EPA Proposed Amendments to Air Regulations for the Oil and Natural Gas Industry (Full Copy)

1. Error! Main Document Only.The EPA Administrator, Lisa P. Jackson,signed the following notice on 07/28/2011, and EPA is submitting it forpublication in the Federal Register (FR). While we have taken steps toensure the accuracy of this Internet version of the rule, it is not theofficial version of the rule for purposes of compliance. Please referto the official version in a forthcoming FR publication, which willappear on the Government Printing Offices FDSys website(http://fdsys.gpo.gov/fdsys/search/home.action) and on Regulations.gov(http://www.regulations.gov) in Docket No. EPA-HQ-OAR-2010-0505. Oncethe official version of this document is published in the FR, thisversion will be removed from the Internet and replaced with a link tothe official version. 6560-50-PENVIRONMENTAL PROTECTION AGENCY40 CFR Parts 60 and 63[EPA-HQ-OAR-2010-0505; FRL- ]RIN 2060-AP76Standards of Performance for New Stationary Sources: Oiland Natural Gas Production and Natural Gas Transmission andDistribution; National Emission Standards for Hazardous AirPollutants From Oil and Natural Gas Production Facilities;and National Emission Standards for Hazardous AirPollutants From Natural Gas Transmission and StorageFacilitiesAGENCY: Environmental Protection Agency (EPA).ACTION: Proposed rule.SUMMARY: This action announces how the EPA will address thereviews of the new source performance standards forvolatile organic compound and sulfur dioxide emissions fromnatural gas processing plants. We are adding to the sourcecategory list any oil and gas operation not covered by thecurrent listing. This action also includes proposedamendments to the existing new source performance standardsfor volatile organic compounds from natural gas processingplants and proposed standards for operations that are not

2. Page 2 of 604covered by the existing new source performance standards.In addition, this action proposes how the EPA will addressthe residual risk and technology review conducted for theoil and natural gas production and natural gas transmissionand storage national emission standards for hazardous airpollutants. This action further proposes standards foremission sources within these two source categories thatare not currently addressed, as well as amendments toimprove aspects of these national emission standards forhazardous air pollutants related to applicability andimplementation. Finally, this action addresses provisionsin these new source performance standards and nationalemission standards for hazardous air pollutants related toemissions during periods of startup, shutdown andmalfunction.DATES: Comments must be received on or before [INSERT DATE60 DAYS FROM DATE OF PUBLICATION IN THE FEDERAL REGISTER].Public Hearing. Three public hearings will be held toprovide the public an opportunity to provide comments onthis proposed rulemaking. One will be held in the Dallas,Texas area, on [INSERT DATE], one in Pittsburgh,Pennsylvania, on [INSERT DATE], and one in Denver,Colorado, on [INSERT DATE]. Each hearing will convene at This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

3. Page 3 of 60410:00 a.m. local time. For additional information on thepublic hearings and requesting to speak, see theSUPPLEMENTARY INFORMATION section of this preamble.ADDRESSES: Submit your comments, identified by Docket IDNumber EPA-HQ-OAR-2010-0505, by one of the followingmethods: • Federal eRulemaking Portal: http://www.regulations.gov: Follow the instructions for submitting comments. • Agency Website: http://www.epa.gov/oar/docket.html. Follow the instructions for submitting comments on the Air and Radiation Docket website. • Email: a-and-r-docket@epa.gov. Include Docket ID Number EPA-HQ-OAR-2010-0505 in the subject line of the message. • Facsimile: (202) 566-9744. • Mail: Attention Docket ID Number EPA-HQ-OAR-2010- 0505, 1200 Pennsylvania Ave., NW, Washington, DC 20460. Please include a total of two copies. In addition, please mail a copy of your comments on the information collection provisions to the Office of Information and Regulatory Affairs, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

4. Page 4 of 604 Office of Management and Budget (OMB), Attn: Desk Officer for the EPA, 725 17th Street, NW, Washington, DC 20503. • Hand Delivery: United States Environmental Protection Agency, EPA West (Air Docket), Room 3334, 1301 Constitution Ave., NW, Washington, DC 20004, Attention Docket ID Number EPA-HQ-OAR- 2010-0505. Such deliveries are only accepted during the Docket’s normal hours of operation, and special arrangements should be made for deliveries of boxed information.Instructions: Direct your comments to Docket ID Number EPA-HQ-OAR-2010-0505. The EPA’s policy is that all commentsreceived will be included in the public docket withoutchange and may be made available online athttp://www.regulations.gov, including any personalinformation provided, unless the comment includesinformation claimed to be confidential business information(CBI) or other information whose disclosure is restrictedby statute. Do not submit information that you consider tobe CBI or otherwise protected through www.regulations.govor email. The www.regulations.gov website is an “anonymousaccess” system, which means the EPA will not know your This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

5. Page 5 of 604identity or contact information unless you provide it inthe body of your comment. If you send an email commentdirectly to the EPA without going throughwww.regulations.gov, your email address will beautomatically captured and included as part of the commentthat is placed in the public docket and made available onthe Internet. If you submit an electronic comment, the EPArecommends that you include your name and other contactinformation in the body of your comment and with any diskor CD-ROM you submit. If the EPA cannot read your commentdue to technical difficulties and cannot contact you forclarification, the EPA may not be able to consider yourcomment. Electronic files should avoid the use of specialcharacters, any form of encryption, and be free of anydefects or viruses. For additional information about theEPA’s public docket, visit the EPA Docket Center homepageat http://www.epa.gov/epahome/dockets.htm. For additionalinstructions on submitting comments, go to section II.C ofthe SUPPLEMENTARY INFORMATION section of this preamble.Docket: All documents in the docket are listed in thewww.regulations.gov index. Although listed in the index,some information is not publicly available, e.g., CBI orother information whose disclosure is restricted by This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

6. Page 6 of 604statute. Certain other material, such as copyrightedmaterial, is not placed on the Internet and will bepublicly available only in hard copy. Publicly availabledocket materials are available either electronicallythrough www.regulations.gov or in hard copy at the U.S.Environmental Protection Agency, EPA West (Air Docket),Room 3334, 1301 Constitution Ave., NW, Washington, DC20004. The Public Reading Room is open from 8:30 a.m. to4:30 p.m., Monday through Friday, excluding legal holidays.The telephone number for the Public Reading Room is (202)566-1744, and the telephone number for the Air Docket is(202) 566-1742.FOR FURTHER INFORMATION CONTACT: Bruce Moore, SectorPolicies and Programs Division, Office of Air QualityPlanning and Standards (E143-01), Environmental ProtectionAgency, Research Triangle Park, North Carolina 27711,telephone number: (919) 541-5460; facsimile number: (919)685-3200; email address: moore.bruce@epa.gov.SUPPLEMENTARY INFORMATION:Organization of This Document. The following outline isprovided to aid in locating information in this preamble.I. Preamble Acronyms and AbbreviationsII. General InformationA. Does this action apply to me? This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

7. Page 7 of 604B. Where can I get a copy of this document and other related information?C. What should I consider as I prepare my comments for the EPA?D. When will a public hearing occur?III. Background InformationA. What are standards of performance and NSPS?B. What are NESHAP?C. What litigation is related to this proposed action?D. What is a sector-based approach?IV. Oil and Natural Gas SectorV. Summary of Proposed Decisions and ActionsA. What are the proposed revisions to the NSPS?B. What are the proposed decisions and actions related to the NESHAP?C. What are the proposed notification, recordkeeping and reporting requirements for this proposed action?D. What are the innovative compliance approaches being considered?E. How does the NSPS relate to permitting of sources?VI. Rationale for Proposed Action for NSPSA. What did we evaluate relative to NSPS?B. What are the results of our evaluations and proposed actions relative to NSPS?VII. Rationale for Proposed Action for NESHAPA. What data were used for the NESHAP analyses?B. What are the proposed decisions regarding certain unregulated emissions sources?C. How did we perform the risk assessment and what are the results and proposed decisions?D. How did we perform the technology review and what are the results and proposed decisions?E. What other actions are we proposing?VIII. What are the cost, environmental, energy and economic impacts of the proposed 40 CFR part 60, subpart OOOO and amendments to subparts HH and HHH of 40 CFR part 63?A. What are the affected sources?B. How are the impacts for this proposal evaluated?C. What are the air quality impacts?D. What are the water quality and solid waste impacts?E. What are the secondary impacts?F. What are the energy impacts?G. What are the cost impacts?H. What are the economic impacts? This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

8. Page 8 of 604I. What are the benefits?IX. Request for CommentsX. Submitting Data CorrectionsXI. Statutory and Executive Order ReviewsA. Executive Order 12866: Regulatory Planning and Review and Executive Order 13563: Improving Regulation and Regulatory ReviewB. Paperwork Reduction ActC. Regulatory Flexibility ActD. Unfunded Mandates Reform ActE. Executive Order 13132: FederalismF. Executive Order 13175: Consultation and Coordination with Indian Tribal GovernmentsG. Executive Order 13045: Protection of Children from Environmental Health and Safety RisksH. Executive Order 13211: Actions Concerning Regulations That Significantly Affect Energy Supply, Distribution, or UseI. National Technology Transfer and Advancement ActJ. Executive Order 12898: Federal Actions to Address Environmental Justice in Minority Populations and Low- Income PopulationsI. Preamble Acronyms and Abbreviations Several acronyms and terms used to describe industrialprocesses, data inventories and risk modeling are includedin this preamble. While this may not be an exhaustive list,to ease the reading of this preamble and for referencepurposes, the following terms and acronyms are definedhere:ACGIH American Conference of Governmental Industrial HygienistsADAF Age-Dependent Adjustment FactorsAEGL Acute Exposure Guideline LevelsAERMOD The air dispersion model used by the HEM-3 modelAPI American Petroleum InstituteBACT Best Available Control TechnologyBID Background Information Document This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

9. Page 9 of 604BPD Barrels Per DayBSER Best System of Emission ReductionBTEX Benzene, Ethylbenzene, Toluene and XyleneCAA Clean Air ActCalEPA California Environmental Protection AgencyCBI Confidential Business InformationCEM Continuous Emissions MonitoringCEMS Continuous Emissions Monitoring SystemCFR Code of Federal RegulationsCIIT Chemical Industry Institute of ToxicologyCO Carbon MonoxideCO2 Carbon DioxideCO2e Carbon Dioxide EquivalentDOE Department of EnergyECHO Enforcement and Compliance History Onlinee-GGRT Electronic Greenhouse Gas Reporting ToolEJ Environmental JusticeEPA Environmental Protection AgencyERPG Emergency Response Planning GuidelinesERT Electronic Reporting ToolGCG Gas Condensate GlycolGHG Greenhouse GasGOR Gas to Oil RatioGWP Global Warming PotentialHAP Hazardous Air PollutantsHEM-3 Human Exposure Model, version 3HI Hazard IndexHP HorsepowerHQ Hazard QuotientH2S Hydrogen SulfideICR Information Collection RequestIPCC Intergovernmental Panel on Climate ChangeIRIS Integrated Risk Information Systemkm KilometerkW KilowattsLAER Lowest Achievable Emission Ratelb PoundsLDAR Leak Detection and RepairMACT Maximum Achievable Control TechnologyMACT Code Code within the NEI used to identify processes included in a source categoryMcf Thousand Cubic FeetMg/yr Megagrams per yearMIR Maximum Individual Risk This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

10. Page 10 of 604MIRR Monitoring, Inspection, Recordkeeping and ReportingMMtCO2e Million Metric Tons of Carbon Dioxide EquivalentsNAAQS National Ambient Air Quality StandardsNAC/AEGL National Advisory Committee for Acute Exposure Guideline Levels for Hazardous SubstancesNAICS North American Industry Classification SystemNAS National Academy of SciencesNATA National Air Toxics AssessmentNEI National Emissions InventoryNEMS National Energy Modeling SystemNESHAP National Emissions Standards for Hazardous Air PollutantsNGL Natural Gas LiquidsNIOSH National Institutes for Occupational Safety and HealthNOx Oxides of NitrogenNRC National Research CouncilNSPS New Source Performance StandardsNSR New Source ReviewNTTAA National Technology Transfer and Advancement ActOAQPS Office of Air Quality Planning and StandardsOMB Office of Management and BudgetPB-HAP Hazardous air pollutants known to be persistent and bio-accumulative in the environmentPFE Potential for Flash EmissionsPM Particulate MatterPM2.5 Particulate Matter (2.5 microns and less)POM Polycyclic Organic MatterPPM Parts Per MillionPPMV Parts Per Million by VolumePSIG Pounds per square inch gaugePTE Potential to EmitQA Quality AssuranceRACT Reasonably Available Control TechnologyRBLC RACT/BACT/LAER ClearinghouseREC Reduced Emissions CompletionsREL CalEPA Reference Exposure LevelRFA Regulatory Flexibility ActRfC Reference ConcentrationRfD Reference DoseRIA Regulatory Impact AnalysisRICE Reciprocating Internal Combustion Engines This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

11. Page 11 of 604RTR Residual Risk and Technology ReviewSAB Science Advisory BoardSBREFA Small Business Regulatory Enforcement Fairness ActSCC Source Classification CodesSCFH Standard Cubic Feet Per HourSCFM Standard Cubic Feet Per MinuteSCM Standard Cubic MetersSCMD Standard Cubic Meters Per DaySCOT Shell Claus Offgas TreatmentSIP State Implementation PlanSISNOSE Significant Economic Impact on a Substantial Number of Small EntitiesS/L/T State and Local and Tribal AgenciesSO2 Sulfur DioxideSSM Startup, Shutdown and MalfunctionSTEL Short-term Exposure LimitTLV Threshold Limit ValueTOSHI Target Organ-Specific Hazard IndexTPY Tons per YearTRIM Total Risk Integrated Modeling SystemTRIM.FaTE A spatially explicit, compartmental mass balance model that describes the movement and transformation of pollutants over time, through a user-defined, bounded system that includes both biotic and abiotic compartmentsTSD Technical Support DocumentUF Uncertainty FactorUMRA Unfunded Mandates Reform ActURE Unit Risk EstimateVCS Voluntary Consensus StandardsVOC Volatile Organic CompoundsVRU Vapor Recovery UnitII. General InformationA. Does this action apply to me? The regulated industrial source categories that arethe subject of this proposal are listed in Table 1 of thispreamble. These standards and any changes considered inthis rulemaking would be directly applicable to sources as This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

12. Page 12 of 604a Federal program. Thus, Federal, state, local and tribalgovernment entities are not affected by this proposedaction. TABLE 1. INDUSTRIAL SOURCE CATEGORIES AFFECTED BY THIS PROPOSED ACTION Category NAICS Examples of regulated code 1 entitiesIndustry . . . . Crude Petroleum and Natural 211111 Gas Extraction 211112 Natural Gas Liquid Extraction 221210 Natural Gas Distribution 486110 Pipeline Distribution of Crude Oil 486210 Pipeline Transportation of Natural GasFederal government . . . . Not affected.State/local/tribal . . . . Not affected.government1 North American Industry Classification System. This table is not intended to be exhaustive, butrather provides a guide for readers regarding entitieslikely to be affected by this action. To determine whetheryour facility would be regulated by this action, you shouldexamine the applicability criteria in the regulations. Ifyou have any questions regarding the applicability of thisaction to a particular entity, contact the person listed inthe preceding FOR FURTHER INFORMATION CONTACT section. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

13. Page 13 of 604B. Where can I get a copy of this document and otherrelated information? In addition to being available in the docket, anelectronic copy of this proposal will also be available onthe EPA’s website. Following signature by the EPAAdministrator, a copy of this proposed action will beposted on the EPA’s website at the following address:http://www.epa.gov/airquality/oilandgas. Additional information is available on the EPA’sResidual Risk and Technology Review (RTR) website athttp://www.epa.gov/ttn/atw/rrisk/oarpg.html. Thisinformation includes the most recent version of the rule,source category descriptions, detailed emissions and otherdata that were used as inputs to the risk assessments.C. What should I consider as I prepare my comments for theEPA?Submitting CBI. Do not submit information containing CBI tothe EPA through www.regulations.gov or email. Clearly markthe part or all of the information that you claim to beCBI. For CBI information on a disk or CD ROM that you mailto the EPA, mark the outside of the disk or CD ROM as CBIand then identify electronically within the disk or CD ROMthe specific information that is claimed as CBI. In This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

14. Page 14 of 604addition to one complete version of the comment thatincludes information claimed as CBI, a copy of the commentthat does not contain the information claimed as CBI mustbe submitted for inclusion in the public docket. If yousubmit a CD ROM or disk that does not contain CBI, mark theoutside of the disk or CD ROM clearly that it does notcontain CBI. Information not marked as CBI will be includedin the public docket and the EPA’s electronic public docketwithout prior notice. Information marked as CBI will not bedisclosed except in accordance with procedures set forth in40 CFR part 2. Send or deliver information identified asCBI only to the following address: Roberto Morales, OAQPSDocument Control Officer (C404-02), EnvironmentalProtection Agency, Office of Air Quality Planning andStandards, Research Triangle Park, North Carolina 27711,Attention Docket ID Number EPA-HQ-OAR-2010-0505.D. When will a public hearing occur? We will hold three public hearings, one in the Dallas,Texas area, one in Pittsburgh, Pennsylvania, and one inDenver, Colorado. If you are interested in attending orspeaking at one of the public hearings, contact Ms. JoanRogers at (919) 541-4487 by [INSERT DATE 11 DAYS FROM DATEOF PUBLICATION IN THE FEDERAL REGISTER]. Details on the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

15. Page 15 of 604public hearings will be provided in a separate notice andwe will specify the time and date of the public hearings onhttp://www.epa.gov/airquality/oilandgas. If no one requeststo speak at one of the public hearings [INSERT DATE 11 DAYSFROM DATE OF PUBLICATION IN THE FEDERAL REGISTER], thenthat public hearing will be cancelled without furthernotice.III. Background InformationA. What are standards of performance and NSPS?1. What is the statutory authority for standards ofperformance and NSPS? Section 111 of the Clean Air Act (CAA) requires theEPA Administrator to list categories of stationary sources,if such sources cause or contribute significantly to airpollution, which may reasonably be anticipated to endangerpublic health or welfare. The EPA must then issueperformance standards for such source categories. Aperformance standard reflects the degree of emissionlimitation achievable through the application of the “bestsystem of emission reduction” (BSER) which the EPAdetermines has been adequately demonstrated. The EPA mayconsider certain costs and nonair quality health andenvironmental impact and energy requirements when This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

16. Page 16 of 604establishing performance standards. Whereas CAA section 112standards are issued for existing and new stationarysources, standards of performance are issued for new andmodified stationary sources. These standards are referredto as new source performance standards (NSPS). The EPA hasthe authority to define the source categories, determinethe pollutants for which standards should be developed,identify the facilities within each source category to becovered and set the emission level of the standards. CAA section 111(b)(1)(B) requires the EPA to “at leastevery 8 years review and, if appropriate, revise”performance standards unless the “Administrator determinesthat such review is not appropriate in light of readilyavailable information on the efficacy” of the standard.When conducting a review of an existing performancestandard, the EPA has discretion to revise that standard toadd emission limits for pollutants or emission sources notcurrently regulated for that source category. In setting or revising a performance standard, CAAsection 111(a)(1) provides that performance standards areto “reflect the degree of emission limitation achievablethrough the application of the best system of emissionreduction which (taking into account the cost of achieving This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

17. Page 17 of 604such reduction and any nonair quality health andenvironmental impact and energy requirements) theAdministrator determines has been adequately demonstrated.”In this notice, we refer to this level of control as theBSER. In determining BSER, we typically conduct atechnology review that identifies what emission reductionsystems exist and how much they reduce air pollution inpractice. Next, for each control system identified, weevaluate its costs, secondary air benefits (or disbenefits)resulting from energy requirements and nonair qualityimpacts such as solid waste generation. Based on ourevaluation, we would determine BSER. The resultant standardis usually a numerical emissions limit, expressed as aperformance level (i.e., a rate-based standard or percentcontrol), that reflects the BSER. Although such standardsare based on the BSER, the EPA may not prescribe aparticular technology that must be used to comply with aperformance standard, except in instances where theAdministrator determines it is not feasible to prescribe orenforce a standard of performance. Typically, sourcesremain free to elect whatever control measures that theychoose to meet the emission limits. Upon promulgation, a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

