Document EvYV0gmvkpzYL721X0y497KXx

i K <I ABD00078219 Natural Resources Defense Council, Inc A Scu York Office 122 EAST 4 2 N D STREET NEW YORK. N.Y. 10l68 212949-OO49 135O NEW YORK AVENUE. N.W. SUITE 3OO WASHINGTON, D.C. 2OOO5 202 783-7800 Western Office 25 KEARNY STREET SAN FRANCISCO. CALIF. 94I03 415421-6561 In Re: National Emission Standard for Vinyl Chloride Petition for Reconsideration of the Withdrawal of the June 1977 Proposed Amendments and- Ccmmants on the January 1935 Proposed Amendments 50 Fed. Reg. 1132 (Jan. 9, 1985) Prepared oh behalf of Natural Resources Defense Council and Environmental Defense Fund by David D. Dcniger Senior Staff Attorney Natural Resources Defense Council March 25, 1985 Rfcvced Pioer Sew England Odice: 850 boston post road slobury. ma. 01776 017 237-0472 Public Lzr.ds lr.s;iiu:c: 1720 race street bznver. co. S0200 303 377--9740 kn ABD00078220 A On January 9, 1985, the Environmental Protection Agency (EPA) published a notice of proposed revisions to the 1976 national emission standard for the hazardous air pollutant vinyl chloride. Included in the notice was the withdrawal of amendments to strengthen the standard, proposed in 1977 pursuant to the settlement agreement in Environmental Defense Fund v. Train.1 The Natural Resources Defense Council (NRDC) and the Environmental Defense Fund (EDF) strenuously object to the withdrawal and hereby petition for reconsideration of it.2 The notice also announces proposed revisions principally concerning requirements for relief valve discharges, as well as certain other matters. NRDC and EDF submit that these revisions substantially weaken the current standard, and we hereby file comments in opposition to them. I. The Withdrawal of the Proposed 1977 strengthening Amendments: Petition for Reconsideration A. 3--ac-kcround In 1976, EPA promulgated a standard for vinyl chloride (VC) which v/as significantly compromised by consideration of factors that EPA is not authorized to consider under Section 112 of the Clean Air Act. EPA rejected certain requirements which would have further curbed VC emissions either on the basis that the agency - No. 76-2045 (D.C. Cir., filed Nov. 19, 1976). 2 Because the withdrawal appears to be a final agency act icr., NRDC has filed a petition for review in the Court of Appeals for the District of Columbia Circuit. NRDC v. EPA, No. 35-1150 (D.C. Cir., filed Mar. 3, 1985). ABD00078221 c2 A deemed them not to be adequately demonstrated or deemed their cost to be too high. Certain control requirements were rejected even though they were already i * use on some sources, or even though they were nothing more than reasonably forecasted improvements that the sources were fully capable of making. These technological and economic tests are inconsistent with the "ample margin of safety" requirement of Section 112(b)(1)(B). Because E?A had not complied with the requirements of Section 112, EDF filed a petition for review in the U.S., Court of Appeals for the D.C. Circuit.3 After significant discussions between the parties, the case was settled with an agreement under which EPA obligated itself to propose amendments to strengthen standard in specified ways. At the heart of' the agreement was EPA's recognition of a "zero emission goal" to be approached as closely as possible, and EPA's agreement to propose an immediate strengthening revision and to undertake a review of the standard in three years leading to a further strengthening revision. As part of the compromise settlement, EDF agreed to dismissal of this specific case. EDF, however, conceded none of irs legal objections to the construction of Section 112 advanced by EPA in the 1976 standard. B. The 1977 Proposed Amendments The improvements to the standard contemplated by the settlement agreement were published or. June 2, 1977.4 EPA J See ncce 1. Fed. Rec. 2S134. % i- ABD00078222 -3- A specifically recognized a "zero emissions goal. " E?A restated its conclusion (with which we agree, of course) that VC, as a carcinogen, has no known threshold of effect and endangers public health at any level of exposure. The agency then stated: In order to assure that the standard continues to approach the only level of emissions which is known to be absolutely protective of health, namely zero emissions, EPA is proposing amendments which require more efficient use of existing control technology at existing plants, and which encourage technology to reach this goal without banning vinyl chloride.5 EPA proposed the following specific requirements: (1) To lower the emission limit for emissions from equipment