the generic standards for all those organic chemicals which can be regulated R&S 113602 VI-2 \J^\ J>y the proposed standards and which axe likely to ultimately be listed under S11'2. Under this approach, the Agency would rephrase ~ the regulation to apply to any organic chemical which met certain criteria, one of which would be listing under $112. This type of "fill in the blank" regulation would thus allow SPA to resolve all the pertinent legal and factual issues in one rulemaking and make r proposal of the standard for each pollutant unnecessary. A second alternative which would accomplish the same end, and one which EDF and NRDC intend to pursue at a later point in this rulemaking, is to request the Agency to propose the generic standards for a specific set of pollutants which we believe should now be - listed under S112. EDF and NRDC intend to submit such a list of candidates for listing either during the public hearing or shortly .thereafter. At that time, we will also submit a subset of these chemicals which we believe are suitable for quick regulation under the proposed generic standards.* * EDF and NRDC feel that this is the minimum which EPA should agree to since many of the generic controls proposed are very modest and will provide considerably less control than even PACT for existing . sources of hydrocarbons which are regulated because the hydrocarbons are precursors in the formation of photochemical oxidant smog. EDF and NRDC also suspect, based on discussions with various industry representatives, that many of these measures have already been implemented and that any objections can be adequately * dealt with in a single generic rulemaking proceeding. Such a single rulemaking proceeding will also reduce the possibility of con flicting court decisions should different companies plan to conted^ the requirements for specific chemicals. " 3. Improvements in the Substantive Requirements * There are a number of avenues for improving the proposed generic standards, some of which involve a fundamental change in the. nature of the Agency fs thinking and others of which involve relatively modest but significant improvements. In this section we will briefly outline both types of improvements. First, EDF and NRDC question EPA's decision to limit the generic standards to measures which require no significant capital investment. We believe that once a substance has satisfied the qualitative criteria for listing as a 112 pollutant that the Agency should move as expeditiously as possible to require compliance with whatever generic standards can be developed, whether those standards apply to process or fugitive emissions. Regrettably, the standards proposed by ZFA are limited to housekeeping measures and do not require even min imam control of emissions from process sources and storage sources using widely accepted technology. R&S 113603 For example, the EPA synthetic organic chemical manufacturing industry (SOCMI) study has already concluded that controls for fugitive and storage emissions are available which can be applied to most sources of carcinogenic air pollution. Given the fact that these requirements are unlikely to change as a result of further investigation and the fact that they generally make economic sense as well, we see no reason for deferring these requirements until the proposal and promulgation of process source oriented regulations. For example, EPA could include a requirement for floating roofs in storage tanks rather than simply requiring that existing tanks be painted ' VI-4 ,* * " white. Floating roofs are already being widely adopted by industry to reduce emissions, both for air pollution and economic reason's. Other examples of widely accepted generic control devices are double mechanical seals for pumps, and the venting of certain leak sources t control systems. Once again, most of these and other measures will be included in the control technology guideline for the organic chemical industry when it is finally released. It makes little sense to delay controls for carcinogenic hydrocarbons .until a second rulemaking when the Agency is preparing to require these same types of controls for hydrocarbons which are precursors to ozone. Prompt implementation of at least these "baseline" generic process and fugitive emission controls is especially important since it may be many years before EPA promulgates ultimate control - requirements for even a small number of the chemicals which are likely to be listed under 112. As a result, these first line controls become even more important in providing at least a minimal d gree of health protection to Americans *who continue to be exposed to highly dangerous air pollutants. In general, EDF and NRDC submit that EPA should include in these generic controls all those measures which are, or will be, required for existing sources of hydrocarbons under the ozone component of state implementation plans, unless there is some reason t believe that the technology which would be used to control these emissions once a substance is listed under 112 will be substantially more effec-- * tive than the technology the Agency is planning to use for ozone attainment. In other words, any control measure, such as the use of floating roofs, which is both generic to the chemical industry and cannot be significantly improved upon as a result of further rulemaking R&S 113604 VI-5 should be required as soon as possible. R&S 113605 In short, the procedure would be as follows. First, -. E?A would identify all generic process and fugitive emission con- tr Is which it is currently considering for regulating organic - chemicals. These controls would then be broken into two groups: (1) those controls which are likely to be required in any events and which raise few marginal cost issues as control requirements are increased; and '(2) those measures where the degree of control or performance expected of a particular device will be an issue. Thus, the issue of whether to require one or two activated carbon beds in a particular control device as well as