18. Page 18 of 604NSPS becomes a national standard to which all new, modifiedor reconstructed sources must comply.2. What is the regulatory history regarding performancestandards for the oil and natural gas sector? In 1979, the EPA listed crude oil and natural gasproduction on its priority list of source categories forpromulgation of NSPS (44 FR 49222, August 21, 1979). OnJune 24, 1985 (50 FR 26122), the EPA promulgated a NSPS forthe source category that addressed volatile organiccompound (VOC) emissions from leaking components at onshorenatural gas processing plants (40 CFR part 60, subpartKKK). On October 1, 1985 (50 FR 40158), a second NSPS waspromulgated for the source category that regulates sulfurdioxide (SO2) emissions from natural gas processing plants(40 CFR part 60, subpart LLL). Other than natural gasprocessing plants, EPA has not previously set NSPS for avariety of oil and natural gas operations.B. What are NESHAP?1. What is the statutory authority for NESHAP? Section 112 of the CAA establishes a two-stageregulatory process to address emissions of hazardous airpollutants (HAP) from stationary sources. In the firststage, after the EPA has identified categories of sources This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

19. Page 19 of 604emitting one or more of the HAP listed in section 112(b) ofthe CAA, section 112(d) of the CAA calls for us topromulgate national emission standards for hazardous airpollutants (NESHAP) for those sources. “Major sources” arethose that emit or have the potential to emit (PTE) 10 tonsper year (tpy) or more of a single HAP or 25 tpy or more ofany combination of HAP. For major sources, thesetechnology-based standards must reflect the maximum degreeof emission reductions of HAP achievable (after consideringcost, energy requirements and nonair quality health andenvironmental impacts) and are commonly referred to asmaximum achievable control technology (MACT) standards. MACT standards are to reflect application of measures,processes, methods, systems or techniques, including, butnot limited to, measures which, (1) reduce the volume of oreliminate pollutants through process changes, substitutionof materials or other modifications, (2) enclose systems orprocesses to eliminate emissions, (3) capture or treatpollutants when released from a process, stack, storage orfugitive emissions point, (4) are design, equipment, workpractice or operational standards (including requirementsfor operator training or certification) or (5) are acombination of the above. CAA section 112(d)(2)(A)-(E). The This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

20. Page 20 of 604MACT standard may take the form of a design, equipment,work practice or operational standard where the EPA firstdetermines either that, (1) a pollutant cannot be emittedthrough a conveyance designed and constructed to emit orcapture the pollutant or that any requirement for or use ofsuch a conveyance would be inconsistent with law or (2) theapplication of measurement methodology to a particularclass of sources is not practicable due to technologicaland economic limitations. CAA sections 112(h)(1)-(2). The MACT “floor” is the minimum control level allowedfor MACT standards promulgated under CAA section 112(d)(3),and may not be based on cost considerations. For newsources, the MACT floor cannot be less stringent than theemission control that is achieved in practice by the best-controlled similar source. The MACT floors for existingsources can be less stringent than floors for new sources,but they cannot be less stringent than the average emissionlimitation achieved by the best-performing 12 percent ofexisting sources in the category or subcategory (or thebest-performing five sources for categories orsubcategories with fewer than 30 sources). In developingMACT standards, we must also consider control options thatare more stringent than the floor. We may establish This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

21. Page 21 of 604standards more stringent than the floor based on theconsideration of the cost of achieving the emissionsreductions, any nonair quality health and environmentalimpacts and energy requirements. The EPA is then required to review these technology-based standards and to revise them “as necessary (takinginto account developments in practices, processes, andcontrol technologies)” no less frequently than every 8years, under CAA section 112(d)(6). In conducting thisreview, the EPA is not obliged to completely recalculatethe prior MACT determination. NRDC v. EPA, 529 F.3d 1077,1084 (D.C. Cir. 2008). The second stage in standard-setting focuses onreducing any remaining “residual” risk according to CAAsection 112(f). This provision requires, first, that theEPA prepare a Report to Congress discussing (among otherthings) methods of calculating risk posed (or potentiallyposed) by sources after implementation of the MACTstandards, the public health significance of those risks,and the EPA’s recommendations as to legislation regardingsuch remaining risk. The EPA prepared and submitted thisreport (Residual Risk Report to Congress, EPA–453/R–99–001)in March 1999. Congress did not act in response to the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

22. Page 22 of 604report, thereby triggering the EPA’s obligation under CAAsection 112(f)(2) to analyze and address residual risk. CAA section 112(f)(2) requires us to determine forsource categories subject to MACT standards, whether theemissions standards provide an ample margin of safety toprotect public health. If the MACT standards for HAP“classified as a known, probable, or possible humancarcinogen do not reduce lifetime excess cancer risks tothe individual most exposed to emissions from a source inthe category or subcategory to less than 1-in-1 million,”the EPA must promulgate residual risk standards for thesource category (or subcategory), as necessary, to providean ample margin of safety to protect public health. Indoing so, the EPA may adopt standards equal to existingMACT standards if the EPA determines that the existingstandards are sufficiently protective. NRDC v. EPA, 529F.3d 1077, 1083 (D.C. Cir. 2008). (“If EPA determines thatthe existing technology-based standards provide an “amplemargin of safety,” then the Agency is free to readopt thosestandards during the residual risk rulemaking.”) The EPAmust also adopt more stringent standards, if necessary, to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

23. Page 23 of 604prevent an adverse environmental effect,1 but must considercost, energy, safety and other relevant factors in doingso. Section 112(f)(2) of the CAA expressly preserves ouruse of a two-step process for developing standards toaddress any residual risk and our interpretation of “amplemargin of safety” developed in the National EmissionStandards for Hazardous Air Pollutants: Benzene Emissionsfrom Maleic Anhydride Plants, Ethylbenzene/Styrene Plants,Benzene Storage Vessels, Benzene Equipment Leaks, and CokeBy-Product Recovery Plants (Benzene NESHAP) (54 FR 38044,September 14, 1989). The first step in this process is thedetermination of acceptable risk. The second step providesfor an ample margin of safety to protect public health,which is the level at which the standards are set (unless amore stringent standard is required to prevent, taking intoconsideration costs, energy, safety, and other relevantfactors, an adverse environmental effect).1 “Adverse environmental effect” is defined in CAA section112(a)(7) as any significant and widespread adverse effect, whichmay be reasonably anticipated to wildlife, aquatic life ornatural resources, including adverse impacts on populations ofendangered or threatened species or significant degradation ofenvironmental qualities over broad areas. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

24. Page 24 of 604 The terms “individual most exposed,” “acceptablelevel,” and “ample margin of safety” are not specificallydefined in the CAA. However, CAA section 112(f)(2)(B)preserves the interpretation set out in the Benzene NESHAP,and the United States Court of Appeals for the District ofColumbia Circuit in NRDC v. EPA, 529 F.3d 1077, concludedthat the EPA’s interpretation of subsection 112(f)(2) is areasonable one. See NRDC v. EPA, 529 F.3d at 1083 (D.C.Cir., “[S]ubsection 112(f)(2)(B) expressly incorporatesEPA’s interpretation of the Clean Air Act from the Benzenestandard, complete with a citation to the FederalRegister”). (D.C. Cir. 2008). See also, A LegislativeHistory of the Clean Air Act Amendments of 1990, volume 1,p. 877 (Senate debate on Conference Report). We notifiedCongress in the Residual Risk Report to Congress that weintended to use the Benzene NESHAP approach in making CAAsection 112(f) residual risk determinations (EPA–453/R–99–001, p. ES–11). In the Benzene NESHAP, we stated as an overallobjective: * * * in protecting public health with an ample margin of safety, we strive to provide maximum feasible protection against risks to health from hazardous air This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

25. Page 25 of 604 pollutants by, (1) protecting the greatest number of persons possible to an individual lifetime risk level no higher than approximately 1-in-1 million; and (2) limiting to no higher than approximately 1-in-10 thousand [i.e., 100-in-1 million] the estimated risk that a person living near a facility would have if he or she were exposed to the maximum pollutant concentrations for 70 years. The Agency also stated that, “The EPA also considersincidence (the number of persons estimated to suffer canceror other serious health effects as a result of exposure toa pollutant) to be an important measure of the health riskto the exposed population. Incidence measures the extent ofhealth risk to the exposed population as a whole, byproviding an estimate of the occurrence of cancer or otherserious health effects in the exposed population.” TheAgency went on to conclude that “estimated incidence wouldbe weighed along with other health risk information injudging acceptability.” As explained more fully in ourResidual Risk Report to Congress, the EPA does not define“rigid line[s] of acceptability,” but considers ratherbroad objectives to be weighed with a series of otherhealth measures and factors (EPA–453/R–99–001, p. ES–11). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

26. Page 26 of 604The determination of what represents an “acceptable” riskis based on a judgment of “what risks are acceptable in theworld in which we live” (Residual Risk Report to Congress,p. 178, quoting the Vinyl Chloride decision at 824 F.2d1165) recognizing that our world is not risk-free. In the Benzene NESHAP, we stated that “EPA willgenerally presume that if the risk to [the maximum exposed]individual is no higher than approximately 1-in-10thousand, that risk level is considered acceptable.” 54 FR38045. We discussed the maximum individual lifetime cancerrisk (or maximum individual risk (MIR)) as being “theestimated risk that a person living near a plant would haveif he or she were exposed to the maximum pollutantconcentrations for 70 years.” Id. We explained that thismeasure of risk “is an estimate of the upper bound of riskbased on conservative assumptions, such as continuousexposure for 24 hours per day for 70 years.” Id. Weacknowledge that maximum individual lifetime cancer risk“does not necessarily reflect the true risk, but displays aconservative risk level which is an upper-bound that isunlikely to be exceeded.” Id. Understanding that there are both benefits andlimitations to using maximum individual lifetime cancer This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

27. Page 27 of 604risk as a metric for determining acceptability, weacknowledged in the 1989 Benzene NESHAP that “considerationof maximum individual risk * * * must take into account thestrengths and weaknesses of this measure of risk.” Id.Consequently, the presumptive risk level of 100-in-1million (1-in-10 thousand) provides a benchmark for judgingthe acceptability of maximum individual lifetime cancerrisk, but does not constitute a rigid line for making thatdetermination. The Agency also explained in the 1989 Benzene NESHAPthe following: “In establishing a presumption for MIR,rather than a rigid line for acceptability, the Agencyintends to weigh it with a series of other health measuresand factors. These include the overall incidence of canceror other serious health effects within the exposedpopulation, the numbers of persons exposed within eachindividual lifetime risk range and associated incidencewithin, typically, a 50-kilometer (km) exposure radiusaround facilities, the science policy assumptions andestimation uncertainties associated with the risk measures,weight of the scientific evidence for human health effects,other quantified or unquantified health effects, effects This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

28. Page 28 of 604due to co-location of facilities and co-emission ofpollutants.” Id. In some cases, these health measures and factors takentogether may provide a more realistic description of themagnitude of risk in the exposed population than thatprovided by maximum individual lifetime cancer risk alone.As explained in the Benzene NESHAP, “[e]ven though therisks judged “acceptable” by the EPA in the first step ofthe Vinyl Chloride inquiry are already low, the second stepof the inquiry, determining an “ample margin of safety,”again includes consideration of all of the health factors,and whether to reduce the risks even further.” In the amplemargin of safety decision process, the Agency againconsiders all of the health risks and other healthinformation considered in the first step. Beyond thatinformation, additional factors relating to the appropriatelevel of control will also be considered, including costsand economic impacts of controls, technologicalfeasibility, uncertainties and any other relevant factors.Considering all of these factors, the Agency will establishthe standard at a level that provides an ample margin ofsafety to protect the public health, as required by CAAsection 112(f). 54 FR 38046. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

29. Page 29 of 6042. How do we consider the risk results in making decisions? As discussed in the previous section of this preamble,we apply a two-step process for developing standards toaddress residual risk. In the first step, the EPAdetermines if risks are acceptable. This determination“considers all health information, including riskestimation uncertainty, and includes a presumptive limit onmaximum individual lifetime [cancer] risk (MIR)2 ofapproximately 1-in-10 thousand [i.e., 100-in-1 million].”54 FR 38045. In the second step of the process, the EPAsets the standard at a level that provides an ample marginof safety “in consideration of all health information,including the number of persons at risk levels higher thanapproximately 1-in-1 million, as well as other relevantfactors, including costs and economic impacts,technological feasibility, and other factors relevant toeach particular decision.” Id. In past residual risk determinations, the EPApresented a number of human health risk metrics associatedwith emissions from the category under review, including:2 Although defined as “maximum individual risk,” MIR refers onlyto cancer risk. MIR, one metric for assessing cancer risk, is theestimated risk were an individual exposed to the maximum level ofa pollutant for a lifetime. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

30. Page 30 of 604The MIR; the numbers of persons in various risk ranges;cancer incidence; the maximum noncancer hazard index (HI);and the maximum acute noncancer hazard. In estimatingrisks, the EPA considered source categories under reviewthat are located near each other and that affect the samepopulation. The EPA provided estimates of the expecteddifference in actual emissions from the source categoryunder review and emissions allowed pursuant to the sourcecategory MACT standard. The EPA also discussed andconsidered risk estimation uncertainties. The EPA isproviding this same type of information in support of theseactions. The Agency acknowledges that the Benzene NESHAPprovides flexibility regarding what factors the EPA mightconsider in making our determinations and how they might beweighed for each source category. In responding to commenton our policy under the Benzene NESHAP, the EPA explainedthat: “The policy chosen by the Administrator permitsconsideration of multiple measures of health risk. Not onlycan the MIR figure be considered, but also incidence, thepresence of noncancer health effects, and the uncertaintiesof the risk estimates. In this way, the effect on the mostexposed individuals can be reviewed as well as the impact This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

31. Page 31 of 604on the general public. These factors can then be weighed ineach individual case. This approach complies with the VinylChloride mandate that the Administrator ascertain anacceptable level of risk to the public by employing [her]expertise to assess available data. It also complies withthe Congressional intent behind the CAA, which did notexclude the use of any particular measure of public healthrisk from the EPAs consideration with respect to CAAsection 112 regulations, and, thereby, implicitly permitsconsideration of any and all measures of health risk whichthe Administrator, in [her] judgment, believes areappropriate to determining what will ‘protect the publichealth.’” For example, the level of the MIR is only one factorto be weighed in determining acceptability of risks. TheBenzene NESHAP explains “an MIR of approximately 1-in-10thousand should ordinarily be the upper end of the range ofacceptability. As risks increase above this benchmark, theybecome presumptively less acceptable under CAA section 112,and would be weighed with the other health risk measuresand information in making an overall judgment onacceptability. Or, the Agency may find, in a particularcase, that a risk that includes MIR less than the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

32. Page 32 of 604presumptively acceptable level is unacceptable in the lightof other health risk factors.” Similarly, with regard tothe ample margin of safety analysis, the Benzene NESHAPstates that: “EPA believes the relative weight of the manyfactors that can be considered in selecting an ample marginof safety can only be determined for each specific sourcecategory. This occurs mainly because technological andeconomic factors (along with the health-related factors)vary from source category to source category.”3. What is the regulatory history regarding NESHAP for theoil and natural gas sector? On July 16, 1992 (57 FR 31576), the EPA published alist of major and area sources for which NESHAP are to bepublished (i.e., the source category list). Oil and naturalgas production facilities were listed as a category ofmajor sources. On February 12, 1998 (63 FR 7155), the EPAamended the source category list to add Natural GasTransmission and Storage as a major source category. On June 17, 1999 (64 FR 32610), the EPA promulgatedMACT standards for the Oil and Natural Gas Production andNatural Gas Transmission and Storage major sourcecategories. The Oil and Natural Gas Production NESHAP (40CFR part 63, subpart HH) contains standards for HAP This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

33. Page 33 of 604emissions from glycol dehydration process vents, storagevessels and natural gas processing plant equipment leaks.The Natural Gas Transmission and Storage NESHAP (40 CFRpart 63, subpart HHH) contains standards for glycoldehydration process vents. In addition to these NESHAP for major sources, the EPAalso promulgated NESHAP for the Oil and Natural GasProduction area source category on January 3, 2007 (72 FR26). These area source standards, which are based ongenerally available control technology, are also containedin 40 CFR part 63, subpart HH. This proposed action doesnot impact these area source standards.C. What litigation is related to this proposed action? On January 14, 2009, pursuant to section 304(a)(2) ofthe CAA, WildEarth Guardians and the San Juan CitizensAlliance filed a Complaint alleging that the EPA failed tomeet its obligations under CAA sections 111(b)(1)(B),112(d)(6) and 112(f)(2) to take actions relative to thereview/revision of the NSPS and the NESHAP with respect tothe Oil and Natural Gas Production source category. OnFebruary 4, 2010, the Court entered a consent decree This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

34. Page 34 of 604requiring the EPA to sign by July 28, 2011,3 proposedstandards and/or determinations not to issue standardspursuant to CAA sections 111(b)(1)(B), 112(d)(6) and112(f)(2) and to take final action by February 28, 2012.D. What is a sector-based approach? Sector-based approaches are based on integratedassessments that consider multiple pollutants in acomprehensive and coordinated manner to manage emissionsand CAA requirements. One of the many ways we can addresssector-based approaches is by reviewing multiple regulatoryprograms together whenever possible, consistent with allapplicable legal requirements. This approach essentiallyexpands the technical analyses on costs and benefits ofparticular technologies, to consider the interactions ofrules that regulate sources. The benefit of multi-pollutantand sector-based analyses and approaches includes theability to identify optimum strategies, consideringfeasibility, cost impacts and benefits across the differentpollutant types while streamlining administrative andcompliance complexities and reducing conflicting and3 On April 27, 2011, pursuant to paragraph 10(a) of the ConsentDecree, the parties filed with the Court a written stipulationthat changes the proposal date from January 31, 2011, to July 28,2011, and the final action date from November 30, 2011, toFebruary 28, 2012. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

35. Page 35 of 604redundant requirements, resulting in added certainty andeasier implementation of control strategies for the sectorunder consideration. In order to benefit from a sector-based approach for the oil and gas industry, the EPAanalyzed how the NSPS and NESHAP under consideration relateto each other and other regulatory requirements currentlyunder review for oil and gas facilities. In this analysis,we looked at how the different control requirements thatresult from these requirements interact, including thedifferent regulatory deadlines and control equipmentrequirements that result, the different reporting andrecordkeeping requirements and opportunities for states toaccount for reductions resulting from this rulemaking intheir State Implementation Plans (SIP). The requirementsanalyzed affect criteria pollutant, HAP and methaneemissions from oil and natural gas processes and cover theNSPS and NESHAP reviews. As a result of the sector-basedapproach, this rulemaking will reduce conflicting andredundant requirements. Also, the sector-based approachfacilitated the streamlining of monitoring, recordkeepingand reporting requirements, thus, reducing administrativeand compliance complexities associated with complying withmultiple regulations. In addition, the sector-based This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

36. Page 36 of 604approach promotes a comprehensive control strategy thatmaximizes the co-control of multiple regulated pollutantswhile obtaining emission reductions as co-benefits.IV. Oil and Natural Gas Sector The oil and natural gas sector includes operationsinvolved in the extraction and production of oil andnatural gas, as well as the processing, transmission anddistribution of natural gas. Specifically for oil, thesector includes all operations from the well to the pointof custody transfer at a petroleum refinery. For naturalgas, the sector includes all operations from the well tothe customer. The oil and natural gas operations cangenerally be separated into four segments: (1) Oil andnatural gas production, (2) natural gas processing, (3)natural gas transmission and (4) natural gas distribution.Each of these segments is briefly discussed below. Oil and natural gas production includes both onshoreand offshore operations. Production operations include thewells and all related processes used in the extraction,production, recovery, lifting, stabilization, separation ortreating of oil and/or natural gas (including condensate).Production components may include, but are not limited to,wells and related casing head, tubing head and “Christmas This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