used in VC and polyvinyl chloride (PVC) production from the level of 10 ppm promulgated in 1976 to 5 ppm. The proposed 5 ppm limit would take effect immediately for new sources and within three years for existing sources. (2) To eliminate the special 0.2 g kg limit for oxychlorination reactors as it applies to new reactors and require new reactors to meet the 5 ppm limit. This could be accomplished, EPA found, by using oxygen, rather than air, as the feed. (3) To strengthen PVC stripping requirements for new PVC resins by limiting residual VC after stripping to one fourth the levels allowed by the 1976 standard. (4) To require a VC emissions offset when a new source of VC emissions is located within 5 miles of an existing VC source. (5) To review and strengthen the standard within three years of Dromulcation of these amendments. 5 Id. 231 ABD00078223 -4- A As we discuss below, EPA's own studies and other evidence clearly demonstrate that these requirements are achievable.6 C. The Withdrawal EPA, however, has now rejected the proposed amendments for a combination of reasons that are not legally cognizable and that ignore the evidence in the record. The January notice states that the June 1977 proposal is being withdrawn because the proposed amendments are not new considered "appropriate." The notice then reviews each of the requirements proposed in 1977 and rejects them on the basis of a combination of two types of arguments: (1) that the technology to accomplish these requirements allegedly has not been demonstrated, and (2) that the requirements are allegedly not justified on a cost-benefit basis.7 We first address the legal problems with this notice, and then the evidentiary problems. 1. Legal Violations (a) Failure to Propose the Withdrawal. The withdrawal was announced as a final action, without being preceded by a proposal. This course of action contrasts sharply with the course of action followed by EPA in the case of benzene. when EPA decided to withdraw its benzene proposals for three source categories last year, the agency issued a proposal and offered the public an 6 In the years following the proposal, EOF repeatedly urged EPA to promulgate the amendments. See letter from Robert Rauch, EPF Staff Attorney, to Douglas Costle, EPA Administrator, :2b. 1973; letter from Rauch to David Hawkins, EPA Assistant Administrator, Aug. 15, 1973; letters from Larry Corcoran, EPF Staff Attorney, to Costle, Apr. 13 St June 13, 1930. 7 50 Fed. R.ec. 1133-36. */ ABD00078224 -5- opportunity to .comment. NRDC and EDF object to the agency's refusal to do so in this case. (b) The Role of Cost Considerations. EPA opens its discussion by stating its current view of the basis for the current standard, as well as the basis which guides EPA's current actions. EPA states: The current VC standard was .established based on judgments concerning the costs and benefits of the standard to society. The standard is not designed to eliminate VC exposure risk entirely. Rather, it strikes a balance between public health protection and the cost. of that orotection.8 This balancing of costs and benefits, however, is contrary to the requirements cf Section 112 of the Clean Air Act. There are at least three decisions of the D.C. Circuit whichreject assertions that cost considerations are relevant or permissible under a "margin of safety" statute such as Section 112. The D.C. Circuit has twice held that under Section 109(b)(1) ofthe Act (which requires EPA to set NAAQSs at the level which protects public health with "an adequate margin of safety"), standards may not be compromised by cost cons iterations.9 In addition, interpreting the "ample margin of safety" requirement in Section 307(a) of the Clean Water Act, the D.C. Circuit rejected the claim 8 Ic. at 1183. 9 Lead Industries Ass'n v. : 647 F.2d 1130, 1145-51, 1153 D.C. Cir. 19S0), cert, denied U.S. 1042 (1930); American Petroleum Ir.st. v. Ccstie, co5 J.2c 1176, 1185-36 (D.C. Car. 1981) , cert, denied 102 S.Ct. 1737 (1932). ABD00078225 .