the decision as to the replacement frequency of the carbon itself would be placed in this second category. Similarly, if EPA felt that it was possible that a pressurized tank as opposed to the use of a floating roof might be required for certain chemicals, a decision would be deferred until the second round of rulemaking. The object is to avoid requiring investment in control devices which later would have to be replaced with even more sophisticated devices as a result of the second rulemaking. We sub"" t that this criterion of "no duplicative investment" is a far superior one to the "no capital investment** criterion pro posed by the Agency. Furthermore, the "hurdle** for listing of a * chemical under S112 is sufficiently high that industry is not in a position to argue that immediate imposition of these generic .control requirements is unjustified until further rulemaking takes place. Even the generic controls cannot be applied until the first rulemaking is conducted and an opportunity for public hearing provided. With regard to the proposed housekeeping or fugitive emission VI-6 * * R&S 113606 controls themselves, EDF and NRDC have a number of suggestions for improving the proposed system. First, we are concerned that EPA __ is placing undue reliance on industry reporting requirements to ensure compliance with the proposed controls. In its preamble SPA has acknowledged the potential importance of independent enforce ment either through SPA field inspectors or through other mechanisms but has failed to develop an adequate alternative enforcpm*yyt- proposal. EDF and NRDC appreciate the fact that the Agency's enforce ment resources for this program are limited and believe that any proposal must make optimal use of those resources. To accomplish this objective, two possible approaches suggest themselves. First, EPA could require affected industries to retain an "environmental auditor" who would not be employed by the company, but would be responsible for reviewing the company * s compllanc with the generic standards and reporting the results of his audit .to EPA. The concept is similar to the use of an outside accountant to perform an annual financial audit on corporations for purposes of SEC requirements. The outside environmental auditor would be responsible for not only reviewing the records required to be k pt by the company but would also be responsible for performing periodic field spotchecks himself to determine adequate compliance with the reporting requirements. Since an environmental auditing profession does not presently exist, it is possible that engineering consulting firms could be retained to perform this task initially . Some of the large "big eight" accounting firms have management consulting staffs which could also be retained to do the job. This approach would significantly reduce the burden on EPA'sown enforc^^ ment resourc s, while providing a relatively high .quality cheek on company compliance with the generic requirements. .- A second possible approach would be to use EPA enforcement personnel/ but use them selectively to ensure that certain performance standards were satisfied by industry. In the preamble/ EPA has pointed out that CMA has argued for a more flexible system which would reduce reporting requirements on industry and rely more heavily on performance standards to reduce fugitive emissions. EDF and NRDC believe that EPA should give consideration to a modified version of the CMA proposal. Under this modified approach, industry w uld be given somewhat more flexibility in developing control plans but would be held to a higher standard of performance. For example, a series of performance standards relating to the frequency and mag nitude of leaks which would be allowed could be developed. During I a random inspection of a selected number of potential sources, the facility would be limited to a specified number of permissible leaks within certain concentration ranges. For example, leaks in excess of 15,000 parts per million (ppm) might be limited to one percent of the fugitive sources tested. Those between 10,000 ppm and 15,000 ppm might be limited to three percent of the sources tested, and those less than 10,000 ppm might be limited to five percent of the sources tested. A minimum sample might be 100 leak sources, a sample which could be completed within a reasonable period of time. R&S 113607 ; After sufficient experience had been gained with the program, the number of permissible leaks could gradually be reduced in each category in order to improve the line of overall performance. To encourage companies to meet the performance standards, .heavy fin s ' VI-8 should be levied on any leaks which are found in excess of the per- * mitted numbers. In addition,, the company could be required to resume weekly or even more frequent reporting of leaks from all possible sources at the facility. In short, by providing strong incentives to fulfill its own management plans, the Agency could not only reduce the resource burden on itself, but also reduce reporting requirements for industry. Naturally, the feasibility of this program would depend on EPA having available at' least enough enforcement personnel to perform these random sample in spections. If:* such personnel are not available, it may be necessary to resort to the extensive routine reporting required in the proposed regulations. ' The key point, in any event, is that both industry and govern ment resources should be focused on reducing emissions from fugitive sources as quickly as possible. Simply reporting and requiring the repair of leaks will not necessarily ensure improved performance over time. Only increasingly restrictive performance standards can achieve-this.goal. We submit that industry should be given the flexibility it desires if it can demonstrate that it can consistently meet stringent performance standards. If it fails to meet those performance standards, the Agency will then have full justification for resorting to a more burdensome set of reporting and-monitoring requirements. To