37. Page 37 of 604tree” piping, as well as pumps, compressors, heatertreaters, separators, storage vessels, pneumatic devicesand dehydrators. Production operations also include thewell drilling, completion and workover processes andincludes all the portable non-self-propelled apparatusassociated with those operations. Production sites includenot only the “pads” where the wells are located, but alsoinclude stand-alone sites where oil, condensate, producedwater and gas from several wells may be separated, storedand treated. The production sector also includes the lowpressure, small diameter, gathering pipelines and relatedcomponents that collect and transport the oil, gas andother materials and wastes from the wells to the refineriesor natural gas processing plants. None of the operationsupstream of the natural gas processing plant are covered bythe existing NSPS. Offshore oil and natural gas productionoccurs on platform structures that house equipment toextract oil and gas from the ocean or lake floor and thatprocess and/or transfer the oil and gas to storage,transport vessels or onshore. Offshore production can alsoinclude secondary platform structures connected to theplatform structure, storage tanks associated with the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

38. Page 38 of 604platform structure and floating production and offloadingequipment. There are three basic types of wells: Oil wells, gaswells and associated gas wells. Oil wells can have“associated” natural gas that is separated and processed orthe crude oil can be the only product processed. Once thecrude oil is separated from the water and other impurities,it is essentially ready to be transported to the refineryvia truck, railcar or pipeline. We consider the oilrefinery sector separately from the oil and natural gassector. Therefore, at the point of custody transfer at therefinery, the oil leaves the oil and natural gas sector andenters the petroleum refining sector. Natural gas is primarily made up of methane. However,whether natural gas is associated gas from oil wells ornon-associated gas from gas or condensate wells, itcommonly exists in mixtures with other hydrocarbons. Thesehydrocarbons are often referred to as natural gas liquids(NGL). They are sold separately and have a variety ofdifferent uses. The raw natural gas often contains watervapor, hydrogen sulfide (H2S), carbon dioxide (CO2),helium, nitrogen and other compounds. Natural gasprocessing consists of separating certain hydrocarbons and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

39. Page 39 of 604fluids from the natural gas to produced “pipeline quality”dry natural gas. While some of the processing can beaccomplished in the production segment, the completeprocessing of natural gas takes place in the natural gasprocessing segment. Natural gas processing operationsseparate and recover NGL or other non-methane gases andliquids from a stream of produced natural gas throughcomponents performing one or more of the followingprocesses: Oil and condensate separation, water removal,separation of NGL, sulfur and CO2 removal, fractionation ofnatural gas liquid and other processes, such as the captureof CO2 separated from natural gas streams for deliveryoutside the facility. Natural gas processing plants are theonly operations covered by the existing NSPS. The pipeline quality natural gas leaves the processingsegment and enters the transmission segment. Pipelines inthe natural gas transmission segment can be interstatepipelines that carry natural gas across state boundaries orintrastate pipelines, which transport the gas within asingle state. While interstate pipelines may be of a largerdiameter and operated at a higher pressure, the basiccomponents are the same. To ensure that the natural gas This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

40. Page 40 of 604flowing through any pipeline remains pressurized,compression of the gas is required periodically along thepipeline. This is accomplished by compressor stationsusually placed between 40 and 100 mile intervals along thepipeline. At a compressor station, the natural gas entersthe station, where it is compressed by reciprocating orcentrifugal compressors. In addition to the pipelines and compressor stations,the natural gas transmission segment includes undergroundstorage facilities. Underground natural gas storageincludes subsurface storage, which typically consists ofdepleted gas or oil reservoirs and salt dome caverns usedfor storing natural gas. One purpose of this storage is forload balancing (equalizing the receipt and delivery ofnatural gas). At an underground storage site, there aretypically other processes, including compression,dehydration and flow measurement. The distribution segment is the final step indelivering natural gas to customers. The natural gas entersthe distribution segment from delivery points located oninterstate and intrastate transmission pipelines tobusiness and household customers. The delivery point wherethe natural gas leaves the transmission segment and enters This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

41. Page 41 of 604the distribution segment is often called the “citygate.”Typically, utilities take ownership of the gas at thecitygate. Natural gas distribution systems consist ofthousands of miles of piping, including mains and servicepipelines to the customers. Distribution systems sometimeshave compressor stations, although they are considerablysmaller than transmission compressor stations. Distributionsystems include metering stations, which allow distributioncompanies to monitor the natural gas in the system.Essentially, these metering stations measure the flow ofgas and allow distribution companies to track natural gasas it flows through the system. Emissions can occur from a variety of processes andpoints throughout the oil and natural gas sector.Primarily, these emissions are organic compounds such asmethane, ethane, VOC and organic HAP. The most commonorganic HAP are n-hexane and BTEX compounds (benzene,toluene, ethylbenzene and xylenes). Hydrogen sulfide (H2S)and sulfur dioxide (SO2) are emitted from production andprocessing operations that handle and treat “sour gas.”Sour gas is defined as natural gas with a maximum H2S This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

42. Page 42 of 604content of 0.25 gr/100 scf (4ppmv) along with the presenceof CO2. In addition, there are significant emissionsassociated with the reciprocating internal combustionengines and combustion turbines that power compressorsthroughout the oil and natural gas sector. However,emissions from internal combustion engines and combustionturbines are covered by regulations specific to engines andturbines and, thus, are not addressed in this action.V. Summary of Proposed Decisions and Actions Pursuant to CAA sections 111(b), 112(d)(2), 112(d)(6)and 112(f), we are proposing to revise the NSPS and NESHAPrelative to oil and gas to include the standards andrequirements summarized in this section. More details ofthe rationale for these proposed standards and requirementsare provided in sections VI and VII of this preamble. Inaddition, as part of these rationale discussions, wesolicit public comment and data relevant to several issues.The comments we receive during the public comment periodwill help inform the rule development process as we worktoward promulgating a final action.A. What are the proposed revisions to the NSPS? We reviewed the two NSPS that apply to the oil and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

43. Page 43 of 604natural gas industry. Based on our review, we believe thatthe requirements at 40 CFR part 60, subpart KKK, should beupdated to reflect requirements in 40 CFR part 60, subpartVVa for controlling VOC equipment leaks at processingplants. We also believe that the requirements at 40 CFRpart 60, subpart LLL, for controlling SO2 emissions fromnatural gas processing plants should be strengthened forfacilities with the highest sulfur feed rates and thehighest H2S concentrations. For a more detailed discussion,please see section VI.B.1 of this preamble. In addition, there are significant VOC emissions fromoil and natural gas operations that are not covered by thetwo existing NSPS, including other emissions at processingplants and emissions from upstream production, as well astransmission and storage facilities. In the 1984 noticethat listed source categories (including Oil and NaturalGas) for promulgation of NSPS, we noted that there werediscrepancies between the source category names on the listand those in the background document, and we clarified ourintent to address all sources under an industry heading atthe same time. See 44 FR 49222, 49224-49225.4 We, therefore,4 The Notice further states that “The Administrator may alsoconcurrently develop standards for sources which are not on the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

44. Page 44 of 604believe that the currently listed Oil and Natural Gassource category covers all operations in this industry(i.e., production, processing, transmission, storage anddistribution). To the extent there are oil and gasoperations not covered by the currently listed Oil andNatural Gas source category, pursuant to CAA section111(b), we hereby modify the category list to include alloperations in the oil and natural gas sector. Section111(b) of the CAA gives the EPA broad authority anddiscretion to list and establish NSPS for a category that,in the Administrator’s judgment, causes or contributessignificantly to air pollution which may reasonably beanticipated to endanger public health or welfare. Pursuantto CAA section 111(b), we are modifying the source categorylist to include any oil and gas operation not covered bythe current listing and evaluating emissions from all oiland gas operations at the same time. We are also proposing standards for several new oiland natural gas affected facilities. The proposed standardswould apply to affected facilities that commenceconstruction, reconstruction or modification after [INSERTpriority list.” 44 FR at 49225. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

45. Page 45 of 604DATE OF PUBLICATION IN THE FEDERAL REGISTER]. Thesestandards, which include requirements for VOC, would becontained in a new subpart, 40 CFR part 60, subpart OOOO.Subpart OOOO would incorporate 40 CFR part 60, subpart KKKand 40 CFR part 60, subpart LLL, thereby having in this onesubpart, all standards that are applicable to the new andmodified affected facilities described above. We alsopropose to amend the title of subparts KKK and LLL,accordingly, to apply only to affected facilities alreadysubject to those subparts. Those operations would notbecome subject to subpart OOOO unless they triggeredapplicability based on new or modified affected facilitiesunder subpart OOOO. We are proposing operational standards for completionsof hydraulically fractured gas wells. Based on our review,we identified two subcategories of fractured gas wells forwhich well completions are conducted. For non-exploratoryand non-delineation wells, the proposed operationalstandards would require reduced emission completion (REC),commonly referred to as “green completion,” in combinationwith pit-flaring of gas not suitable for entering thegathering line. For exploratory and delineation wells(these wells generally are not in close proximity to a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

46. Page 46 of 604gathering line), we proposed an operational standard thatwould require pit flaring. Well completions subject to thestandards would be limited to gas well completionsfollowing hydraulic fracturing operations. Thesecompletions include those conducted at newly drilled andfractured wells, as well as completions conducted followingrefracturing operations at various times over the life ofthe well. We have determined that a completion associatedwith refracturing performed at an existing well (i.e., awell existing prior to (INSERT DATE OF PUBLICATION IN THEFEDERAL REGISTER)) is considered a modification under CAAsection 111(a), because physical change occurs to theexisting well resulting in emissions increase during therefracturing and completion operation. A detaileddiscussion of this determination is presented in theTechnical Support Document (TSD) in the docket. Therefore,the proposed standards would apply to completions at newgas wells that are fractured or refractured along withcompletions associated with fracturing or refracturing ofexisting gas wells. The modification determination andresultant applicability of NSPS to the completion operationfollowing fracturing or refracturing of existing gas wells(i.e., wells existing before [INSERT DATE OF PUBLICATION IN This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

47. Page 47 of 604THE FEDERAL REGISTER] would be limited strictly to thewellhead, well bore, casing and tubing, and any conveyancethrough which gas is vented to the atmosphere and not beextended beyond the wellhead to other ancillary componentsthat may be at the well site such as existing storagevessels, process vessels, separators, dehydrators or anyother components or apparatus. We are also proposing VOC standards to reduceemissions from gas-driven pneumatic devices. We areproposing that each pneumatic device is an affectedfacility. Accordingly, the proposed standards would applyto each newly installed pneumatic device (includingreplacement of an existing device). At gas processingplants, we are proposing a zero emission limit for eachindividual pneumatic controller. The proposed emissionstandards would reflect the emission level achievable fromthe use of non-gas-driven pneumatic controllers. At otherlocations, we are proposing a bleed limit of 6 standardcubic feet of gas per hour for an individual pneumaticcontroller, which would reflect the emission levelachievable from the use of low bleed gas-driven pneumaticcontrollers. In both cases, the standards provide This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

48. Page 48 of 604exemptions for certain applications based on functionalconsiderations. In addition, the proposed rule would require measuresto reduce VOC emissions from centrifugal and reciprocatingcompressors. As explained in more detail below in sectionVI.B.4, we are proposing equipment standards forcentrifugal compressors. The proposed standards wouldrequire the use of dry seal systems. However, we are awarethat some owners and operators may need to use centrifugalcompressors with wet seals, and we are soliciting commenton the suitability of a compliance option allowing the useof wet seals combined with routing of emissions from theseal liquid through a closed vent system to a controldevice as an acceptable alternative to installing dryseals. Our review of reciprocating compressors found thatpiston rod packing wear produces fugitive emissions thatcannot be captured and conveyed to a control device. As aresult, we are proposing operational standards forreciprocating compressors, such that the proposed rulewould require replacement of the rod packing based on hoursof usage. The owner or operator of a reciprocatingcompressor affected facility would be required to monitor This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

49. Page 49 of 604the duration (in hours) that the compressor is operated.When the hours of operation reaches 26,000 hours, the owneror operator would be required to change the rod packingimmediately. However, to avoid unscheduled shutdowns when26,000 hours is reached, owners and operators could trackhours of operation such that packing replacement could becoordinated with planned maintenance shutdowns before hoursof operation reached 26,000. Some operators may prefer toreplace the rod packing on a fixed schedule to ensure thatthe hours of operation would not reach 26,000 hours. Wesolicit comment on the appropriateness of a fixedreplacement frequency and other considerations that wouldbe associated with regular replacement. We are also proposing VOC standards for new ormodified storage vessels. The proposed rule, which wouldapply to individual vessels, would require that vesselsmeeting certain specifications achieve at least 95-percentreduction in VOC emissions. Requirements would apply tovessels with a throughput of 1 barrel of condensate per dayor 20 barrels of crude oil per day. These thresholds areequivalent to VOC emissions of about 6 tpy. For gas processing plants, we are updating therequirements for leak detection and repair (LDAR) to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

50. Page 50 of 604reflect procedures and leak thresholds established by 40CFR 60, subpart VVa. The existing NSPS requires 40 CFR part60, subpart VV procedures and thresholds. For 40 CFR part 60, subpart LLL, which regulates SO2emissions from natural gas processing plants, we determinedthat affected facilities with sulfur feed rate of at least5 long tons per day or H2S concentration in the acid gasstream of at least 50 percent can achieve up to 99.9-percent SO2 control, which is greater than the existingstandard. Therefore, we are proposing revision to theperformance standards in subpart LLL as a result of thisreview. For a more detailed discussion of this proposeddetermination, please see section VI.B.1 of this preamble. We are proposing to address compliance requirementsfor periods of startup, shutdown and malfunction (SSM) for40 CFR part 60, subpart OOOO. The SSM changes are discussedin detail in section VI.B.5 below. In addition, we areproposing to incorporate the requirements in 40 CFR part60, subpart KKK and 40 CFR part 60, subpart LLL into thenew subpart OOOO so that all requirements applicable to thenew and modified facilities would be in one subpart. Thiswould simplify and streamline compliance efforts on the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

51. Page 51 of 604part of the oil and natural gas industry and could minimizeduplication of notification, recordkeeping and reporting.B. What are the proposed decisions and actions related tothe NESHAP? This section summarizes the results of our RTR for theOil and Natural Gas Production and the Natural GasTransmission and Storage source categories and our proposeddecisions concerning these two 1999 NESHAP.1. Addressing Unregulated Emissions Sources Pursuant to CAA sections 112(d)(2) and (3), we areproposing MACT standards for subcategories of glycoldehydrators for which standards were not previouslydeveloped (hereinafter referred to as the “smalldehydrators”). In the Oil and Natural Gas Production sourcecategory, the subcategory consists of glycol dehydratorswith an actual annual average natural gas flowrate lessthan 85,000 standard cubic meters per day (scmd) or actualaverage benzene emissions less than 0.9 megagrams per year(Mg/yr). In the Natural Gas Transmission and Storage sourcecategory, the subcategory consists of glycol dehydratorswith an actual annual average natural gas flowrate lessthan 283,000 scmd or actual average benzene emissions lessthan 0.9 Mg/yr. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

52. Page 52 of 604 The proposed MACT standards for the subcategory ofsmall dehydrators at oil and gas production facilitieswould require that existing affected sources meet a unit-specific BTEX limit of 1.10x10-4 grams BTEX/standard cubicmeters (scm)-parts per million by volume (ppmv) and thatnew affected sources meet a BTEX limit of 4.66 x10-6 gramsBTEX/scm-ppmv. At natural gas transmission and storageaffected sources, the proposed MACT standard for thesubcategory of small dehydrators would require thatexisting affected sources meet a unit-specific BTEXemission limit of 6.42x10-5 grams BTEX/scm-ppmv and that newaffected sources meet a BTEX limit of 1.10 x10-5 gramsBTEX/scm-ppmv. We are also proposing MACT standards for storagevessels that are currently not regulated under the Oil andNatural Gas Production NESHAP. The current MACT standardsapply only to storage vessels with the potential for flashemissions (PFE). As explained in section VII, the originalMACT analysis accounted for all storage vessels. We are,therefore, proposing to apply the current MACT standards of95-percent emission reduction to every storage vessel atmajor source oil and natural gas production facilities. Inconjunction with this change, we are proposing to amend the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

53. Page 53 of 604definition of associated equipment to exclude all storagevessels, and not just those with the PFE, from beingconsidered “associated equipment.” This means thatemissions from all storage vessels, and not just those fromstorage vessels with the PFE, are to be included in themajor source determination.2. What are the proposed decisions and actions related tothe risk review? For both the Oil and Natural Gas Production and theNatural Gas Transmission and Storage source categories, wefind that the current levels of emissions allowed by theMACT reflect acceptable levels of risk; however, the levelof emissions allowed by the alternative compliance optionfor glycol dehydrator MACT (i.e., the option of reducingbenzene emissions to less than 0.9 Mg/yr in lieu of theMACT standard of 95-percent control) reflects anunacceptable level of risk. We are, therefore, proposing toeliminate the 0.9 Mg/yr alternative compliance option. In addition, we are proposing that the MACT for thesetwo oil and gas source categories, as revised per above,provide an ample margin of safety to protect public healthand prevent adverse environmental effects.3. What are the proposed decisions and actions related to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

54. Page 54 of 604the technology reviews of the existing NESHAP? For both the Oil and Natural Gas Production and theNatural Gas Transmission and Storage source categories, weare proposing no revisions to the existing NESHAP pursuantto section 112(d)(6) of the CAA.4. What other actions are we proposing? We are proposing an alternative performance test fornon-flare, combustion control devices. This test is to beconducted by the combustion control device manufacturer todemonstrate the destruction efficiency achieved by aspecific model of combustion control device. This wouldallow a source to purchase a performance tested device forinstallation at their site without being required toconduct a site-specific performance test. A definition for“flare” is being proposed in the NESHAP to clarify whichcombustion control devices fall under the manufacturers’performance testing alternative, and to clarify whichdevices must be performance tested. We are also proposing to: Revise the parametricmonitoring calibration provisions; require periodicperformance testing where applicable; remove the allowanceof a design analysis for all control devices other thancondensers; remove the requirement for a minimum residence This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

55. Page 55 of 604time for an enclosed combustion device; and addrecordkeeping and reporting requirements to document carbonreplacement intervals. These changes are being proposed tobring the NESHAP up-to-date based on what we have learnedregarding control devices and compliance since the originalpromulgation date. In addition, we are proposing the elimination of theSSM exemption in the Oil and Natural Gas Production and theNatural Gas Transmission and Storage NESHAP. As discussedin more detail below in section VII, consistent with SierraClub v. EPA, 551 F.3d 1019 (D.C. Cir. 2010), the EPA isproposing that the established standards in these twoNESHAP apply at all times. We are proposing to revise Table2 to both 40 CFR part 63, subpart HH and 40 CFR part 63,subpart HHH to indicate that certain 40 CFR part 63 generalprovisions relative to SSM do not apply, including: 40 CFR63.6 (e)(1)(i)5 and (ii), 40 CFR 63.6(e)(3) (SSM planrequirement), 40 CFR 63.6(f)(1); 40 CFR 63.7(e)(1), 40 CFR63.8(c)(1)(i) and (iii), and the last sentence of 40 CFR5 40 CFR 63.6(e)(1)(i) requires owners or operators to actaccording to the general duty to “operate and maintain anyaffected source, including associated air pollution controlequipment and monitoring equipment, in a manner consistent withsafety and good air pollution control practices for minimizingemissions.” This general duty to minimize is included in ourproposed standard at 40 CFR 63.783(b)(1). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