-6 - that EPA may consider costs when establishing an ample margin of safety standard for a toxic water pollutant.10 Likewise, the Supreme Court has at least three times rejected efforts to imply the relevance of cost-benefit considerations under statutes which direct agencies to base decisions on public health or environmental protection factors.il Acknowledging the legal limitations expressed in the terms of the law and their legislative history, EPA has from time to time nonetheless tried to defend the use of cost-effectiveness or costbenefit analysis under Section 112 with the argument that Congress did not mean what it plainly said. According to this argument, the Congressional policy choice was, in EPA's view, so unwise that it should be ignored. In its place, EPA substitutes the agency's own policy choice. This sort of administrative attempt to rewrite a law with which the agency does not agree has been uniformly rejected by the courts. In all six of the cases cited above, the Court of Appeals and the Supreme Court have emphasized that neither an agency' r.cr a court may second-guess Congressional limitations on the role of cost considerations under health and safety statutes. EPA cannot violate this restriction. 10 Hercules, Inc, v. EPA, 593 F.2d 111-12 (D.C. Cir. 1973). 11 Union Electric Co. v. EPA, 427 U.S. 246 (1976) (no requirement will be implied for EPA roconsider costs when reviewir.c a state implementation plan requirement adopted to meec a primary standard); TVA v. Hill, 437 U.S. 153 (1973) (no "reasonableness test to be implied into the Endangered Specie Act); American Textile Mfrs Inst, v. Donovan, 452 U.S. 490 (1 (no cost-benefit test under the Occupational Safety and Health Act) . a) in ABD00078226 A (c) Technology Demonstration Requirements. In rejecting the proposed lowering of the 10 ppm emission limit to 5 ppm, EPA states simply that based on experience to date, "10 ppmv represents the lowest level of control which has been consistentlv achieved."12 Before addressing this statement empirically, we must take issue with its legal premises. EPA apparently presumes that Section 112 of the Act establishes a requirement for demonstrating that a level of control has been "consistently achieved" in the past. While a demonstration requirement under Section 111,' there is no such requirement under Section 112. None of the key-words of Section 111 which establish such a test for new source performance standardsl3 are present in Section 112. The Supreme Court made clear in Union Electric that the omission from sections of the 1970 Clean Air Act of words relating to technological and economic showings reflects a deliberate Congressional choice to which EPA and the courts must give effect.14 Sections of the law where such requirements are lacking are "technology-forcing. " In the Court's words, allowing considerations of technological infeasibility claims: 12 50 Fed. Reg. at 1184 (emphasis added). 13 An NSPS shall reflect "the degree of emission reduction achievable through the application of the best system of continuous emission reduction which (taking into consideration theccst of achieving such emission reduction and any nonair cualitv health and environmental impacts and energy requirements) the Administrator determines has been adequately demcnstrataec for that category of sources." Section 111(a) (1 ) (C } . 14 427 U.S. at . [3 IRC at 2145-47.] ABD00(g8227 would frustrate congressional intent. It would permit a proposed plan to be struck down as infeasible before it is given a chance to work, even though Congress clearly contemplated that some plans would be infeasible when proposed. . . . Technologyforcing is a concept somewhat new to our national experience and it necessarily entails some risks. But Congress considered those risks in passing the 1970 amendments and decided the dangers posed by uncontrolled air pollution made them worth taking.15 The Court's reasoning applies equally to a hazardous air pollutant standard issued under Section 112. EPA's apparent test -- that a control requirement must have been achieved in the past before it may be required for the future -- is also inconsistent with tyhe -agency's clear technology forcing responsibility under sections such as Section 202. Section 202(a)(1), for example, instructs E?A to set a standard for a future application date which accomplishes such improvements in controls as are reasonably forecasted by then. EPA has the authority, indeed the responsibility, to project such progress and base the future standard on it. As the D.C. Circuit stated in NRDC v. EPA, the diesels case; "The Clean Air Act recu ires the EPA to look to the future in setting standards ...."16 And as stated as far back as the 1970 Senate Report: [EPA] is exoected to press for the development and application of improved technology, rather than be limited by that which exists. In other words, standards should be a function of 15 .Q-, ERC at 2151] . 