establish the initial performance standards, we submit that EPA should immediately require the targeted industries to submit data pursuant to 114 of the Clean Air Act to allow the Agency to establish leakage frequency rates for different sources R&S 113608 r &S 113609 a. VI-9 within the industry. Since the development of this baseline data will take considerable time, we suggest that the Agency proceedimmediately to require industries to begin developing this data. Once the data is developed, it will allow the Agency to focus on those sources which appear to be most significant in terms of total emissions. Such data will also give Agency inspectors know ledge of which sources should be spotchecked during inspections. In short, this second alternative might actually encourage industry to perform more frequent checks of potentially leaking valves and other sources than would actually be required by a formal set of reporting requirements. Since industry would not know when an EPA inspector might arrive, it would be required to keep in; a constant state of readiness. Regular reporting require ments might actually produce the opposite incentive. Industry would know that it was only required to have the system in good operating condition at the time of its regular reporting. A system with less certainty as to when checks would be made may actually produce improved compliance. In any event, we submit that EPA ought to consider a hybrid which might involve some routine reporting, with the addition of spotchecks geared to specific performance standards as outlined above. We believe that all of these ideas need further discussion and would welcome a careful review of this area prior * to promulgation of final generic standards. VII-1 < * *i R&S 113610 VII. CONCLOSION5 As noted in the preceding sections, during the remaining .period for supplementary comment NRDC and EOF will submit addi tional material in support of and in further explanation of the analysis and recommendations contained in these comments. We -have summarized here the major elements we believe are necessary to generic rules to implement the hazardous air pollutants pro visions of the Clean Air Act with respect to airborne carcinogens. v` . First, EPA must establish a candidate list of carcino gens which may be emitted from stationary sources as air pollu tants. At a minimum, this list must contain several hundred substances which' are now known to cause cancer in humans or in animals .and which are used in industrial facilities and may be emitted as air pollutants. Second, from the candidate list EPA must select.a min n T-rnm of 20 substances per year to list as hazardous air pollutants and regulate as Section 112 requires. Considering that EPA has regulated only four substances to date, and that there are several hundred potential hazardous air pollutants, the Agency has grossly failed to carry out its mandate under the Clean Air Act. Nothing less than a commitment to regulating 20 substances per year will satisfy that mandate, because nothing less will guarantee that within a reasonable time the public will be protected from cancer caused by airborne chemical and radio active substances released from stationary sources. VII-2 R&S 113611 Third, the Agency must resolve in this generic pro ceeding certain essential scientific issues which arise in the same manner and must be resolved on the same evidence in each proceeding involving an individual carcinogen. These scientific ** issues would be foreclosed from further consideration in pro- ** ceedings to regulate individual substances. Only when signifi cant new information is presented which calls for a different resolution of the generic scientific issues, will such issues be reconsidered, and only in the context of a general proceeding to amend the general rules. This resolving and foreclosing of the generic scientific issues is essential if a rational and 'efficient consistent airborne carcinogens policy is to be imple- mented, and a significant number of substances brought under control. a* * Fourth, the Agency must establish in this general proceeding that when a carcinogen is listed as a hazardous air pollutant, within 3GO days a standard of no measurable emissions will automatically go into effect, unless Industry shows that it cannot be achieved through process changes and product substi tution and unless Z?A has established by that time an alternative ** * standard'which deviates from no measurable emissions to the mini mum extent required. The substitute standard must provide for a phased reduction to zero emissions at a reasonable point in the future. The presumption necessarily is harder to overcome for new plants than old, since new sources of carcinogenic emissions VIX-3 . have ccusiderably greater emissions control flexibility and, of course, need not be built at all. Dispersal, -of sources of I carcinogenic air pollutants is not an acceptable alternative to controlling their emissions. ,* *. . ,, . ` Fifth/ EPA must abandon its arbitrary and unlawful - - m - *. * reliance on quantitative risk assessment- methods. Quantitative risk estimates can be wrong by a factor of 5 million times or acre and thus are too unreliable and imprecise to play a rational role in'-determining the level of controls applied to a hazardous pollutant. Such estimates may have a role in the grossest fora of priority setting/ but that will be valid only if the Agency much nore explicitly recognizes the uncertainties of the esti mates and comits itself to not using them in any way in subse quent standard setting. * ' 9* Sixth, EPA should expand the scope of the -proposed "first step* generic controls that will be applied quickly' to . include those controls on process emission points which are not *' m expensive, or likely to be inconsistent or duplicative of more effective technological controls to be considered thereafter. Hcreover, EPA should consider restructuring these interim rules . in ways that stay enhance voluntary compliance and permit more effective enforcement. R&S CO O)