56. Page 56 of 60463.8(d)(3); 40 CFR 63.10(b)(2)(i),(ii), (iv) and (v); 40CFR 63.10(c)(10), (11) and (15); and 40 CFR 63.10(d)(5). Weare also proposing to: (1) Revise 40 CFR 63.771(d)(4)(i)and 40 CFR 63.1281(d)(4)(i) regarding operation of thecontrol device to be consistent with the SSM compliancerequirements; and (2) revise the SSM-associated reportingand recordkeeping requirements in 40 CFR 63.774, 40 CFR63.775, 40 CFR 63.1284 and 40 CFR 63.1285 to requirereporting and recordkeeping for periods of malfunction. Inaddition, as explained below, we are proposing to add anaffirmative defense to civil penalties for exceedances ofemission limits caused by malfunctions, as well as criteriafor establishing the affirmative defense. The EPA has attempted to ensure that we have neitheroverlooked nor failed to propose to remove from theexisting text any provisions that are inappropriate,unnecessary or redundant in the absence of the SSMexemption, nor included any such provisions in the proposednew regulatory language. We are specifically seekingcomment on whether there are any such provisions that wehave inadvertently overlooked or incorporated. We are also revising the applicability provisions of40 CFR part 63, subpart HH to clarify requirements This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

57. Page 57 of 604regarding PTE determination and the scope of a facilitysubject to subpart HH. Lastly, we are proposing severaleditorial corrections and plain language revisions toimprove these rules.C. What are the proposed notification, recordkeeping andreporting requirements for this proposed action?1. What are the proposed notification, recordkeeping andreporting requirements for the proposed NSPS? The proposed 40 CFR part 60, subpart OOOO includes newrequirements for several operations for which there are noexisting Federal standards. Most notably, as discussed insections V.A and VI.B of this preamble, the proposed NSPSwill cover completions and recompletions of hydraulicallyfractured gas wells. We estimate that over 20,000completions and recompletions annually will be subject tothe proposed requirements. Given the number of theseoperations, we believe that notification and reporting mustbe streamlined to the extent possible to minimize undueburden on owners and operators, as well as state, local andtribal agencies. In section V.D of this preamble, wediscuss some innovative implementation approaches beingconsidered and seek comment on these and other potential This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

58. Page 58 of 604methods of streamlining notification and reporting for wellcompletions covered by the proposed rule. Owners or operators are required to submit initialnotifications and annual reports, and to retain records toassist in documenting that they are complying with theprovisions of the NSPS. These notification, recordkeepingand reporting activities include both requirements of the40 CFR part 60 General Provisions, as well as requirementsspecific to 40 CFR part 60, subpart OOOO. Owners or operators of affected facilities (except forpneumatic controller and gas wellhead affected sources)must submit an initial notification within 1 year afterbecoming subject to 40 CFR part 60, subpart OOOO or by 1year after the publication of the final rule in the FederalRegister, whichever is later. For pneumatic controllers,owners and operators are not required to submit an initialnotification, but instead are required to report theinstallation of these affected facilities in theirfacility’s annual report. Owners or operators of wellheadaffected facilities (well completions) would also berequired to submit a 30-day advance notification of eachwell completion subject to the NSPS. In addition, annualreports are due 1 year after initial startup date for your This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

59. Page 59 of 604affected facility or 1 year after the date of publicationof the final rule in the Federal Register, whichever islater. The notification and annual reports must includeinformation on all affected facilities owned or operatedthat were new, modified or reconstructed sources during thereporting period. A single report may be submitted coveringmultiple affected facilities, provided that the reportcontains all the information required by 40 CFR 60.5420(b).This information includes general information on thefacility (i.e., company name and address, etc.), as well asinformation specific to individual affected facilities. For wellhead affected facilities, this informationincludes details of each well completion during the period,including duration of periods of gas recovery, flaring andventing. For centrifugal compressor affected facilities,information includes documentation that the compressor isfitted with dry seals. For reciprocating compressors,information includes the cumulative hours of operation ofeach compressor and records of rod packing replacement. Information for pneumatic device affected facilitiesincludes location and manufacturer specifications of eachpneumatic controller installed during the period anddocumentation that supports any exemption claimed allowing This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

60. Page 60 of 604use of high bleed controllers. For controllers installed atgas processing plants, the owner or operator would documentthe use of non-gas driven devices. For controllersinstalled in locations other than at gas processing plants,owners or operators would provide manufacturer’sspecifications that document bleed rate not exceeding 6cubic feet per hour. For storage vessel affected facilities, requiredreport information includes information that documentscontrol device compliance, if applicable. For vessels withthroughputs below 1 barrel of condensate per day and 21barrels of crude oil per day, required information alsoincludes calculations or other documentation of thethroughput. For onshore gas processing plants, semi-annualreports are required, and include information on number ofpressure relief devices, number of pressure relief devicesfor which leaks were detected and pressure relief devicesfor which leaks were not repaired, as required in 40 CFR60.5396 of subpart OOOO. Records must be retained for 5 years and generallyconsist of the same information required in the initialnotification and annual and semiannual reports.2. What are the proposed amendments to notification, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

61. Page 61 of 604recordkeeping and reporting requirements for the NESHAP? We are proposing to revise certain recordkeepingrequirements of 40 CFR part 63, subpart HH and 40 CFR part63, subpart HHH. Specifically, we are proposing thatfacilities using carbon adsorbers as a control device keeprecords of their carbon replacement schedule and recordsfor each carbon replacement. In addition, owners andoperators are required to keep records of the occurrenceand duration of each malfunction or operation of the airpollution control equipment and monitoring equipment. In addition, in conjunction with the proposed MACTstandards for small glycol dehydration units and storagevessels that do not have the PFE in the proposed amendmentto 40 CFR part 63, subpart HH, we are proposing that ownersand operators of affected small glycol dehydration unitsand storage vessels submit an initial notification within 1year after becoming subject to subpart HH or by 1 yearafter the publication of the final rule in the FederalRegister, whichever is later. Similarly, in conjunction with the proposed MACTstandards for small glycol dehydration units in theproposed 40 CFR part 63, subpart HHH amendments, we areproposing that owners and operators of small glycol This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

62. Page 62 of 604dehydration units submit an initial notification within 1year after becoming subject to subpart HHH or by 1 yearafter the publication of the final rule in the FederalRegister, whichever is later. Affected sources under either40 CFR part 63, subpart HH or subpart HHH that plan to bearea sources by the compliance dates will be required tosubmit a notification describing their schedule for theactions planned to achieve area source status. The proposed amendments to the NESHAP also includeadditional requirements for the contents of the periodicreports. For both 40 CFR part 63, subpart HH and 40 CFRpart 63, subpart HHH, we are proposing that the periodicreports also include periodic test results and informationregarding any carbon replacement events that occurredduring the reporting period.3. How is information submitted using the ElectronicReporting Tool (ERT)? Performance test data are an important source ofinformation that the EPA uses in compliance determinations,developing and reviewing standards, emission factordevelopment, annual emission rate determinations and otherpurposes. In these activities, the EPA has found itineffective and time consuming, not only for owners and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

63. Page 63 of 604operators, but also for regulatory agencies, to locate,collect and submit performance test data because of variedlocations for data storage and varied data storage methods.In recent years, though, stack testing firms have typicallycollected performance test data in electronic format,making it possible to move to an electronic data submittalsystem that would increase the ease and efficiency of datasubmittal and improve data accessibility. Through this proposal, the EPA is taking a step toincrease the ease and efficiency of data submittal andimprove data accessibility. Specifically, the EPA isproposing that owners and operators of oil and natural gassector facilities submit electronic copies of requiredperformance test reports to the EPA’s WebFIRE database. TheWebFIRE database was constructed to store performance testdata for use in developing emission factors. A descriptionof the WebFIRE database is available athttp://cfpub.epa.gov/oarweb/index.cfm?action=fire.main. As proposed above, data entry would be through anelectronic emissions test report structure called theElectronic Reporting Tool (ERT). The ERT will be able totransmit the electronic report through the EPA’s CentralData Exchange network for storage in the WebFIRE database This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

64. Page 64 of 604making submittal of data very straightforward and easy. Adescription of the ERT can be found athttp://www.epa.gov/ttn/chief/ert/ert_tool.html. The proposal to submit performance test dataelectronically to the EPA would apply only to thoseperformance tests conducted using test methods that will besupported by the ERT. The ERT contains a specificelectronic data entry form for most of the commonly usedEPA reference methods. A listing of the pollutants and testmethods supported by the ERT is available athttp://www.epa.gov/ttn/chief/ert/ert_tool.html. We believethat industry would benefit from this proposed approach toelectronic data submittal. Having these data, the EPA wouldbe able to develop improved emission factors, make fewerinformation requests, and promulgate better regulations. One major advantage of the proposed submittal ofperformance test data through the ERT is a standardizedmethod to compile and store much of the documentationrequired to be reported by this rule. Another advantage isthat the ERT clearly states testing information that wouldbe required. Another important benefit of submitting thesedata to the EPA at the time the source test is conducted isthat it should substantially reduce the effort involved in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

65. Page 65 of 604data collection activities in the future. When the EPA hasperformance test data in hand, there will likely be feweror less substantial data collection requests in conjunctionwith prospective required residual risk assessments ortechnology reviews. This would result in a reduced burdenon both affected facilities (in terms of reduced manpowerto respond to data collection requests) and the EPA (interms of preparing and distributing data collectionrequests and assessing the results). State, local and tribal agencies could also benefitfrom more streamlined and accurate review of electronicdata submitted to them. The ERT would allow for anelectronic review process rather than a manual dataassessment making review and evaluation of the sourceprovided data and calculations easier and more efficient.Finally, another benefit of the proposed data submittal toWebFIRE electronically is that these data would greatlyimprove the overall quality of existing and new emissionsfactors by supplementing the pool of emissions test datafor establishing emissions factors and by ensuring that thefactors are more representative of current industryoperational procedures. A common complaint heard fromindustry and regulators is that emission factors are This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

66. Page 66 of 604outdated or not representative of a particular sourcecategory. With timely receipt and incorporation of datafrom most performance tests, the EPA would be able toensure that emission factors, when updated, represent themost current range of operational practices. In summary, inaddition to supporting regulation development, controlstrategy development and other air pollution controlactivities having an electronic database populated withperformance test data would save industry, state, local,tribal agencies and the EPA significant time, money andeffort while also improving the quality of emissioninventories and, as a result, air quality regulations.D. What are the innovative compliance approaches beingconsidered? Given the potential number and diversity of sourcesaffected by this action, we are exploring optionalapproaches to provide the regulated community, theregulators and the public a more effective mechanism thatmaximizes compliance and transparency while minimizingburden. Under a traditional approach, owners or operatorswould provide notifications and keep records of informationrequired by the NSPS. In addition, they would certify This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

67. Page 67 of 604compliance with the NSPS as part of a required annualreport that would include compliance-related information,such as details of each well completion event andinformation documenting compliance with other requirementsof the NSPS. The EPA, state or local agency would thenphysically inspect the affected facilities and/or audit therecords retained by the owner or operator. As analternative to the traditional approach, we are seeking aninnovative way to provide for more transparency to thepublic and less burden on the regulatory agencies andowners and operators, especially as it relates tomodification of existing sources through recompletions ofhydraulically fractured gas wells. These innovativeapproaches would provide compliance assurance in light ofthe absence of requirements for CAA title V permitting ofnon-major sources. Section V.E of this preamble discusses permittingimplications associated with the NSPS and presents aproposed rationale for exempting non-major sources subjectto the NSPS from title V permitting requirements. Asdiscussed in sections V.A, V.C and VI.B of this preamble,the proposed NSPS will cover completions and recompletionsof hydraulically fractured gas wells. We estimate that over This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

68. Page 68 of 60420,000 completions and recompletions annually will besubject to the proposed requirements. As a result, webelieve that notification and reporting associated withwell completions must be streamlined to the extent possibleto minimize undue burden on owners and operators, as wellas state, local and tribal agencies. Though therequirements being proposed here are based on thetraditional approach to compliance and do not includespecific regulatory provisions for innovative compliancetools, we have included discussions below that describe howsome of these optional tools could work, and we willconsider providing for such options in the final action.Further, we request comments and suggestions on all aspectsof the innovative compliance approaches discussed below andhow they may be implemented appropriately. We are seekingcomment regarding the scope of application of one or moreof these approaches, i.e., which provisions of thestandards being proposed here would be suitable forspecific compliance approaches, and whether the approachesshould be alternatives to the requirements in theregulations. The guiding principles we are following in consideringthese approaches to compliance are: (1) Simplicity and ease This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

69. Page 69 of 604of understanding and implementation; (2) transparency andpublic accessibility; (3) electronic implementation whereappropriate; and (4) encouragement of compliance by makingcompliance easier than noncompliance. Below are some toolsthat, when used in tandem with emissions limits andoperational standards, the Agency believes could bothassure compliance and transparency, while minimizing burdenon affected sources and regulatory agencies.1. Registration of Wells and Advance Notification ofPlanned Completions Although the proposed NSPS will not require approvalto drill or complete wells, it is important that regulatoryagencies know when completions of hydraulically fracturedwells are to be performed. Notification should occursufficiently in advance to allow for inspections or auditsto certify or verify that the operator will have in placeand use the appropriate controls during the completion. Tothat end, the proposed NSPS requires a 30-day advancenotification of each completion or recompletion of ahydraulically fractured gas well. The advance notificationwould require that owners or operators provide theanticipated date of the completion, the geographiccoordinates of the well and identifying information This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

70. Page 70 of 604concerning the owner or operator and responsible companyofficial. We believe this notification requirement servesas the registration requirement and could be streamlinedthrough optional electronic reporting with web-based publicaccess or other methods. We seek comment on potentialmethodologies that would minimize burden on operators,while providing timely and useful information forregulators and the public. We also solicit comment onprovisions for a follow-up notification one or two daysbefore an impending completion via telephone or byelectronic means, since it is difficult to predict exactlywhen a well will be ready for completion a month inadvance. However, we would expect an owner or operator toprovide the follow-up notification only in cases where thecompletion date was expected to deviate from the originaldate provided. We ask for suggestions regarding how muchadvance notification is needed and the most effectivemethod of providing sufficient and accurate advancenotification of well completions.2. Third Party Verification To complement the annual compliance certificationrequired under the proposed NSPS, we are considering andseeking comment on the potential use of third party This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

71. Page 71 of 604verification to assure compliance. Since the emissionsources in the oil and natural gas sector, especially wellcompletions, are widely geographically dispersed (often invery remote locations), compliance assurance can be verydifficult and burdensome for state, local and tribalagencies and EPA permitting staff, inspectors andcompliance officers. Additionally, we believe thatverification of the data collection, compilation andcalculations by an independent and impartial third partycould facilitate the demonstration of compliance for thepublic. Verification of emissions data can also bebeneficial to owners and operators by providing certaintyof compliance status. As mentioned above, notification and reportingrequirements associated with well completions are likelyapplications for third party verification used in tandemwith the required annual compliance certification. Thethird party verification program could be used in a varietyof ways to ease regulatory burden on the owners andoperators and to leverage compliance assurance efforts ofthe EPA and state, local and tribal agencies. The thirdparty agent could serve as a clearinghouse fornotifications, records and annual compliance certifications This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

72. Page 72 of 604submitted by owners and operators. This would provideonline access to completion information by regulatoryagencies and the public. Having notifications submitted tothe clearinghouse would relieve state, local and tribalagencies of the burden of receiving thousands of paper oremail well completion notifications each year, yet stillprovide them quick access to the information. Using a thirdparty agent, it is possible that notifications of wellcompletions could be submitted with an advance period muchless than 30 days that could make a 2 day follow-upnotification unnecessary. The clearinghouse could alsohouse information on past completions and copies ofcompliance certifications. We seek comment on whetherannual reports for well completions would be needed if asuitable third party verification program was in place andalready housed that same information. We also solicitcomment on the range of potential activities the thirdparty verification program could handle with regard to wellcompletions. In this proposed action, there are also provisions forapplying third party verification to the requiredelectronic reporting using the ERT (see section V.C.3 abovefor a discussion of the ERT). As stated above, all sources This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

73. Page 73 of 604must use the ERT to submit all performance test reports(required in 40 CFR parts 60, 61 and 63) to the EPA. Thereis an option in the ERT for state, local and tribalagencies to review and verify that the informationsubmitted to the EPA is truthful, accurate and complete.Third party verifiers could be contractors or otherpersonnel familiar with oil and natural gas exploration andproduction. We are seeking comment on appropriate thirdparty reviewers and qualifications and registrationrequirements under such a program. We want to state clearlyhere that third party verification would not supersede orsubstitute for inspections or audit of data and informationby state, local and tribal agencies and the EPA. Potential issues with third party verification includecosts incurred by industry and approval of third partyverifiers. The cost of third party verification would beborne by the affected industries. We are seeking comment onwhether third party verification paid for by industry wouldresult in impartial, accurate and complete datainformation. The EPA, working with state, local and tribalagencies and industry, would expect to develop guidance forthird party verifiers. We are seeking comment on whether ornot the EPA should approve third party verifiers. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

74. Page 74 of 6043. Electronic Reporting Using Existing Mechanisms The proposed 40 CFR part 60, subpart OOOO and finalGreenhouse Gas (GHG) Mandatory Reporting Rule, 40 CFR part98, subpart W, provide details on flare and vented emissionsources and how to estimate their emissions. We solicitcomment on requiring sources to electronically submit theiremissions data for the oil and gas rules proposed here. TheEPA’s Electronic Greenhouse Gas Reporting Tool (e-GGRT) for40 CFR part 98, subpart W, while used to report emissionsat the emissions source level (e.g., well completions, wellunloading, compressors, gas plant leaks, etc.), willaggregate emissions at the basin level for e-reportingpurposes. As a result, it may be difficult to mergereporting under NSPS subpart OOOO with GHG Reporting Rulesubpart W methane reporting, especially if manual reportingis used. However, since the operator would have theseemissions details at the individual well level (becausethat will be how they would develop their basin-wideestimates), we do not believe it would be a significantburden to require owners or operators to report the datathey already have for subpart W in an ERT for NSPS andNESHAP compliance purposes. However, if the e-GGRT is notstructured to provide for reporting of other pollutants This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

75. Page 75 of 604besides GHG (e.g., VOC and HAP), then there may be somemodification of the database required to accommodate theother pollutants.4. Provisions for Encouraging Innovative Technology The oil and natural gas industry has a long history ofinnovation in developing new exploration and productionmethods, along with techniques to minimize product lossesand reduce adverse environmental impacts. These efforts areoften undertaken with tremendous amounts of research,including pilot applications at operating facilities in thefield. Absent regulation, these developmental activities,some of which ultimately are not successful, can proceedwithout risk of violation of any standards. However, asmore emission sources in this source category are coveredby regulation, as in the case of the action being proposedhere, there likely will be situations where innovation anddevelopment of new control techniques potentially could bestifled by risk of violation. We believe it is important to facilitate, not hinder,innovation and continued development of new technology thatcan result in enhanced environmental performance offacilities and sources affected by the EPA’s regulations.However, any approaches to accommodate technology This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

76. Page 76 of 604development must be designed and implemented in accordancewith the CAA and other statutes. We seek comment onapproaches that may be suitable for allowing temporaryfield testing of technology in development. Theseapproaches could include not only established proceduresunder the CAA and its implementing regulations, but newways to apply or interpret these provisions to avoidimpeding innovation while remaining environmentallyresponsible and legal.E. How does the NSPS Relate to Permitting of Sources?1. How does this action affect permitting requirements? The proposed rules do not change the Federalrequirements for determining whether oil and gas sourcesare major sources for purposes of nonattainment major NewSource Review (NSR), prevention of significantdeterioration, CAA title V, or HAP major sources pursuantto CAA section 112. Specifically, if an owner or operatoris not currently required to get a major NSR or title Vpermit for oil and gas sources, including well completions,it would not be required to get a major NSR or title Vpermit as a result of these proposed standards. EPA-approved state and local major source permitting programswould not be affected. That is, state and local agencies This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