16 NRDC v. EPA, 655 E 313, 323 .C. Cir. 1921), cert, denies sub nom General Motors v. C-crsuch, 16 ER.C 1616 (1981  # ABD00078228 -9- A the degree of control required, not the degree of technology available today.17 The 1977 proposal, which contemplated setting a 5 ppm emission limit with a three year leadtime for compliance by existing sources, is precisely this sort of standard. If technology-forc.i ng authority clearly exists under sections of the Act which are not driven by health considerations and which explicitly provide for considering costs, then certainly such authority exists under Section 112, which is driven solely by the mandate to protect public health with an ample margin of safety. EPA recognized this in 1977, but is denying it now. The withdrawal is based on an incorrect statutory construction and is therefore invalid. (2) Conflict with the Evidence When examined in accordance with Section 112' s requirements, the evidence in EPA's own support documents will not support EPA1s rejection of the 1977 proposed amendments. (a) The 5 ppm Standard. The bases given for .'withdrawing the proposed 5 ppm emission limit are: (1) that the 5 pern limit "was not based on data for control technology different from that analysed" in 1976; (2) that it was opposed by industrial commenters on the grounds that it would require maintaining an average level lower than 5 ppm, and that it would result i they (ana apparently EPA) view- as little* 1 overall e miss ion reduction; and 3) that the commenters objected to the " zero emission goal. "IS 17 S . R ep . NO . 1196, 91st Cong., 2d Sess. 2 4 {1970 13 50 Fed. at 1134. ABDQ0Ql7229 A EPA then stated that the 5 ppm limit was withdrawn because it had not been "consistently achieved" on existing facilities. As stated above, this is not the legally correct test. And when the record is examined, it fully supports the achievability of the 5 ppm limit by new sources upon commencement of operations, and by existing sources within three years of promulgation. In contrast to the agency's current position, the 1977 proposal stated that the 5 ppm limit would not require the installation of new equipment. Rather, its purpose was "to force owners and operators to maximize the effectiveness of existing control 'systems. "19 The evidence amassed by EPA since then show's that the 5 ppm level is in fact achievable both by better use of existing equipment and by new and upgraded equipment. The document "Vinyl Chloride - A Review of National Emission Standards "20 states: "Primary control devices [e.g., incineration, solvent absorption, refrigeration] are reducing emissions, in most cases, well below the 10 ppm level for exhaust gases. "21 Specifically: o A series of thermal incineration tests at one plant in Kentucky demonstrated VC levels consistently at 0.26 pom VC or below, less than three percent of the 10 ppm limit, and less than six percent of the proposed 5 ppm limit.22 19 42 Fed. Ref. at 28155. 20 EPA-450/3-32-Q02. 21 Id. at 4-1 (emphasis added). 22 Id . at 4-3.  /' ABD00078230 . -11 - o Flares are now rated at at least 90 percent effectiveness in destroying VC, which allows a 5 opm limit to be met for any streams under 50 ppm VC. 23 o Two PVC plants have demonstrated that a double bed carbon adsorptionsvstem can achieve compliance with a 5 ppm limit. While the Review states that they are effective in meeting the 10 ppm limit, it also states: "When a probe at the outlet of the bed indicates an approaching 5 ppm level (as an indicator of breakthrough), the waste stream is diverted to the other bed while the first bed is being regenerated."24 This indicates the achievability of the 5 ppm- level. o Solvent absorption is available for both PVC and EDC/VC plants. The Review describes a B.F. Goodrich system, improved since the original 1976- support document was written. A 99.99 percent recovery efficiency is reported for this system. This should be sufficient to meet a 5 ppm limit on most, if not all, streams. The solvent is described as "proprietary, commercially available,'inexpensive, and reputed to be low in toxicity."25 o Tenneco has developed a process of reacting VC with ozone in the presence of activated carbon which reducers VC to less than approximately 1 ppm from streams containing 10-10,000 ppm VC.36 The availability of these measures belies any claim that the 5 ppm limit cannot be achieved by new sources, or by existing ones given a three year lead time.27 As for industry's argument that the 5 pom limit would result in "little" overall reduction in VC, these measures coul reduce 23 Id . at 4-12 . 24 Id. at 4-13 -- 4-14. 25 Id. at 4-14 -- 4-15 . 26 Tr # a ~ 4-16--- 4-17. 27 It be a rs emphasizing that the 1977 proposal contemplated a proc sdure thrcugh which an existing scurc'e could s n ow Z ? A t r. a a 5 DD m limit was still not achievable and obtain an extension or hicher limit. ABD00078231 - 12 - A the estimated 680 kg/vr from a 68 Gg/yr PVC plant by half.23 Given the carcinogenic nature of VC and the requirement of Section 112 for protection of public health with an ample margin of safety, this reduction cannot legally be foregone. As for the industry commenters objections to the "zero emission goal" toward which the 1977 proposed amendments would go, the ample margin of safety provision requires no less. (b) Oxychlorination Vent Standard. The proposed 1977 amendments would have required the oxychlorination vent to meet the 5 ppm level on new sources. The 1977 proposal states that this can be accomplished by the use of oxygen as a feed instead cf air. The 1977 proposal also cites studies showing that this has been found to be economical.29 The 1985 withdrawal notice cites no discussion of the use of oxygen as a feed. We can find no such discussion in the VC Review either. CPA has failed to justify the withdrawal of this proposal. (c) Striooinc. The 1977 proposal would have cut the residual VC limit for new dispersion resins from 2000 ppm tc 500 ppm, and the limit for ether new resins from 4-00 ppm to 100 ppm. The oroDOsal stated that CPA believed most new resins could meet these levels, but if one could'not, a manufacturer had the option of choosing not to begin making it. CPA new states tnat oecause of difficulty administering the proposed distinction between new and existinc resins, the agency chose to examine whether all 2 8 VC Review at 4-6 , 29 42 Pec. Reg. bie 4-4. 2 S1 5 7 . ABD00078232 - 13 - ' A resins could be stripped to these lower limits. (The January 1985 notice reflects no effort to change or fine-tune the "new resin" definition so that it would work.) The notice states: The review study found that resin stripping has improved since the current standard was promulgated, and that, some processors can achieve lower resih residual VC levels than those required in the original standard. In certain cases, some resins can meet the more stringent levels specified in the previously proposed amendments. But because it has not been shewn that all resins can at this moment be stripped satisfactorily to the lower levels, no levels lower than those specified by the current standard have been "demonstrated, " and no lower level will be required for any resins, new or old.30 This is, first of all, an application of the wrong legal test. There is no such demonstration requirement. The Review study in fact shows that an impressive percentage of ?VC production can already meet the lower limits, and that much of the remaining production is close.31 if EPA set this target cut for new facilities, or for existing ones with a three year lead time, it clearly could be achieved simply by copying the equipment anc procedures used by the leading facilities new. D. Conclu sion For the reasons given above, NRDC and EDF submit that the withdrawal is not supported by the lav/. Nor is it supported by the evidence, when it is evaluated according to the requirements 30 50 Fed. Reg. at 1135. 31 See Review at 4-55 --4-56, Table 4-11 ABD0OOI7S233 * A and limitations of the law. We therefore request that EPA reconsider the withdrawal and publish the proposed amendments. This should be done, we submit, no later than at the time final action is taken on the new amendments proposed in January, to which we now turn. II. The 1985 Proposed Amendments: Comments The principal feature of the proposed "administrative revisions" to the standard is the change in the relief valve discharge requirements. Because the effect of this change is to substantially weaken the current standard, NRDC and ED? submit comments in opposition below. We also comment on two other changes: the amendments proposed for the reporting requirements, and the le^c . detection and elimination requirements. A. Relief Valve Discharges 1. The Proposal to Grant an Allowable Discharge Rate The' current standard bars all relief valve discharges except in "emergency" situations. An emergency does not exist if the discharge is "preventable." Currently, EPA enforcement personnel examine relief valve discharge reports made by companies to determine if they were preventable through the use of available eauipment or procedures. If they are found to be preventable, thev are not considered emergencies and constitute violations of the standard. In recent years, enforcement of the relief valve discharce standard bv ERA and the Department of Justice has picked up. ABD0OO*78234 A