77. Page 77 of 604with EPA-approved programs will still make case-by-casemajor source determinations for purposes of major NSR andtitle V, relying on the regulatory criteria, as explainedin the McCarthy Memo.6 Consistent with the McCarthy Memo,whether or not a permitting authority should aggregate twoor more pollutant-emitting activities into a single majorstationary source for purposes of NSR and title V remains acase-by-case decision in which permitting authoritiesretain the discretion to consider the factors relevant tothe specific circumstances of the permitted activities. In addition, the proposed standards would not changethe requirements for determining whether oil and gassources are subject to minor NSR. Nor would the proposedstandards affect existing EPA-approved state and localminor NSR rules, as well as policies and practicesimplementing those rules. Many state and local agencieshave already adopted minor NSR permitting programs thatprovide for control of emissions from relatively smallemission sources, including various pieces of equipmentused in oil and gas fields. State and local agencies would6 Withdrawal of Source Determinations for Oil and Gas Industries,September 22, 2009. This memo continues to articulate theAgency’s interpretation for major NSR and title V permitting ofoil and gas sources. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

78. Page 78 of 604be able to continue to use any EPA-approved GeneralPermits, Permits by Rule, and other similar streamliningmechanisms to permit oil and gas sources such as wells. Werecently promulgated the final Tribal Minor NSR rules foruse in issuing minor issue permits on tribal lands, wheremany oil and gas sources are located. The proposed standards will lead to better control ofand reduced emissions from oil and gas production, gasprocessing and transmission and storage, including wells.In some instances, we anticipate that complying with theNSPS would reduce emissions from these smaller sources tobelow the minor source applicability thresholds. In thosecases, sources that would otherwise have been subject tominor NSR would not need to get minor NSR permits as aresult of being subject to the NSPS. Accordingly, thenumber of minor NSR permits, as well as the Agencyresources needed to issue them, would be reduced. We expect the emission reductions achieved from theproposed standards to significantly improve ozonenonattainment problems in areas where oil and gasproduction occurs. Strategies for attaining and maintainingthe national ambient air quality standards (NAAQS) are afunction of SIP (or, in some instances, Federal This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

79. Page 79 of 604Implementation Plans and Tribal Implementation Plans)pursuant to CAA section 110. In developing plans to attainand maintain the NAAQS, EPA works with state, local orTribal agencies to account for growth and develop overallcontrol strategies that address existing and expectedemissions. The reductions achieved by the standards willmake it easier for state and local agencies to plan for andto attain and maintain the ozone NAAQS.2. How does this action affect applicability of CAA titleV? Under section 502(a) of the CAA, the EPA may exemptone or more non-major sources7 subject to CAA section 111(NSPS) standards from the requirements of title V if theEPA finds that compliance with such requirements is"impracticable, infeasible, or unnecessarily burdensome" onsuch sources. The EPA determine whether to exempt a non-major source from title V at the time we issue the relevantCAA section 111 standards (40 CFR 70.3(b)(2)). We areproposing in this action to exempt from the requirements oftitle V non-major sources that would be subject to theproposed NSPS for well completions, pneumatic devices,7 CAA section 502(a) prohibits title V exemption for any majorsource, which is defined in CAA section 501(2) and 40 CFR 70.2. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

80. Page 80 of 604compressors, and/or storage vessels. These non-majorsources (hereinafter referred to as the “oil and gas NSPSnon-major sources”) would not be required to obtain title Vpermits solely as a result of being subject to one or moreof the proposed NSPS identified above (hereinafter referredto as the “proposed NSPS”); however, if they were otherwiserequired to obtain title V permits, such requirement(s)would not be affected by the proposed exemption. Consistent with the statute, the EPA believes thatcompliance with title V permitting is "unnecessarilyburdensome" for the oil and gas NSPS non-major sources. TheEPAs inquiry into whether this criterion was satisfied isbased primarily upon consideration of the following fourfactors: (1) Whether title V would result in significantimprovements to the compliance requirements that we areproposing for the oil and gas NSPS affected non-majorsources; (2) whether title V permitting would impose asignificant burden on these non-major sources and whetherthat burden would be aggravated by any difficulty thesesources may have in obtaining assistance from permittingagencies; (3) whether the costs of title V permitting forthese non-major sources would be justified, taking intoconsideration any potential gains in compliance likely to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

81. Page 81 of 604occur for such sources; and (4) whether there areimplementation and enforcement programs in place that aresufficient to assure compliance with the proposed Oil andNatural Gas NSPS without relying on title V permits. Notall of the four factors must weigh in favor of anexemption. See 70 FR 75320, 75323 (Title V Exemption Rule).Instead, the factors are to be considered in combinationand the EPA determines whether the factors, taken together,support an exemption from title V for the oil and gas non-major sources. Additionally, consistent with the guidanceprovided by the legislative history of CAA section 502(a),8we considered whether exempting the Oil and Natural GasNSPS non-major sources would adversely affect publichealth, welfare or the environment. The first factor iswhether title V would result in significant improvements tothe compliance requirements in the proposed NSPS. A findingthat title V would not result in significant improvementsto the compliance requirements in the proposed NSPS wouldsupport a conclusion that title V permitting is8 The legislative history of section 502(a) suggests that EPAshould not grant title V exemptions where doing so wouldadversely affect public health, welfare or the environment. (SeeChafee-Baucus Statement of Senate Managers, Environment andNatural Resources Policy Division 1990 CAA Leg. Hist. 905,Compiled November 1993.) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

82. Page 82 of 604"unnecessary" for non-major sources subject to the Oil andNatural Gas Production NSPS. One way that title V may improve compliance is byrequiring monitoring (including recordkeeping designed toserve as monitoring) to assure compliance with permit termsand conditions reflecting the emission limitations andcontrol technology requirements imposed in the standard.See 40 CFR 70.6(c)(1) and 40 CFR 71.6(c)(1). The "periodicmonitoring" provisions of 40 CFR 70.6(a)(3)(i)(B) and 40CFR 71.6(a)(3)(i)(B) require new monitoring to be added tothe permit when the underlying standard does not alreadyrequire “periodic testing or instrumental ornoninstrumental monitoring (which may consist ofrecordkeeping designed to serve as monitoring)." Inaddition, title V imposes a number of recordkeeping andreporting requirements that may be important for assuringcompliance. These include requirements for a monitoringreport at least every 6 months, prompt reports ofdeviations, and an annual compliance certification. See 40CFR 70.6(a)(3) and 40 CFR 71.6(a)(3), 40 CFR 70.6(c)(1) and40 CFR 71.6(c)(1), and 40 CFR 70.6(c)(5) and 40 CFR71.6(c)(5). To determine whether title V permits would addsignificant compliance requirements to the proposed NSPS, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

83. Page 83 of 604we compared the title V monitoring, recordkeeping andreporting requirements mentioned above to thoserequirements proposed for the Oil and Natural Gas NSPSaffected facilities. For wellhead affected facilities (well completions),the proposed NSPS would require (1) 30-day advancenotification of each well completion to be performed; (2)noninstrumental monitoring, which is achieved throughdocumentation and recordkeeping of procedures followedduring each completion, including total duration of thecompletion event, amount of time gas is recovered usingreduced emission completion techniques, amount of time gasis combusted, amount of time gas is vented to theatmosphere and justification for periods when gas iscombusted or vented rather than being recovered; (3)reports of cases where well completions were not performedin compliance with the NSPS; (4) annual reports thatdocument all completions performed during the reportingperiod (a single report may be used to document multiplecompletions conducted by a single owner or operator duringthe reporting period); and (5) annual compliancecertifications submitted with the annual report. These monitoring, recordkeeping and reporting This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

84. Page 84 of 604requirements in the proposed NSPS for well completions aresufficient to ensure that the Administrator, the state,local and tribal agencies and the public are aware ofcompletion events before they are performed to provideopportunity for inspection. Sufficient documentation wouldalso be required to be retained and reported to theAdministrator to assure compliance with the NSPS for wellcompletions. In light of the above, we have determined thatadditional monitoring through title V is not needed andthat the monitoring, recordkeeping and reportingrequirements described above are sufficient to assurecompliance with the proposed requirements for wellcompletions. With respect to storage vessels, the proposed NSPSwould require 95-percent control of VOC emissions. Theproposed standard could be met by a vapor recovery unit, aflare control device or other control device. The proposedNSPS would require an initial performance test followed bycontinuous monitoring of the control device used to meetthe 95-percent control. We believe that the monitoringrequirements described above are sufficient to assurecompliance with the proposed NSPS for storage vessels and,therefore, additional monitoring through title V is not This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

85. Page 85 of 604needed. In addition to monitoring, as part of the firstfactor, we have considered the extent to which title Vcould potentially enhance compliance through recordkeepingor reporting requirements. The proposed NSPS would require(1) construction, startup and modification notifications,as required by 40 CFR 60.7(a); and (2) annual reports thatidentify all storage vessel affected facilities of theowner or operator and documentation of periods of non-compliance. The proposed NSPS would also require recordsdocumenting liquid throughput of condensate or crude oil(to determine applicability), as provided for in theproposed rule. Recordkeeping would also include records ofthe initial performance test and other information thatdocument compliance with applicable emission limit. Theserequirements are similar to those under title V. In lightof the above, we believe that the monitoring, recordkeepingand reporting requirements described above are sufficientto assure compliance with the proposed NSPS for storagevessels. For pneumatic controllers, centrifugal compressors andreciprocating compressors, the proposed NSPS are in the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

86. Page 86 of 604form of operational, work practice or equipment standards.9For each of these affected facilities, the proposed NSPSwould require: (1) Construction, startup and modificationnotifications, as required by 40 CFR 60.7(a); (2) annualreports; (3) for each pneumatic controller installed ormodified (including replacement of an existing controller),records of location and date of installation anddocumentation that each controller emits no more than theapplicable emission limit or is exempt (with rationale forthe exemption); (4) for each centrifugal compressor,records that document that each new or modified compressoris equipped with dry seals; and (5) for each new ormodified reciprocating compressor, records of rod packingreplacement, including elapsed operating hours since theprevious rod packing installation. For these other affected sources described above, theproposed NSPS provide monitoring in the form ofrecordkeeping (as described above) that would assurecompliance with the proposed operational, work practice orequipment standards. Monitoring by means other thanrecordkeeping would not be practical or appropriate for9 The proposed numeric standards for pneumatic controllersreflect the use of specific equipment (either non-gas drivendevice or low-bleed device). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

87. Page 87 of 604these standards. Records are required to ensure that thesestandards and practices are followed. We believe that themonitoring, recordkeeping and reporting requirementsdescribed above are sufficient to assure compliance withthe proposed NSPS for pneumatic controllers andcompressors. We acknowledge that title V might provide foradditional compliance requirements for these non-majorsources, but we have determined, as explained above, thatthe monitoring, recordkeeping and reporting requirements inthis proposed NSPS are sufficient to assure compliance withthe proposed standards for well completions, storagevessels, pneumatic controllers and compressors. Further,given the nature of some of the operations and the types ofthe requirements at issue, the additional compliancerequirements under title V would not significantly improvethe compliance requirements in this proposed NSPS. Forinstance, well completions occur over a very short period(generally 3 to 10 days), and the proposed NSPS forpneumatic controllers and centrifugal compressors can bemet by simply installing the equipment that meet theproposed emission limit; therefore, the semi-annualreporting requirement under title V would not improve This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

88. Page 88 of 604compliance with these proposed NSPS and, in fact, may seeminappropriate for such short term operations. For the reasons stated above, we believe that title Vwould not result in significant improvements to thecompliance requirements that are provided in this proposedNSPS. Therefore, the first factor supports a conclusionthat title V permitting is "unnecessary" for non-majorsources subject to the Oil and Natural Gas NSPS. The second factor we considered is whether title Vpermitting would impose significant burdens on the oil andnatural gas NSPS non-major sources and whether that burdenwould be aggravated by any difficulty these sources mayhave in obtaining assistance from permitting agencies.Subjecting any source to title V permitting imposes certainburdens and costs that do not exist outside of the title Vprogram. EPA estimated that the average cost of obtainingand complying with a title V permit was $65,700 per sourcefor a 5-year permit period, including fees. See InformationCollection Request (ICR) for Part 70 Operating PermitRegulations, January 2007, EPA ICR Number 1587.07. EPA doesnot have specific estimates for the burdens and costs ofpermitting the oil and gas NSPS non-major sources; however,there are certain activities associated with the 40 CFR This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

89. Page 89 of 604part 70 and 40 CFR part 71 rules. These activities aremandatory and impose burdens on the any facility subject totitle V. They include reading and understanding permitprogram regulations; obtaining and understanding permitapplication forms; answering follow-up questions frompermitting authorities after the application is submitted;reviewing and understanding the permit; collecting records;preparing and submitting monitoring reports; preparing andsubmitting prompt deviation reports, as defined by thestate, which may include a combination of written, verbaland other communications methods; collecting information,preparing and submitting the annual compliancecertification; preparing applications for permit revisionsevery 5 years; and, as needed, preparing and submittingapplications for permit revisions. In addition, althoughnot required by the permit rules, many sources obtain thecontractual services of consultants to help them understandand meet the permitting programs requirements. The ICR for40 CFR part 70 provides additional information on theoverall burdens and costs, as well as the relative burdensof each activity described here. Also, for a morecomprehensive list of requirements imposed on 40 CFR part70 sources (hence, burden on sources), see the requirements This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

90. Page 90 of 604of 40 CFR 70.3, 40 CFR 70.5, 40 CFR 70.6, and 40 CFR 70.7.The activities described above, which are quite extensiveand time consuming, would be a significant burden on thenon-major sources that would be subject to the proposedNSPS, in particular for well completion and/or pneumaticdevices, considering the short duration of a wellcompletion and the one time equipment installation of apneumatic controller for meeting the proposed NSPS.Furthermore, some of the non-major sources that would besubject to the proposed NSPS may be small entities that maylack the technical resources and, therefore, needassistance from the permitting authorities to comply withthe title V permitting requirements. Based on ourprojections, over 20,000 well completions (for both newhydraulically fractured gas wells and for existing gaswells that are subsequently fractured or re-fractured) willbe performed each year. For pneumatic controller affectedfacilities, we estimate that approximately 14,000 newcontrollers would be subject to the NSPS each year. Ourestimated numbers of affected facilities that would besubject to the proposed NSPS for storage vessels andcompressors are smaller (around 500 compressors and 300storage vessels). Although we do not know the total number This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

91. Page 91 of 604of non-major sources that would be subject to the proposedNSPS, based on the estimated numbers of affectedfacilities, we anticipate a significant increase in thenumber of permit applications that permitting authoritieswould have to process each year. This significant burden onthe permitting authorities raises a concern with thepotential difficulty or delay that the small entities mayface in obtaining sufficient assistance from the permittingauthorities. The third factor we considered is whether the costs oftitle V permitting for these area sources would bejustified, taking into consideration any potential gains incompliance likely to occur for such sources. We concluded,in considering the first factor, that the monitoring,recordkeeping and reporting requirements in this proposedNSPS assure compliance with the proposed standards, thattitle V would not result in significant improvement tothese compliance requirements and, that, in some instances,certain title V compliance requirements may not beappropriate. In addition, as discussed above in ourconsideration of the second factor, we have concerns withthe potential burdens that title V may impose on thesesources. In addition, below in our consideration of the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

92. Page 92 of 604fourth factor, we find that there are adequateimplementation and enforcement programs in place to assurecompliance with the proposed NSPS. In light of the above,we find that the costs of title V permitting are notjustified for the sources we propose to exempt.Accordingly, the third factor supports title V exemptionfor the oil and gas NSPS non-major sources. The fourth factor we considered is whether there areimplementation and enforcement programs in place that aresufficient to assure compliance with the proposed NSPS foroil and gas sources without relying on title V permits.The CAA provides States the opportunity to take delegationof NSPS. Before the EPA will delegate the program, the EPAwill evaluate the state programs to ensure that states haveadequate capability to enforce the CAA section 111regulations and provide assurances that they will enforcethe NSPS. In addition, EPA retains authority to enforcethis NSPS anytime under CAA sections 111, 113 and 114.Accordingly, we can enforce the monitoring, recordkeepingand reporting requirements, which, as discussed under thefirst factor, are adequate to assure compliance with thisNSPS. Also, states and the EPA often conduct voluntarycompliance assistance, outreach and education programs This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

93. Page 93 of 604(compliance assistance programs), which are not required bystatute. We determined that these additional programs willsupplement and enhance the success of compliance with theseproposed standards. We believe that the statutoryrequirements for implementation and enforcement of thisNSPS by the delegated states, the EPA and the additionalassistance programs described above together are sufficientto assure compliance with these proposed standards withoutrelying on title V permitting. Our balance of the four factors strongly supports afinding that title V is unnecessarily burdensome for theoil and gas non-major sources. While title V might addadditional compliance requirements if imposed, we believethat there would not be significant improvements to thecompliance requirements in this proposed rule because theproposed rule requirements are specifically designed toassure compliance with the proposed NSPS and, as explainedabove, some of the title V requirements may not beappropriate for certain operations and/or proposedstandards. We are also concerned with the potential burdenthat title V may impose on some of these sources. In lightof little or no potential gain in compliance if title Vwere required, we do not believe that the costs of title V This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

94. Page 94 of 604permitting is justified in this case. Finally, there areadequate implementation and enforcement programs in placeto assure compliance with these proposed standards. Thus,we propose that title V permitting is "unnecessarilyburdensome" for the oil and gas non-major sources. In addition to evaluating whether compliance withtitle V requirements is "unnecessarily burdensome," EPAalso considered, consistent with guidance provided by thelegislative history of section 502(a), whether exemptingoil and gas NSPS non-major sources from title Vrequirements would adversely affect public health, welfareor the environment. The title V permit program does notimpose new substantive air quality control requirements onsources, but instead requires that certain proceduralmeasures be followed, particularly with respect todetermining compliance with applicable requirements. Asstated in our consideration of factor one, title V wouldnot lead to significant improvements in the compliancerequirements for the proposed NSPS. For the reason statedabove, we believe that exempting these non-major sourcesfrom title V permitting requirements would not adverselyaffect public health, welfare or the environment. On the contrary, we are concerned that requiring title This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

95. Page 95 of 604V in this case could potentially adversely affect publichealth, welfare or the environment. As mentioned above, weanticipate a significant increase in the number of permitapplications that permitting authorities would have toprocess each year. Depending on the number of non-majorsources that would be subject to this rule, requiringpermits for those sources, at least in the first few yearsof implementation, could potentially adversely affectpublic health, welfare or the environment by shifting stateagencies resources away from assuring compliance for majorsources (which cannot be exempt from title V) to issuingnew permits for these non-major sources, potentiallyreducing overall air program effectiveness. Based on the above analysis, we conclude that title Vpermitting would be "unnecessarily burdensome" for oil andgas NSPS non-major sources. We are, therefore, proposingthat these non-major sources be exempt from title Vpermitting requirements.VI. Rationale for Proposed Action for NSPSA. What did we evaluate relative to NSPS? As noted above, there are two existing NSPS thataddress emissions from the Oil and Natural Gas sourcecategory. These NSPS are relatively narrow in scope, as This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

96. Page 96 of 604they address emissions only at natural gas processingplants. Specifically, 40 CFR part 60, subpart KKK addressesVOC emissions from leaking equipment at onshore natural gasprocessing plants and 40 CFR part 60, subpart LLL addressesSO2 emissions from natural gas processing plants. CAA section 111(b)(1)(B) requires the EPA to reviewand revise, if appropriate, NSPS standards. Accordingly, weevaluated whether the existing NSPS reflect the BSER forthe emission sources that they address. This review wasconducted by examining currently used, new and emergingcontrol systems and assessing whether they representadvances in emission reduction techniques from those uponwhich the existing NSPS are based, including advances inLDAR approaches and SO2 control at natural gas processingplants. For each new or emerging control option identified,we then evaluated emission reductions, costs, energyrequirements and nonair quality impacts, such as solidwaste generation. In this package, we have also evaluated whether therewere additional pollutants emitted by facilities in the Oiland Natural Gas source category that warrant regulation andfor which we have adequate information to promulgate This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