While it is not all it could be, it is a substantial effort. It has been a very successful effort so far as well. The proposal states, however, that primarily because of a "significant" use of agency resources in analyzing the preventability of discharges, EPA has chosen to revise the standard to permit an allowable number of relief valve discharges per facility per year. PVC reactor discharges would be allowed 0..35 discharges per 100 polymerization batches (with an upper limit of four allowed discharges per year for the suspension resin process). Nonreactor sourceswould be allowed 0.25 discharges/100 batches,, with an upper limit of three per year. Continuous process units would be allowed 1 discharge per year. EDC/VC plants would be allowed four discharges.per year. NRDC andEDF strongly oppose granting the industry an allowed, "free" rate or number of discharges. As a matter of principal, all preventable discharges must be avoided. They are the single largest remaining type of VC emissions, accounting for thousands of pounds per event, and tens of thousands of pounds par year at some sources. A single discharge can release tens of thousands of pounds of VC. See, e.g., the abysmal record of releases at the Formosa plant in Delaware. EPA enforcement policy states that releases are presumed preventable unless a source demonstrates otherwise. To so demonstrate, the source has to shew that neither, proper employee training, inspection and maintenance of equipment, proper design and coeration of control equipment, nor installation or ail controls needed to mesa the standard would have sufficed to ABB00098235 a A prevent the discharge. If a discharge has occurred before, that fact is to be taken into account in deciding if subsequent discharges could have been anticipated and preveneted. Proper instrumentation and controls are required before a release can be deemed unpreventable. Likewise, the injection of short-stop chemicals must have been tried 'where possible. In addition, enforcement policy treats a release as preventable if a firm could have used a gasholder system to contain releases_until they could be treated.32 As a matter of empirical fact, the proposed discharge allowances constitute a significant weakening of the standard. EPA preventability analyses, conducted pursuant to this long standing EPA enforcement policy, have routinely resulted in the conclusion that the vast majority of releases are preventable. The very support document on which EPA relies for the proposed changes to the relief valve requirements itself indicates that the new allowable release rates will exceed what would occur under the existing standard and enforcement policy. For the purposes of developing these rates, EPA's contractor, Radian, applied the following criteria: Based on information gathered from visits, contacts with regional offices, compliance reports, and vendors, it was that certain types of discharges can be plant 10-day judged 32 See, e.g., Memorandum from Director, Division of Stationary Source Enforcement, to Lawrence Goldman, Chief, Enforcement Branch, Region I (Apr. 24, 1979), re: "Vinyl Chloride Rearer Valve Discharges from ?VC Reactors-Bordan Chemical. " See also Memorandum from Director, Division of Starionary Source Enforcement, to Stuart Roth, Attorney, Enforcement Division, Recion I (Feb. 23, 1978). (These memoranda are attached as Aocendix A) ABD000j7^236 prevented by reasonable measures. For example, discharges caused by operator error, due to operator negligence or failureto follow standard operating procedure (SOP), were considered clearly "preventable." In addition, operator errors resulting,from insufficient training or lack of established SOP were judged clearly "preventable". . . . Another example of 'preventable" discharges is discharges due to premature releases from relief devices. . . . Finally, discharges recurring for the same reasons as previous discharge incidents were considered "preventable. "... Although other causes for specific discharge occurrences may be preventable, this assessment v/ould need to be done on a case-bvcase basis and would require, in many cases, more information than is available in the 10- day compliance reports. Thus a more refined preventabilitv assessment can not be made with the available data. For the purposes of this analysis, only the general discharge categories described above are jaentified as clearly "preventable. "33 In other words, were the current enforcement policy to continue to be followed under the current regulation, more discharges would be found to be preventable. Additional control measures besides those mentioned above v/ould be applied. Further indication that the proposed allowable discharge rates are ir.appropr iate comes fr om the statement s, in th e R I i ~ Valve report, that these rates v; uuld be complied w i th at present by the bulk of the industry.34 In other words, these 1 e ve 1 s 3 3 Memorandum from Reese Hcwle and Karen F idler, Radian Co to File, re: "Vinyl Chloride Standard - Numerical Limits for Relief Valve Discharges" (Acr. 16, 1934) a: A-6, attached as Appendix A to "Vinyl Chloride: Relief Valve Study," ZPA-45G/ 002 (emphasis added). 34 iC a rt-O /. ABDQ00[7217 A would legitimize the current discharge rates of the industry, regardless of the ability to lover them further. The VC Review Study, in fact, describes a variety of additional control measures beyond those mentioned in the Relief Valve report. There have been significant advances in the use of instrumentation and computer controls, and in the development of short-stop chemicals and in the development and use of systems to insert them into reactors which are overpressurizing. Gasholders, about which more will be said below, are in use in at least three of the five plants surveyed for the Relief Valve report, and are capable of containing many, if not all, discharges until they can be treated by primary control devices.35 Several companies' effective preventative systems are described in the VC Review Study.36 Conoco, for example, has developed a hilling agent capable of preventing discharges under worst-case conditions.^/ The only reason offered thus far for the chance is the fact that the present requirements entail use of allegedly "significant" agency resources in case-by-case preventability analyses.33 To our knowledge, however, EPA has presented no 35 See VC Study, 4-13 -- 4-40; Relief Valve Report, 4-2, Table 4-1. " * 36 VC Review Study, 4-40 -- 4-4S. 37 11* at 4-43 " 4-46. 33 EPA also states that the current standard leads to "uncsrtaintv" for the industry, as it does not knew exactly what discharges EPA will consider preventable. This appears to us to be a highly salutory uncertainty; the coal of the current standard is to come as close as possible to eliminating relief valve discharges and the uncertainty about risk of being in violation cf the standard rends to stimulate greater efforts to prevent  ABD9007S238 A analysis to suDoort this resource concern. Moreover, even granting that the case-by-case analyses consume some resources, we submit that EPA cannot legitimately resolve its resource problem by weakening the standard. Rather, if EPA lacks the resources to carry out its duties, its obligation is to tell Congress what it needs to do the job right. 2. The Failure to Require Gasholders NRDC and EDF submit that EPA should be strengthening the standard to specify use of all available equipment and procedures that can prevent discharges or, should they occur, prevent or minimize the escape of VC to the atmosphere. Foremost among the control measures not new required by the standard is a gasholder system. As mentioned above, EPA's support documents show that such a system is feasible and is fact in use on a number (at least three) of facilities. Just this month, the State of Delaware entered a consent agreement with Formosa Plastics under which that company has become obligated to install such a system at its plant.39 EPA might advance two reasons for not requiring the use of a gasholder system on each plant. First, EPA might deem the cost to be too high. As demonstrated above, however, cost considerations may not be considered under Section 112. Moreover, it is difficult for NRDCand EDF to accept such a contention from EPA's reculation-writing- side in light of the fact that the agency's cischarge s. 39 Delaware v. Formosa Plastics Coro., C.A. 34C-DE-73, Para. 1(3) (De. Sup. Ct, consent, order filed Mar. 13, 1935 ). ABD90(M239 A (and the states') enforcement side has proved capable of achieving such requirements in consent orders. The second reason E?A might advance is a contention that some facilities may be able to prevent discharges almost entirely without such systems. To this we respond that no facilities will be able to eliminate such discharges entirely; there will continue to be "emergencies" that are not judged "preventable." There is no reason to allow these emissions to reach the atmosphere, even if the discharge valve releases themselves are unpreventable. For these releases there remains a vital purpose served by gasholder systems . NRDC and ED? therefore request that E?A reevaluate