97. Page 97 of 604standards of performance. Finally, we have identifiedadditional processes in the Oil and Natural Gas sourcecategory for which it may be appropriate to developperformance standards. This would include processes thatemit the currently regulated pollutants, VOC and SO2, aswell as any additional pollutants for which we determinedregulation to be appropriate.B. What are the results of our evaluations and proposedactions relative to NSPS?1. Do the existing NSPS reflect the BSER for sourcescovered? Consistent with our obligations under CAA section111(b), we evaluated whether the control options reflectedin the current NSPS for the Oil and Natural Gas sourcecategory still represent BSER. To evaluate the BSER optionsfor equipment leaks, we reviewed EPA’s current LDARprograms, the Reasonably Available Control Technology(RACT)/Best Available Control Technology (BACT)/LowestAchievable Emission Rate (LAER) Clearinghouse (RBLC)database, and emerging technologies that have beenidentified by partners in the Natural Gas STAR program. The current NSPS for equipment leaks of VOC at naturalgas processing plants (40 CFR part 60, subpart KKK) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

98. Page 98 of 604requires compliance with specific provisions of 40 CFR part60, subpart VV, which is a LDAR program, based on the useof EPA Method 21 to identify equipment leaks. In additionto the subpart VV requirements, we reviewed the LDARrequirements in 40 CFR part 60, subpart VVa. This LDARprogram is considered to be more stringent than the subpartVV requirements, because it has lower component leakthreshold definitions and more frequent monitoring, incomparison to the subpart VV program. Furthermore, subpartVVa requires monitoring of connectors, while subpart VVdoes not. We also reviewed options based on optical gasimaging. As mentioned above, the currently required LDARprogram for natural gas processing plants (40 CFR part 60,subpart KKK) is based on EPA Method 21, which requires theuse of an organic vapor analyzer to monitor components andto measure the concentration of the emissions inidentifying leaks. We recognize that there have beenadvancements in the use of optical gas imaging to detectleaks from these same types of components. Theseinstruments do not yet provide a direct measure of leakconcentrations. The instruments instead provide a measureof a leak relative to an instrument specific calibration This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

99. Page 99 of 604point. Since the promulgation of 40 CFR part 60, subpartKKK (which requires Method 21 leak measurement monthly),the EPA has updated the 40 CFR part 60 General Provisionsto allow the use of advanced leak detection tools, such asoptical gas imaging and ultrasound equipment as analternative to the LDAR protocol based on Method 21 leakmeasurements (see 40 CFR 60.18(g)). The alternative workpractice allowing use of these advanced technologiesincludes a provision for conducting a Method 21-based LDARcheck of the regulated equipment annually to verify goodperformance. In our review, we evaluated 4 options in consideringBSER for VOC equipment leaks at natural gas processingplants. One option we evaluated consists of changing from a40 CFR part 60, subpart VV-level program, which is what 40CFR part 60, subpart KKK currently requires, to a 40 CFRpart 60, subpart VVa program, which applies to newsynthetic organic chemical plants after 2006. Subpart VValowers the leak definition for valves from 10,000 parts permillion (ppm) to 500 ppm, and requires the monitoring ofconnectors. In our analysis of these impacts, we estimatedthat, for a typical natural gas processing plant, theincremental cost effectiveness of changing from the current This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

100. Page 100 of 604subpart VV-level program to a subpart VVa-level programusing Method 21 is $3,352 per ton of VOC reduction. In evaluating 40 CFR part 60, subpart VVa-level LDARat processing plants, we also analyzed separately theindividual types of components (valves, connectors,pressure relief devices and open-ended lines) to determinecost effectiveness for individual components. Detaileddiscussions of these component-by-component analyses areincluded in the TSD in the docket. Cost effectivenessranged from $144 per ton of VOC (for valves) to $4,360 perton of VOC (for connectors), with no change in requirementsfor pressure relief devices and open-ended lines. Another option we evaluated for gas processing plantswas the use of optical gas imaging combined with an annualEPA Method 21 check (i.e., the alternative work practicefor monitoring equipment for leaks at 40 CFR 60.18(g)). Wehad previously determined that the VOC reduction achievedby this combination of optical gas imaging and Method 21would be equivalent to reductions achieved by the 40 CFRpart 60, subpart VVa-level program. Based on that emissionreduction level, we determined the cost effectiveness ofthis option to be $6,462 per ton of VOC reduction. Thisanalysis is based on the facility purchasing an optical gas This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

101. Page 101 of 604imaging system costing $85,000. However, we identified atleast one manufacturer who rents the optical gas imagingsystems. That manufacturer rents the optical gas imagingsystem for $3,950 per week. Using this rental cost in placeof the purchase cost, the VOC cost effectiveness of themonthly optical gas imaging combined with annual Method 21checks is $4,638 per ton of VOC reduction.10 A third optionwe evaluated consisted of monthly optical gas imagingwithout an annual Method 21 check. We estimated the annualcost of the monthly optical gas imaging LDAR program to be$76,581, based on camera purchase, or $51,999, based oncamera rental. However, because we were unable to estimatethe VOC emissions achieved by an optical imaging programalone, we were unable to estimate the cost effectiveness ofthis option. Finally, we evaluated a fourth option similar to thethird option, except that the optical gas imaging would beperformed annually rather than monthly. For this option, weestimated the annual cost to be $43,851, based on camerapurchase, or $18,479, based on camera rental.10 Because optical gas imaging is used to view several pieces ofequipment at a facility at once to survey for leaks, optionsinvolving imaging are not amenable to a component by componentanalysis. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

102. Page 102 of 604 We request comment on the applicability of a LDARprogram based solely on the use of optical gas imaging. Ofmost use to us would be information on the effectiveness ofthis and, potentially, other advanced measurementtechnologies, to detect and repair small leaks on the sameorder or smaller than specified in the 40 CFR part 60,subpart VVa equipment leak requirements and the effects ofincreased frequency of and associated leak detection,recording and repair practices. Because we could not estimate the cost effectivenessof options 3 and 4, we could not identify either of thesetwo options as BSER for reducing VOC leaks at gasprocessing plants. Because options 1 and 2 have achieveequivalent VOC reduction and are both cost effective, webelieve that both options 1 and 2 reflect BSER for LDAR fornatural gas processing plants. As mentioned above, option 1is the LDAR in 40 CFR part 60, subpart VVa and option 2 isthe alternative work practice at 40 CFR 60.18(g) and isalready available to use as an alternative to subpart VVaLDAR. Therefore, we propose that the NSPS for equipmentleaks of VOC at gas processing plants be revised to requirecompliance with the subpart VVa equipment leakrequirements. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

103. Page 103 of 604 For 40 CFR part 60, subpart LLL, we reviewed controlsystems for SO2 emissions from sweetening units located atnatural gas processing plants, including those followed bya sulfur recovery unit. Subpart LLL provides specificstandards for SO2 emission reduction efficiency, on thebasis of sulfur feed rate and the sulfur content of thenatural gas. According to available literature, the most widelyused process for converting H2S in acid gases (i.e., H2Sand CO2) separated from natural gas by a sweetening process(such as amine treating) into elemental sulfur is the Clausprocess. Sulfur recovery efficiencies are higher withhigher concentrations of H2S in the feed stream due to thethermodynamic equilibrium limitation of the Claus process.The Claus sulfur recovery unit produces elemental sulfurfrom H2S in a series of catalytic stages, recovering up to97-percent recovery of the sulfur from the acid gas fromthe sweetening process. Further, sulfur recovery isaccomplished by making process modifications or byemploying a tail gas treatment process to convert theunconverted sulfur compounds from the Claus unit. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

104. Page 104 of 604 We evaluated process modifications and tail gastreatment options when we proposed 40 CFR part 60, subpartLLL. 49 FR 2656, 2659-2660 (1984). As we explained in thepreamble to the proposed subpart LLL, control throughsulfur recovery with tail gas treatment may not always becost effective, depending on sulfur feed rate and inlet H2Sconcentrations. Therefore, other methods of increasingsulfur recovery via process modifications were evaluated.As shown in the original evaluation, the performancecapabilities and costs of each of these technologies arehighly dependent on the ratio of H2S and CO2 in the gasstream and the total quantity of sulfur in the gas streambeing treated. The most effective means of control wasselected as BSER for the different stream characteristics.As a result, separate emissions limitations were developedin the form of equations that calculate the requiredinitial and continuous emission reduction efficiency foreach plant. The equations were based on the designperformance capabilities of the technologies selected asBSER relative to the gas stream characteristics. 49 FR2656, 2663-2664 (1984). The emission limit for sulfur feedrates at or below 5 long tons per day, regardless of H2S This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

105. Page 105 of 604content, was 79 percent. For facilities with sulfur feedrates above 5 long tons per day, the emission limits rangedfrom 79 percent at an H2S content below 10 percent to 99.8percent for H2S contents at or above 50 percent. To review these emission limitations, we performed asearch of the RBLC database and state regulations. No stateregulations identified had emission limitations morestringent than 40 CFR part 60, subpart LLL. However, theRBLC database search identified two entries with SO2emission reductions of 99.9 percent. One entry is for afacility in Bakersfield, California, with a 90 long ton perday sulfur recovery unit followed by an amine-based tail-gas treating unit. The second entry is for a facility inCoden, Alabama, with a sulfur recovery unit with a sulfurfeed rate of 280 long tons per day, followed by selectivecatalytic reduction and a tail gas incinerator. However,neither of these entries contained information regardingthe H2S contents of the feed stream. Because the sulfurrecovery efficiency of these large sized plants was greaterthan 99.8 percent, we reevaluated the original data. Basedon the available cost information, it appears that a 99.9-percent efficiency is cost effective for facilities with a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

106. Page 106 of 604sulfur feed rate greater than 5 long tons per day and H2Scontent equal to or greater than 50 percent. Based on ourreview, we are proposing that the maximum initial andcontinuous efficiency for facilities with a sulfur feedrate greater than 5 long tons per day and a H2S contentequal to or greater than 50 percent be raised to 99.9percent. We are not proposing to make changes to theequations. Our search of the RBLC database did not uncoverinformation regarding costs and achievable emissionreductions to suggest that the emission limitations forfacilities with a sulfur feed rate less than 5 long tonsper day or H2S content less than 50 percent should bemodified. Therefore, we are not proposing any changes tothe emissions limitations for facilities with sulfur feedrate and H2S content less than 5 long tons per day and 50percent, respectively.2. What pollutants are being evaluated in this Oil andNatural Gas NSPS package? The two current NSPS for the Oil and Natural Gassource category address emissions of VOC and SO2. Inaddition to these pollutants, sources in this source This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

107. Page 107 of 604category also emit a variety of other pollutants, mostnotably, air toxics. As discussed elsewhere in this notice,there are NESHAP that address air toxics from the oil andnatural gas sector. In addition, processes in the Oil and Natural Gassource category emit significant amounts of methane. The1990 - 2009 U.S. GHG Inventory estimates 2009 methaneemissions from Petroleum and Natural Gas Systems (notincluding petroleum refineries) to be 251.55 MMtCO2e(million metric tons of CO2-equivalents (CO2e)).11 Theemissions estimated from well completions and recompletionsexclude a significant number of wells completed in tightsand plays, such as the Marcellus, due to availability ofdata when the 2009 Inventory was developed. The estimatein this proposal includes an adjustment for tight sandplays (being considered as a planned improvement indevelopment of the 2010 Inventory). This adjustment wouldincrease the 2009 Inventory estimate by 76.74 MMtCO2e. Thetotal methane emissions from Petroleum and Natural GasSystems, based on the 2009 Inventory, adjusted for tight11 U.S. EPA. Inventory of U.S. Greenhouse Gas Inventory andSinks. 1990 - 2009.http://www.epa.gov/climatechange/emissions/downloads10/US-GHG-Inventory-2010_ExecutiveSummary.pdf This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

108. Page 108 of 604sand plays and the Marcellus, is 328.29 MMtCO2e. Althoughthis proposed rule does not include standards forregulating the GHG emissions discussed above, we continueto assess these significant emissions and evaluateappropriate actions for addressing these concerns. Becausemany of the proposed requirements for control of VOCemissions also control methane emissions as a co-benefit,the proposed VOC standards would also achieve significantreduction of methane emissions. Significant emissions of oxides of nitrogen (NOx) alsooccur at oil and natural gas sites due to the combustion ofnatural gas in reciprocating engines and combustionturbines used to drive the compressors that move naturalgas through the system, and from combustion of natural gasin heaters and boilers. While these engines, turbines,heaters and boilers are co-located with processes in theoil and natural gas sector, they are not in the Oil andNatural Gas source category and are not being addressed inthis action. The NOx emissions from engines and turbinesare covered by the Standards of Performance for StationarySpark Internal Combustion Engines (40 CFR part 60, subpartJJJJ) and Standards of Performance for Stationary This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

109. Page 109 of 604Combustion Turbines (40 CFR part 60, subpart KKKK),respectively. An additional source of NOx emissions would be pitflaring of VOC emissions from well completions duringperiods where REC is not feasible, as would be requiredunder our proposed operational standards for wellheadaffected facilities. As discussed below in section VI.B.4(well completion), pit flaring is the only way weidentified of controlling VOC emissions during theseperiods. Because there is no way of directly measuring theNOx produced, nor is there any way of applying controlsother than minimizing flaring, we propose to allow flaringonly when REC is not feasible. We have included ourestimates of NOx formation from pit flaring in ourdiscussion of secondary impacts in section VI.B.4.3. What emission sources are being evaluated in this Oiland Natural Gas NSPS package? The current NSPS only cover emissions of VOC and SO2from one type of facility in the oil and natural gassector, which is the natural gas processing plant. This isthe only type of facility in the Oil and Natural Gas sourcecategory where we would expect SO2 to be emitted directly, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

110. Page 110 of 604although H2S contained in sour gas, when oxidized in theatmosphere or combusted in boilers and heaters in thefield, forms SO2 as a product of oxidation. These fieldboilers and heaters are not part of the Oil and Natural Gassource category and are generally too small to be regulatedby the NSPS covering boilers (i.e., they have a heat inputof less than 10 million British Thermal Units per hour).However, we may consider addressing them as part of afuture sector-based strategy for the oil and natural gassector. In addition to VOC emissions from gas processingplants, there are numerous sources of VOC throughout theoil and natural gas sector that are not addressed by thecurrent NSPS. As explained above in section V.A, pursuantto CAA section 111(b), to the extent necessary, we aremodifying the listed category to include all segments ofthe oil and natural gas industry for regulation. We arealso proposing VOC standards to cover additional processesat oil and natural gas operations. These include NSPS forVOC from gas well completions, pneumatic controllers,compressors and storage vessels. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

111. Page 111 of 604 We believe that produced water ponds are also apotentially significant source of emissions, but we haveonly limited information. We, therefore, solicit commentson produced water ponds, particularly in the followingsubject areas: (a) We are requesting comments pertaining to methodsfor calculating emissions. The State of Colorado currentlyuses a mass balance that assumes 100 percent of the VOCcontent is emitted to the atmosphere. Water9, an airemissions model, is another option that has somelimitations, including poor methanol estimation. (b) We are requesting additional information ontypical VOC content in produced water and any availablechemical analyses, including data that could help clarifyseasonal variations or differences among gas fields.Additionally, we request data that increase ourunderstanding of how changing process variables or age ofwells affect produced water output and VOC content. (c) We solicit information on the size and throughputcapacity of typical evaporation pond facilities and requestsuggestions on parameters that could be used to defineaffected facilities or affected sources. We also seekinformation on impacts of smaller evaporation pits that are This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

112. Page 112 of 604co-located with drilling operations, whether those warrantcontrol and, if so, how controls should be developed. (d) An important factor is cost of emission reductiontechnologies, including recovery credits or cost savingsrealized from recovered salable product. We are seekinginformation on these considerations as well. (e) We are also seeking information on any limitationsfor emission reduction technologies such as availability ofelectricity, waste generation and disposal and throughputand concentration constraints. (f) Finally, we solicit information on separatortechnologies that are able to improve the oil-waterseparation efficiency.4. What are the rationales for the proposed NSPS? We have provided below our rationales for the proposedBSER determinations and performance standards for a numberof VOC emission sources in the Oil and Natural Gas sourcecategory that are not covered by the existing NSPS. Ourgeneral process for evaluating systems of emissionreduction for the emission sources discussed belowincluded: (1) Identification of available control measures;(2) evaluation of these measures to determine emissionreductions achieved, associated costs, nonair environmental This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

113. Page 113 of 604impacts, energy impacts and any limitations to theirapplication; and (3) selection of the control techniquesthat represent BSER based on the information we considered. We identified the control options discussed in thispackage through our review of relevant state and localrequirements and mitigation measures developed and reportedby the EPA’s Natural Gas STAR program. The EPA’s NaturalGas STAR program has worked with industry partners since1993 to identify cost effective measures to reduceemissions of methane and other pollutants from natural gasoperations. We relied heavily on this wealth of informationin conducting this review. We also identified stateregulations, primarily in Colorado and Wyoming, whichrequire mitigation measures for some emission sources inthe Oil and Natural Gas source category.a. NSPS for Well Completions Well completion activities are a significant source ofVOC emissions, which occur when natural gas and non-methanehydrocarbons are vented to the atmosphere during flowbackof a hydraulically fractured gas well. Flowback emissionsare short-term in nature and occur over a period of severaldays following fracturing of a new well or refracturing ofan existing well. Well completions include multiple steps This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

114. Page 114 of 604after the well bore hole has reached the target depth.These steps include inserting and cementing-in well casing,perforating the casing at one or more producing horizons,and often hydraulically fracturing one or more zones in thereservoir to stimulate production. Well recompletions mayalso include hydraulic fracturing. Hydraulic fracturing isone technique for improving gas production where thereservoir rock is fractured with very high pressure fluid,typically water emulsion with a proppant (generally sand)that “props open” the fractures after fluid pressure isreduced. Emissions are a result of the backflow of thefracture fluids and reservoir gas at high volume andvelocity necessary to lift excess proppant and fluids tothe surface. This multi-phase mixture is often directed toa surface impoundment where natural gas and VOC vaporsescape to the atmosphere during the collection of water,sand and hydrocarbon liquids. As the fracture fluids aredepleted, the backflow eventually contains more volume ofnatural gas from the formation. Wells that are fracturedgenerally have great amounts of emissions because of theextended length of the flowback period required to purgethe well of the fluids and sand that are associated withthe fracturing operation. Along with the fluids and sand This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

115. Page 115 of 604from the fracturing operation, the 3 to 10-day flowbackperiod also results in emissions of natural gas and VOCthat would not occur in large quantities at oil wells or atnatural gas wells that are not fractured. Thus, we estimatethat gas well completions involving hydraulic fracturingvent substantially more VOC, approximately 200 times more,than completions not involving hydraulic fracturing.Specifically, we estimate that uncontrolled well completionemissions for a hydraulically fractured gas well areapproximately 23 tons of VOC, where emissions for aconventional gas well completion are around 0.12 tons VOC.These estimates are explained in detail in the TSDavailable in the docket. Based on our review, we believethat emissions from recompletions of previously completedwells that are fractured or refractured to stimulateproduction or to begin production from a new productionhorizon are of similar magnitude and composition asemissions from completions of new wells that have beenhydraulically fractured. EPA has based the NSPS impacts analysis on bestavailable emission data. However, we recognize that thereis uncertainty associated with our estimates. For both newcompletions and recompletions, there are a variety of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

116. Page 116 of 604factors that will determine the length of the flowbackperiod and actual volume of emissions such as the number ofzones, depth, pressure of the reservoir, gas composition,etc. This variability means there will be some wells whichemit more than the estimated emission factor and some wellsthat emit less. During our review, we examined information from theNatural Gas STAR program and the Colorado and Wyoming staterules covering well completions. We identified twosubcategories of fractured gas wells: (1) Non-exploratoryand non-delineation wells; and (2) exploratory anddelineation wells. An exploratory well is the first welldrilled to determine the presence of a producing reservoirand the well’s commercial viability. A delineation well isa well drilled to determine the boundary of a field orproducing reservoir. Because exploratory and delineationwells are generally isolated from existing producing wells,there are no gathering lines available for collection ofgas recovered during completion operations. In contrast,non-exploratory and non-delineation wells are located whereexisting, producing wells are connected to gathering linesand are, therefore, able to be connected to a gathering This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

117. Page 117 of 604line to collect recovered salable natural gas product thatwould otherwise be vented to the atmosphere or combusted. For subcategory 1, we identified “green” completion,which we refer to as REC, as an option for reducing VOCemissions during well completions. REC are performed byseparating the flowback water, sand, hydrocarbon condensateand natural gas to reduce the portion of natural gas andVOC vented to the atmosphere, while maximizing recovery ofsalable natural gas and VOC condensate. In some cases, fora portion of the completion operation, such as when CO2 ornitrogen is injected with the fracture water, initial gasproduced is not of suitable quality to introduce into thegathering line due to CO2 or nitrogen content or otherundesirable characteristic. In such cases, for a portion ofthe flowback period, gas cannot be recovered, but must beeither vented or combusted. In practice, REC are oftencombined with combustion to minimize the amount of gas andcondensate being vented. This combustion process is rathercrude, consisting of a horizontal pipe downstream of theREC equipment, fitted with a continuous ignition source anddischarging over a pit near the wellhead. Because of thenature of the flowback (i.e., with periods of water, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

118. Page 118 of 604condensate, and gas in slug flow), conveying the entireportion of this stream to a traditional flare controldevice or other control device, such as a vapor recoveryunit, is not feasible. These control devices are notdesigned to accommodate the multiphase flow consisting ofwater, sand and hydrocarbon liquids, along with the gas andvapor being controlled. Although “pit flaring” does notemploy a traditional flare control device, and is notcapable of being tested or monitored for efficiency due tothe multiphase slug flow and intermittent nature of thedischarge of gas, water and sand over the pit, it doesprovide a means of minimizing vented gas and is preferableto venting. Because of the rather large exposed flame, openpit flaring can present a fire hazard or other undesirableimpacts in some situations (e.g., dry, windy conditions,proximity to residences, etc.). As a result, we are awarethat owners and operators may not be able to pit flareunrecoverable gas safely in every case. In some cases, pitflaring may be prohibited by local ordinance. Equipment required to conduct REC may include tankage,special gas-liquid-sand separator traps and gasdehydration. Equipment costs associated with REC will varyfrom well to well. Typical well completions last between 3 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

119. Page 119 of 604and 10 days and costs of performing REC are projected to bebetween $700 and $6,500 per day, including a cost ofapproximately $3,523 per completion event for the pitflaring equipment. However, there are savings associatedwith the use of REC because the gas recovered can beincorporated into the production stream and sold. In fact,we estimate that REC will result in an overall net costsavings in many cases. The emission reductions for a hydraulically fracturedwell are estimated to be around 22 tons of VOC. Based on anaverage incremental cost of $33,237 per completion, thecost effectiveness of REC, without considering any costsavings, is around $1,516 per ton of VOC (which we havepreviously found to be cost effective on average). When thevalue of the gas recovered (approximately 150 tons ofmethane per completion) is considered, the costeffectiveness is estimated as an average net savings of $99per ton VOC reduced, using standard discount rates. Webelieve that these costs are very reasonable, given theemission reduction that would be achieved. Aside from thepotential hazards associated with pit flaring, in somecases, we did not identify any nonair environmentalimpacts, health or energy impacts associated with REC This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

120. Page 120 of 604combined with combustion. However, pit flaring wouldproduce NOx emissions. Because we believe that theseemissions cannot be controlled or measured directly due tothe open combustion process characteristic of pit flaring,we used published emission factors (EPA Emission GuidelinesAP-42)to estimate the NOx emissions for purposes ofassessing secondary impacts. For category 1 wellcompletions, we estimated that 0.02 tons of NOx are producedper event. This is based on the assumption that 5 percentof the flowback gas is combusted by the combustion device.The 1.2 tons of VOC controlled during the pit flaringportion of category 1 well completions is approximately 57times greater than the NOx produced by pit flaring. Thus,we believe that the benefit of the VOC reduction faroutweighs the secondary impact of NOx formation during pitflaring. We believe that, based on the analysis above, REC incombination with combustion is BSER for subcategory 1wells. We considered setting a numerical performancestandard for subcategory 1 wells. However, it is notpracticable to measure the emissions during pit flaring orventing because the gas is discharged over the pit along This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

121. Page 121 of 604with water and sand in multiphase slug flow. Therefore, webelieve it is not feasible to set a numerical performancestandard. Pursuant to section 111(h)(2) of the CAA, we areproposing an operational standard for subcategory 1 wellsthat would require a combination of REC and pit flaring tominimize venting of gas and condensate vapors to theatmosphere, with provisions for venting in lieu of pitflaring for situations in which pit flaring would presentsafety hazards or for periods when the flowback gas isnoncombustible due to high concentrations of nitrogen orCO2. The proposed operational standard would be accompaniedby requirements for documentation of the overall durationof the completion event, duration of recovery using REC,duration of combustion, duration of venting, and specificreasons for venting in lieu of combustion. We recognize that there is heterogeneity in welloperations and costs, and that while RECs may be cost-effective on average, they may not be for all operators.Nonetheless, EPA is proposing to require an operationalstandard rather than a performance-based standard (e.g.,requiring that some percentage of emissions be flared orcaptured), because we believe there are no feasible waysfor operators to measure emissions with enough certainty to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

122. Page 122 of 604demonstrate compliance with a performance-based standardfor REC in combination with pit flaring. The EPA requestscomment on this and seeks input on whether alternativeapproaches to requiring REC for all operators with accessto pipelines may exist that would allow operators to meet aperformance-based standard if they can demonstrate that anREC is not cost effective. We have discussed above certain situations whereunrecoverable gas would be vented because pit flaring wouldpresent a fire hazard or is infeasible because gas isnoncombustible due to high concentrations of nitrogen orCO2. We solicit comment on whether there are other suchsituations where flaring would be unsafe or infeasible, andpotential criteria that would support venting in lieu ofpit flaring. In addition, we learned that coalbed methanereservoirs may have low pressure, which would present atechnical barrier for performing a REC because the wellpressure may not be substantial enough to overcomegathering line pressure. In addition, we identified thatcoalbed methane wells often have low to almost no VOCemissions, even following the hydraulic fracturing process.We solicit comment on criteria and thresholds that could beused to exempt some well completion operations occurring in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

123. Page 123 of 604coalbed methane reservoirs from the requirements forsubcategory 1 wells. Of the 25,000 new and modified fractured gas wellscompleted each year, we estimate that approximately 3,000to 4,000 currently employ reduced emission completion. Weexpect this number to increase to over 21,000 REC annuallyas operators comply with the proposed NSPS. We estimatethat approximately 9,300 new wells and 12,000 existingwells will be fractured or refractured annually that wouldbe subject to subcategory 1 requirements under the NSPS. Webelieve that there will be a sufficient supply of RECequipment available by the time the NSPS becomes effective.However, energy availability could be affected if ashortage of REC equipment was allowed to cause delays inwell completions. We request comment on whether sufficientsupply of this equipment and personnel to operate it willbe available to accommodate the increased number of REC bythe effective date of the NSPS. We also request specificestimates of how much time would be required to get enoughequipment in operation to accommodate the full number ofREC performed annually. In the event that public comments indicate thatavailable equipment would likely be insufficient to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

124. Page 124 of 604accommodate the increase in number of REC performed, we areconsidering phasing in requirements for well completionsthat would achieve an overall comparable level ofenvironmental benefit. For example, operators performingcompletions of fractured or refractured existing wells(i.e., modified wells) could be allowed to controlemissions through pit flaring instead of REC for someperiod of time. After some date certain, all modified wellswould be subject to REC. We solicit comment on the phasingof requirements for REC along with suggestions for otherways to address a potential short-term REC equipmentshortage that may hinder operators’ compliance with theproposed NSPS, while also achieving a comparable level ofreduced emissions to the air. Although we have determined that, on average, reducedemission completions are cost effective, well and reservoircharacteristics could vary, such that some REC are morecost effective than others. Unlike most stationary sourcecontrols, REC equipment is used only for a 3 to 10 dayperiod. Our review found that most operators contract withservice companies to perform REC rather than purchase theequipment themselves, which was reflected in our economicanalysis. It is also possible that the contracting costs of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

125. Page 125 of 604supplying and operating REC equipment may rise in the shortterm with the increased demand for those services. Werequest comment and any available technical information tojudge whether our assumption of $33,237 per well completionfor this service given the projected number of wells in2015 subject to this requirement is accurate. We believe that the proposed rule regulates onlysignificant emission sources for which controls are cost-effective. Nevertheless, we solicit comment and supportingdata on appropriate thresholds (e.g., pressure, flowrate)that we should consider in specifying which wellcompletions are subject to the REC requirements forsubcategory 1 wells. Comments specifying thresholds shouldinclude an analysis of why sources below these thresholdsare not cost effective to control. In addition, there may be economic, technical or otheropportunities or barriers associated with performing costeffective REC that we have not identified in our review.For example, some small regulated entities may have anincreased source of revenue due to the captured product. Onthe other hand, some small regulated entities may have lessaccess to REC than larger regulated entities might have. Werequest information on such opportunities and barriers that This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

126. Page 126 of 604we should consider and suggestions for how we may take theminto account in structuring the NSPS. The second subcategory of fractured gas wells includesexploratory wells or delineation wells. Because these typesof wells generally are not in proximity to existinggathering lines, REC is not an option, since there is noinfrastructure in place to get the recovered gas to marketor further processing. For these wells, the only potentialcontrol option we were able to identify is pit flaring,described above. As explained above, because of the slugflow nature of the flowback gas, water and sand, control bya traditional flare control device or other controldevices, such as vapor recovery units, is infeasible, whichleaves pit flaring as the only practicable control systemfor subcategory 2 wells. As also discussed above, open pitflaring can present a fire hazard or other undesirableimpacts in some situations. Aside from the potentialhazards associated with pit flaring, in some cases, we didnot identify any nonair environmental impacts, health orenergy impacts associated with pit flaring. However, pitflaring would produce NOx emissions. As in the case ofcategory 1 wells, we believe that these emissions cannot becontrolled or measured directly due to the open combustion This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

127. Page 127 of 604process characteristic of pit flaring. We again usedpublished emission factors to estimate the NOx emissionsfor purposes of assessing secondary impacts. For category 2well completions, we estimated that 0.32 tons of NOx areproduced as secondary emissions per completion event. Thisis based on the assumption that 95 percent of flowback gasis combusted by the combustion device. The 22 tons of VOCreduced during the pit flaring used to control category 2well completions is approximately 69 times greater than theNOx produced. Thus, we believe that the benefit of the VOCreduction far outweighs the secondary impact of NOxformation during pit flaring. In light of the above, we propose to determine thatBSER for subcategory 2 wells would be pit flaring. As weexplained above, it is not practicable to measure theemissions during pit flaring or venting because the gas isdischarged during flowback mixed with water and sand inmultiphase slug flow. It is, therefore, not feasible to seta numerical performance standard. Pursuant to CAA section 111(h)(2), we are proposing anoperational standard for subcategory 2 wells that requiresminimization of venting of gas and hydrocarbon vapors This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

128. Page 128 of 604during the completion operation through the use of pitflaring, with provisions for venting in lieu of pit flaringfor situations in which flaring would present safetyhazards or for periods when the flowback gas isnoncombustible due to high concentrations of nitrogen orcarbon dioxide. Consistent with requirements for subcategory 1 wells,owners or operators of subcategory 2 wells would berequired to document completions and provide justificationfor periods when gas was vented in lieu of combustion. Wesolicit comment on whether there are other such situationswhere flaring would be unsafe or infeasible and potentialcriteria that would support venting in lieu of pit flaring. For controlling completion emissions at oil wells andconventional (non-fractured) gas wells, we have identifiedand evaluated the following control options: REC inconjunction with pit flaring and pit flaring alone. Due tothe low uncontrolled VOC emissions of approximately 0.007ton per completion and, therefore, low potential emissionreductions from these events, the cost per ton of reductionbased on REC would be extremely high (over $700,000 per tonof VOC reduced). We evaluated the use of pit flaring aloneas a system for controlling emissions from oil wells and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

129. Page 129 of 604conventional gas wells and determined that the cost cost-effectiveness would be approximately $520,000 per ton foroil wells and approximately $32,000 per ton forconventional gas wells. In light of the high cost per tonof VOC reduction, we do not consider either of thesecontrol options to be BSER for oil wells and conventionalwells. We propose that fracturing (or refracturing) andcompletion of an existing well (i.e., a well existing priorto (INSERT DATE OF PUBLICATION IN THE FEDERAL REGISTER)) isconsidered a modification under CAA section 111(a), becausephysical change occurs to the existing well, which includesthe wellbore, casing and tubing, resulting in an emissionsincrease during the completion operation. The physicalchange, in this case, would be caused by the reperforationof the casing and tubing, along with the refracturing ofthe wellbore. The increased VOC emissions would occurduring the flowback period following the fracturing orrefracturing operation. Therefore, the proposed standardsfor category 1 and category 2 wells would apply tocompletions at existing fractured or refractured wells. EPA seeks comment on the 10 percent per year rate ofrefracturing for natural gas wells assumed in the impacts This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

130. Page 130 of 604analysis found in the TSD. EPA has received anecdotalinformation suggesting that refracturing could be occurringmuch less frequently, while others suggest that the percentof wells refractured in a given year could be greater. Weseek comment and comprehensive data and information on therate of refracturing and key factors that influence ordetermine refracturing frequency. In addition to well completions, we considered VOCemissions occurring at the wellhead affected facilityduring subsequent day-to-day operations during wellproduction. As discussed below in section VI.B.1.e, VOCemissions from wellheads are very small during productionand account for about 2.6 tons VOC per year. We are notaware of any cost effective controls that can be used toaddress these relatively small emissions.b. NSPS for Pneumatic Controllers Pneumatic controllers are automated instruments usedfor maintaining a process condition, such as liquid level,pressure, pressure differential and temperature. Pneumaticcontrollers are widely used in the oil and natural gassector. In many situations across all segments of the oiland gas industry, pneumatic controllers make use of theavailable high-pressure natural gas to operate. In these This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

131. Page 131 of 604“gas-driven” pneumatic controllers, natural gas may bereleased with every valve movement or continuously from thevalve control pilot. The rate at which this release occursis referred to as the device bleed rate. Bleed rates aredependent on the design of the device. Similar designs willhave similar steady-state rates when operated under similarconditions. Gas-driven pneumatic controllers are typicallycharacterized as “high-bleed” or “low-bleed,” where a high-bleed device releases more than 6 standard cubic feet perhour (scfh) of gas, with 18 scfh bleed rate being what weused in our analyses below. There are three basic designs:(1) Continuous bleed devices (high or low-bleed) are usedto modulate flow, liquid level or pressure and gas isvented at a steady-state rate; (2) actuating/intermittentdevices (high or low-bleed) perform quick control movementsand only release gas when they open or close a valve or asthey throttle the gas flow; and (3) self-contained devicesrelease gas to a downstream pipeline instead of to theatmosphere. We are not aware of any add-on controls thatare or can be used to reduce VOC emissions from gas-drivenpneumatic devices. For an average high-bleed pneumatic controller locatedin production (where the content of VOC in the raw product This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

132. Page 132 of 604stream is relatively high), the difference in VOC emissionsbetween a high-bleed controller and a low-bleed controlleris around 1.8 tpy. For the transmission and storage segment(where the content of VOC in the pipeline quality gas isrelatively low), the difference in VOC emissions between ahigh-bleed controller and a low-bleed controller is around0.89 tpy. We have developed projections that estimate thatapproximately 13,600 new gas-driven units in the productionsegment and 67 new gas-driven units in the transmission andstorage segment will be installed each year, includingreplacement of old units. Not all pneumatic controllers aregas driven. These “non-gas driven” pneumatic controllersuse sources of power other than pressurized natural gas,such as compressed “instrument air.” Because these devicesare not gas driven, they do not release natural gas or VOCemissions, but they do have energy impacts becauseelectrical power is required to drive the instrument aircompressor system. Electrical service of at least 13.3kilowatts (kW) is required to power a 10 horsepower (hp)instrument air compressor, which is a relatively smallcapacity compressor. At sites without available electricalservice sufficient to power an instrument air compressor,only gas driven pneumatic devices can be used. During our This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

133. Page 133 of 604review, we determined that gas processing plants are theonly facilities in the oil and natural gas sector highlylikely to have electrical service sufficient to power aninstrument air system, and that approximately half ofexisting gas processing plants are using non-gas drivendevices. For devices at gas processing plants, we evaluated theuse of non-gas driven controllers and low-bleed controllersas options for reducing VOC emissions, with high-bleedcontrollers being the baseline. As mentioned above, non-gasdriven devices themselves have zero emissions, but they dohave energy impacts because electrical power is required todrive the instrument air compressor system. In our costanalysis, we determined that the annualized cost ofinstalling and operating a fully redundant 10 hp (13.3 kW)instrument air system (systems generally are designed withredundancy to allow for system maintenance and failurewithout loss of air pressure), including duplicatecompressors, air tanks and dryers, would be $11,090. Asystem of this size is capable of serving 15 control loopsand reducing VOC emissions by 4.2 tpy, for a costeffectiveness of $2,659 per ton of VOC reduced. If thesavings of the salable natural gas that would have been This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

134. Page 134 of 604emitted is considered, the value of the gas not emittedwould help offset the cost for this control, bringing thecost per ton of VOC down to $1,824. We also evaluated the use of low-bleed controllers inplace of high-bleed controllers at processing plants. Weevaluated the impact of bleeding 6 standard cubic feet ofnatural gas per hour, which is the maximum bleed rate fromlow-bleed controllers, according to manufacturers of thesedevices. We chose natural gas as a surrogate for VOC,because manufacturers’ technical specifications forpneumatic controllers are stated in terms of natural gasbleed rate rather than VOC. The capital cost differencebetween a new high-bleed controller and a new low-bleedcontroller is estimated to be $165. Without taking intoaccount the savings due to the natural gas losses avoided,the annual costs are estimated to be around $23 per year,which is a cost of $13 per ton of VOC reduced for theproduction segment. If the savings of the salable naturalgas that would have been emitted is considered, there is anet savings of $1,519 per ton of VOC reduced. Although the non-gas-driven controller system is moreexpensive than the low-bleed controller system, it is stillreasonably cost-effective. Furthermore, the non-gas-driven This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

135. Page 135 of 604controller system achieves a 100-percent VOC reduction incontrast to a 66-percent reduction achieved by a low-bleedcontroller. Moreover, we believe the collateral emissionsfrom electrical power generation needed to run thecompressor are very low. Finally, non-gas-driven pneumaticcontrollers avoid potentially explosive concentrations ofnatural gas which can occur as a result of normal bleedingfrom groups of gas-driven pneumatic controllers located inclose proximity, as they often are at gas processingplants. Based on our review described above, we believethat a non-gas-driven controller is BSER for reducing VOCemissions from pneumatic devices at gas processing plants.Accordingly, the proposed standard for pneumatic devices atgas processing plants is a zero VOC emission limit. For the production (other than processing plants) andtransmission and storage segments, where electrical servicesufficient to power an instrument air system is likelyunavailable and, therefore, only gas-driven devices can beused, we evaluated the use of low-bleed controllers inplace of high-bleed controllers. Just as in our analysis oflow-bleed controllers as an option for gas processingplants, we evaluated the impact of bleeding 6 standardcubic feet per minute (scfm) of natural gas per hour This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