its position on gasholder systems and related equipment, and promulgate regulations requiring their installation and use. B. Leah Detection and Elimination NRDC and EDF do not support the proposed adoption of the generic leak detection and repair program promulgated in June 1934 in place of the current leak detection and elimination programs for VC, as it appears that the effect of this change will be to weaken the standards. EPA states that the regulated firms will have the choice whether to stick with their current programs or to adopt the generic one; that option will predictably be exercised only by companies which find their current programs weaker than the generic one. Firms with stronger requirements will likely switch to the generic program. One place where the weakness of the generic program versus the current ones can be seem is the definition or a "leak. " The ABD000t24D generic program, and the proposed revisions to the VC standard, would define a "leak" as a portable monitor reading of 10,000 ppm or more. But current leak detection and elimination programs developed by the companies and approved by E?A incorporate much lower "leak" definitions. The VC Review Study reports that the "leak" definition under the current programs ranges from 300 ppm to 5_ ppm.40 These levels are from three to 0.05 percent of the proposed "leak" definition. .There is no justification for such a change. We request that ERA prepare and air for comment a point-bvpoint analysis of the current leak detection and elimination programs as they compare to generic program proposed for adoption. C. Reporting Requirements NRDC and EDF object to the proposed changes in reporting requirements. In particular, we object to the proposed repeal of the 10-day requirement for reporting relief valve discharges, and the proposed substitution of a quarterly reporting requirement. Frankly, we are surprised E?A would consider such a reduction in the reporting obligation at this time, in the wake of the Bhopal catastrophe. The 10-day reporting requirement serves to give ERA, state officials, and the public contemporaneous knowledge of releases of this carcinogenic compound. Such notice could air in the. orevention of repeat incidents. It also serves the public's rich: 40 4-72, Table 4 ABDOO028241 to know what it is forced to breathe. Finally, allowing a longer interval for reporting these releases would tend to promote a perception that a reported release is merely "past history." We submit that taking these sometimes enormous releases with the seriousness they warrant requires immediate notice. AIn fact, we urge E?A to tighten the 10-day requirement by requiring notice to EPA within 24 hours. Such a requirement would be more consonant with the requirements and spirit of the CERCLA reporting requirements, about which more is said below. NRDC and EDF also object to the proposed repeal of the requirement to report all test results, not just thos .0 that are in excess of standards. An affirmative requirement to report all results promotes accuracy and completeness. It also would permit EPA to build a data base from which to evaluate possible revisions of the substantive requirements of the standards during the periodic reviews mandated by Section 112. . D. CERCLA Reporting Requirements The Vinyl Institute has filed comments questioning the applicability of CERCLA and its reporting requirements, which are discussed in the prop-osal.^i In our view, the releases of VC from relief valve discharges are clearly reportable under CERCLA, except to the extent that they are the result of genuine emergencies. Any non-emergency, preventable release is not permitted by the current standard (and should nor be under any revised standard). Such a release is not a "federally permitted 41 50 Fee. Reg. at 1193. ABD0002S242 release" and therefore is not exempt from the CERCLA reporting requirements. D. Conelusion For the reasons explained above, NRDC and EDF object to the proposed changes in the relief valve discharge standards, the failure to propose and promulgate gasholder systems and other effective controls, the proposed changes in the leak detection and elimination requirments (such as the definition of a "leak"), and the reporting requirements. III. Conclu sion NRDC and EDF request that EPA re-examine the withdrawal of the proposed standard for the reasons given in our petition for reconsideration contained in Part I of this submission. We also request that EPA reconsider and modify its proposed chances in the VC standard in conformity with our comments set forth in Part II.