136. Page 136 of 604contrasted with 18 scfm from a high-bleed unit. Again, thecapital cost difference between a new high-bleed controllerand a new low-bleed controller is estimated to be $165.Without taking into account the savings due to the naturalgas losses avoided, the annual costs are estimated to bearound $23 per year, which is a cost of $13 per ton of VOCreduced for the production segment. If the savings of thesalable natural gas that would have been emitted isconsidered, there is a net savings for this control. In thetransmission and storage segment, where the VOC content ofthe vented gas is much lower than in the productionsegment, the cost effectiveness of a low-bleed pneumaticdevice is estimated to be around $262 per ton of VOCreduced. However, there are no potential offsetting savingsto be realized in the transmission and storage segment,since the operators of transmission and storage stationstypically do not own the gas they are handling. Based onour evaluation of the emissions and costs, we believe thatlow-bleed controllers represent BSER for pneumaticcontrollers in the production (other than processingplants) and transmission and storage segments. Therefore,for pneumatic devices at these locations, we propose a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

137. Page 137 of 604natural gas bleed rate limit of 6.0 scfh to reflect the VOClimit with the use of a low-bleed controller. There may be situations where high-bleed controllersand the attendant gas bleed rate greater than 6 cubic feetper hour, are necessary due to functional requirements,such as positive actuation or rapid actuation. An examplewould be controllers used on large emergency shutdownvalves on pipelines entering or exiting compressionstations. For such situations, we have provided in theproposed rule an exemption where pneumatic controllersmeeting the emission standards discussed above would pose afunctional limitation due to their actuation response timeor other operating characteristics. We are requestingcomments on whether there are other situations that shouldbe considered for this exemption. If you provide suchcomment, please specify the criteria for such situationsthat would help assure that only appropriate exemptions areclaimed. The proposed standards would apply to installation ofa new pneumatic device (including replacing an existingdevice with a new device). We consider that a pneumaticdevice, an apparatus, is an affected facility and eachinstallation is construction subject to the proposed NSPS. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

138. Page 138 of 604See definitions of “affected facility” and “construction”at 40 CFR 60.2.c. NSPS for Compressors There are many locations throughout the oil andnatural gas sector where compression of natural gas isrequired to move it along the pipeline. This isaccomplished by compressors powered by combustion turbines,reciprocating internal combustion engines or electricmotors. Turbine-powered compressors use a small portion ofthe natural gas that they compress to fuel the turbine. Theturbine operates a centrifugal compressor, which compressesthe natural gas for transit through the pipeline. Sometimesan electric motor is used to turn a centrifugal compressor.This type of compressor does not require the use of any ofthe natural gas from the pipeline, but it does require asubstantial source of electricity. Reciprocating sparkignition engines are also used to power many compressors,referred to as reciprocating compressors, since theycompress gas using pistons that are driven by the engine.Like combustion turbines, these engines are fueled bynatural gas from the pipeline. Both centrifugal andreciprocating compressors are sources of VOC emissions andwere evaluated for coverage under the NSPS. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

139. Page 139 of 604 Centrifugal Compressors. Centrifugal compressorsrequire seals around the rotating shaft to minimize gasleakage and fugitive VOC emissions from where the shaftexits the compressor casing. There are two types of sealsystems: Wet seal systems and mechanical dry seal systems. Wet seal systems use oil, which is circulated underhigh pressure between three or more rings around thecompressor shaft, forming a barrier to minimize compressedgas leakage. Very little gas escapes through the oilbarrier, but considerable gas is absorbed by the oil. Theamount of gas absorbed and entrained by the oil barrier isaffected by the operating pressure of the gas beinghandled; higher operating pressures result in higherabsorption of gas into the oil. Seal oil is purged of theabsorbed and entrained gas (using heaters, flash tanks anddegassing techniques) and recirculated to the seal area forreuse. Gas that is purged from the seal oil is commonlyvented to the atmosphere. Degassing of the seal oil emitsan average of 47.7 scfm of gas, depending on the operatingpressure of the compressor. An uncontrolled wet seal systemcan emit, on average, approximately 20.5 tpy of VOC duringthe venting process (production segment) or about 3.5 tpy(transmission and storage segment). We identified two This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

140. Page 140 of 604potential control techniques for reducing emissions fromdegassing of wet seal systems: (1) Routing the gas back toa low pressure fuel stream to be combusted as fuel gas and(2) routing the gas to a flare. We know only of anecdotal,undocumented information on routing of the gas back to afuel stream and, therefore, were unable to assess costs andcost effectiveness of the first option. Although we do nothave specific examples of routing emissions from wet sealdegassing to a flare, we were able to estimate the cost,emission reductions and cost effectiveness of the secondoption using uncontrolled wet seals as a baseline. Based on the average uncontrolled emissions of wetseal systems discussed above and a flare efficiency of 95percent, we determined that VOC emission reductions from awet seal system would be an average of 19.5 tpy (productionsegment) or 3.3 tpy (transmission and storage segment).Using an annualized cost of flare installation andoperation of $103,373, we estimated the incremental costeffectiveness of this option (from uncontrolled wet sealsto controlled wet seals using a flare) to be approximately$5,300/ton and $31,000/ton for the production segment andtransmission and storage segment, respectively. With thisoption, there would be secondary air impacts from This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

141. Page 141 of 604combustion. However we did not identify any nonair qualityor energy impacts associated with this control technique. Dry seal systems do not use any circulating seal oil.Dry seals operate mechanically under the opposing forcecreated by hydrodynamic grooves and springs. Fugitiveemissions occur from dry seals around the compressor shaft.Based on manufacturer studies and engineering designestimates, fugitive emissions from dry seal systems areapproximately 6 scfm of gas, depending on the operatingpressure of the compressor. A dry seal system can havefugitive emissions of, on average, approximately 2.6 tpy ofVOC (production segment) or about 0.4 tpy (transmission andstorage segment). We did not identify any control devicesuitable to capture and control the fugitive emissions fromdry seals around the compressor shaft. Using uncontrolled wet seals as a baseline, weevaluated the reductions and incremental cost effectivenessof dry seal systems. Based on the average fugitiveemissions, we determined that VOC emission reductionsachieved by dry seal systems compared to uncontrolled wetseal systems would be 18 tpy (production segment) and 3.1tpy (transmission and storage segment). Combined with anannualized cost of dry seal systems of $10,678, the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

142. Page 142 of 604incremental cost effectiveness compared to uncontrolled wetseal systems would be $595/ton and $3,495/ton for theproduction segment and transmission and storage segment,respectively. We identified neither nonair quality nor anyenergy impacts associated with this option. In performing our analysis, we estimated theincremental cost of a dry seal compressor over that of anequivalent wet seal compressor to be $75,000. This valuewas obtained from a vendor who represents a large share ofthe market for centrifugal compressors. However, thisnumber likely represents a conservatively high valuebecause wet seal units have a significant amount ofancillary equipment, namely the seal oil system and, thus,additional capital expenses. Dry seal systems have someancillary equipment (the seal gas filtration system), butthe costs are less than the wet seal oil system. We werenot able to directly confirm this assumption with thevendor, however, a search of product literature showed thatseal oil systems and seal gas filtration systems aretypically listed separate from the basic compressorpackage. Using available data on the cost of thisequipment, it is very likely that the cost of purchasing adry seal compressor may actually be lower that a wet seal This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

143. Page 143 of 604compressor. We seek comment on available cost data of a dryseal versus wet seal compressor, including all ancillaryequipment costs. In light of the above analyses, we propose todetermine that dry seal systems are BSER for reducing VOCemissions from centrifugal compressors. We evaluated thepossibility of setting a performance standard that reflectsthe emission limitation achievable through the use of a dryseal system. However, as mentioned above, VOC fromcentrifugal compressors with dry seals are fugitiveemissions from around the compressor shafts. There is nodevice to capture and control these fugitive emissions, norcan reliable measurement of these emissions be conducteddue to difficulty in accessing the leakage area and dangerof contacting the shaft rotating at approximately 30,000revolutions per minute. This not only poses a likely hazardthat would destroy test equipment on contact, it poses asafety hazard to personnel, as well. Therefore, pursuant tosection 111(h)(2) of the CAA, we are proposing an equipmentstandard that would require the use of dry seals to limitthe VOC emissions from new centrifugal compressors. Weconsider that a centrifugal compressor, an apparatus, is anaffected facility and each installation is construction This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

144. Page 144 of 604subject to the proposed NSPS. See definitions of “affectedfacility” and “construction” at 40 CFR 60.2. Accordingly,the proposed standard would apply to installation of newcentrifugal compressors at new locations, as well asreplacement of old compressors. Although we are proposing to determine dry sealsystems to be BSER for centrifugal compressors, we aresoliciting comments on the emission reduction potential,cost and any limitations for the option of routing the gasback to a low pressure fuel stream to be combusted as fuelgas. In addition, we solicit comments on whether there aresituations or applications where wet seal is the onlyoption, because a dry seal system is infeasible orotherwise inappropriate. Reciprocating Compressors. Reciprocating compressorsin the natural gas industry leak natural gas fugitive VOCduring normal operation. The highest volumes of gas lossand fugitive VOC emissions are associated with piston rodpacking systems. Packing systems are used to maintain atight seal around the piston rod, preventing the highpressure gas in the compressor cylinder from leaking, whileallowing the rod to move freely. This leakage rate isdependent on a variety of factors, including physical size This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

145. Page 145 of 604of the compressor piston rod, operating speed and operatingpressure. Under the best conditions, new packing systemsproperly installed on a smooth, well-aligned shaft can beexpected to leak a minimum of 11.5 scfh. Higher leak ratesare a consequence of fit, alignment of the packing partsand wear. We evaluated the possibility of reducing VOC emissionsfrom reciprocal compressors through a control device.However, VOC from reciprocating compressors are fugitiveemissions from around the compressor shafts. Although it ispossible to construct an enclosure around the rod packingarea and vent the emissions outside for safety purposes,connection to a closed vent system and control device wouldcreate back pressure on the leaking gas. This back pressurewould cause the leaked gas instead to be forced inside thecrankcase of the engine, which would dilute lubricatingoil, causing premature failure of engine bearings, pose anexplosion hazard and eventually be vented from thecrankcase breather, defeating the purpose of a controldevice. As mentioned above, as packing wears and deteriorates,leak rates can increase. We, therefore, evaluatereplacement of compressor rod packing systems as an option This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

146. Page 146 of 604for reducing VOC emissions. Conventional bronze-metallicpacking rings wear out and need to be replaced every 3 to 5years, depending on the compressor’s rate of usage (i.e.,the percentage of time that a compressor is in pressurizedmode). Based on industry experience in the Natural Gas STARprogram and other sources, we evaluated the rod packingreplacement costs for reciprocating compressors atdifferent segments of this industry. Usage rates vary bysegment. Usage rates for compressors at wellheads,gathering/boosting stations, processing plants,transmission stations and storage facilities are 100, 79,90, 79 and 68 percent, respectively. Reciprocatingcompressors at wellheads are small and operate at lowerpressures, which limit VOC emissions from these sources.Due to the low VOC emissions from these compressors, about0.044 tpy, combined with an annual cost of approximately$3,700, the cost per ton of VOC reduction is rather high.We estimated that the cost effectiveness of controllingwellhead compressors is over $84,000 per ton of VOCreduced, which we believe to be too high and, therefore,not reasonable. Because the cost effectiveness of replacingpacking wellhead compressor rod systems is not reasonable, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

147. Page 147 of 604and absent other emission reduction measures, we did notfind a BSER for reducing VOC emissions from reciprocalcompressors at wellheads. For reciprocating compressors located at other oil andgas operations, we estimated that the cost effectiveness ofcontrolling compressor VOC emissions by rod packingreplacement would be $870 per ton of VOC for reciprocatingcompressors at gathering and boosting stations, $270 perton of VOC for reciprocating compressors at processingstations, $2,800 per ton of VOC for reciprocatingcompressors at transmission stations and $3,700 per ton ofVOC for reciprocating compressors at underground storagefacilities. We consider these costs to be reasonable. Wedid not identify any nonair quality health or environmentalimpacts or energy impacts associated with rod packingreplacement. In light of the above, we propose to determinethat such control is the BSER for reducing VOC emissionfrom compressors at these other oil and gas operations. Because VOC emitted from reciprocal compressors arefugitive emissions, there is no device to capture andcontrol the emissions. Therefore, pursuant to section111(h) of the CAA, we are proposing an operationalstandard. Based on industry experience reported to the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

148. Page 148 of 604Natural Gas STAR program, we determined that packing rodsshould be replaced every 3 years of operation. However, toaccount for segments of the industry in which reciprocatingcompressors operate in pressurized mode a fraction of thecalendar year (ranging from approximately 68 percent up toapproximately 90 percent), the proposed rule expresses thereplacement requirement in terms of hours of operationrather than on a calendar year basis. One year ofcontinuous operation would be 8,760 hours. Three years ofcontinuous operation would be 26,280 hours, or rounded tothe nearest thousand, 26,000 hours. Accordingly, theproposed rule would require the replacement of the rodpacking every 26,000 hours of operation. The owner oroperator would be required to monitor the hours ofoperation beginning with the installation of thereciprocating compressor affected facility. Cumulativehours of operation would be reported each year in thefacility’s annual report. Once the hours of operationreached 26,000 hours, the owner or operator would berequired to change the rod packing immediately, althoughunexpected shutdowns could be avoided by tracking hours ofoperation and planning for packing replacement at scheduledmaintenance shutdowns before the hours of operation reached This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

149. Page 149 of 60426,000. Some industry partners of the Natural Gas STAR programcurrently conduct periodic testing to determine the leakagerates that would identify economically beneficialreplacement of rod packing based on natural gas savings.Therefore, we are soliciting comments on incorporating amethod similar to that in the Natural Gas STAR’s LessonsLearned document entitled, Reducing Methane Emissions fromCompressor Rod Packing Systems(http://www.epa.gov/gasstar/documents/ll_rodpack.pdf), tobe incorporated in the NSPS. We are soliciting comments onhow to determine a suitable leak threshold above which rodpacking replacement would be cost effective for VOCemission reduction. We are also soliciting comment on theappropriate replacement frequency and other considerationsthat would be associated with regular replacement periods.d. NSPS for Storage Vessels Crude oil, condensate and produced water are typicallystored in fixed-roof storage vessels. Some vessels used forstoring produced water may be open-top tanks. Thesevessels, which are operated at or near atmospheric pressureconditions, are typically located as part of a tankbattery. A tank battery refers to the collection of process This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

150. Page 150 of 604equipment used to separate, treat and store crude oil,condensate, natural gas and produced water. The extractedproducts from productions wells enter the tank batterythrough the production header, which may collect productfrom many wells. Emissions from storage vessels are a result ofworking, breathing and flash losses. Working losses occurdue to the emptying and filling of storage tanks. Breathinglosses are the release of gas associated with dailytemperature fluctuations and other equilibrium effects.Flash losses occur when a liquid with dissolved gases istransferred from a vessel with higher pressure to a vesselwith lower pressure, thus, allowing dissolved gases and aportion of the liquid to vaporize or flash. In the oil andnatural gas production segment, flashing losses occur whenlive crude oils or condensates flow into a storage tankfrom a processing vessel operated at a higher pressure.Typically, the larger the pressure drop, the more flashemissions will occur in the storage stage. Temperature ofthe liquid also influences the amount of flash emissions.The amount of liquid entering the tank during a given time,commonly known as throughput, also affects the emissionrate, with higher throughput tanks having higher annual This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

151. Page 151 of 604emissions, given that other parameters are the same. In analyzing controls for storage vessels, we reviewedcontrol techniques identified in the Natural Gas STARprogram and state regulations. We identified two ways ofcontrolling storage vessel emissions, both of which canreduce VOC emissions by 95 percent. One option would be toinstall a vapor recovery unit (VRU) and recover all thevapors from the tanks. The other option would be to routethe emissions from the tanks to a flare control device.These devices could be “candlestick” flares that are foundat gas processing plants or other larger facilities orenclosed combustors which are commonly found at smallerfield facilities. We estimated the total annual cost for aVRU to be approximately $18,900/yr and for a flare to beapproximately $8,900/yr. Cost effectiveness of thesecontrol options depend on the amount of vapor produced bythe storage vessels being controlled. A VRU has a potentialadvantage over flaring, in that it recovers hydrocarbonvapors that potentially can be used as supplemental burnerfuel, or the vapors can be condensed and collected ascondensate that can be sold. If natural gas is recovered,it can be sold, as well, as long as a gathering line isavailable to convey the recovered salable gas product to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

152. Page 152 of 604market or to further processing. A VRU also does not havesecondary air impacts that flaring does, as describedbelow. However, a VRU cannot be used in all instances. Someconditions that affect the feasibility of VRU are:Availability of electrical service sufficient to power theVRU; fluctuations in vapor loading caused by surges inthroughput and flash emissions from the tank; potential fordrawing air into condensate tanks causing an explosionhazard; and lack of appropriate destination or use for thevapor recovered. Like a VRU, a flare control device can also achieve acontrol efficiency of 95 percent. There are no technicallimitations on the use of flares to control vapors fromcondensate and crude oil tanks. However, flaring has asecondary impact from emissions of NOx and otherpollutants. In light of the technical limitations with theuse of a VRU, we are unable to conclude that a VRU isbetter than flaring. We, therefore, propose to determinethat both a VRU and flare are BSER for reducing VOCemission from storage vessels. We propose an NSPS of 95-percent reduction for storage vessels to reflect the levelof emission reduction achievable by VRU and flares. VOC emissions from storage vessels vary significantly, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

153. Page 153 of 604depending on the rate of liquid entering and passingthrough the vessel (i.e., its throughput), the pressure ofthe liquid as it enters the atmospheric pressure storagevessel, the liquid’s volatility and temperature of theliquid. Some storage vessels have negligible emissions,such as those with very little throughput and/or handlingheavy liquids entering at atmospheric pressure. We do notbelieve that it is cost effective to control these vessels.We believe it is important to control tanks withsignificant VOC emissions under the proposed NSPS. In our analysis, we evaluated storage tanks withvarying condensate or crude oil throughput. We usedemission factors developed for the Texas EnvironmentalResearch Consortium in a study that evaluated VOC emissionsfrom crude oil and condensate storage tanks by performingdirect measurements. The study found that the average VOCemission factor for crude oil storage tanks was 1.6 pounds(lb) VOC per barrel of crude oil throughput. The averageVOC emission factor for condensate tanks was determined tobe 33.3 lb VOC per barrel of condensate throughput.Applying these emission factors and evaluating condensatethroughput rates of 0.5, 1, 2 and 5 barrels per day (bpd),we determined that VOC emissions at these condensate This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

154. Page 154 of 604throughput rates would be approximately 3, 6, 12 and 30tpy, respectively. Similarly, we evaluated crude oilthroughput rates of 1, 5, 20 and 50 bpd. Based on the Texasstudy, these crude oil throughput rates would result in VOCemissions of 0.3, 1.5, 5.8 and 14.6 tpy, respectively. Webelieve that it is important to control tanks withsignificant VOC emissions. Furthermore, we believe it wouldbe easier and less costly for owners and operators todetermine applicability by using a throughput thresholdinstead of an emissions threshold. As a result of the aboveanalyses, we believe that storage vessels with at least 1bpd of condensate or 20 bpd of crude oil should becontrolled. These throughput rates are equivalent to VOCemissions of approximately 6 tpy. Based on an estimatedannual cost of $18,900 for the control device, controllingstorage vessels with these condensate or crude oilthroughputs would result in a cost effectiveness of $3,150per ton of VOC reduced. Based on our evaluation, we propose to determine thatboth a VRU and flare are BSER for reducing VOC emissionfrom storage vessels with throughput of at least 1 barrelof condensate per day or 20 barrels of crude oil per day.We propose an NSPS of 95-percent reduction for these This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

155. Page 155 of 604storage vessels to reflect the level of emission reductionachievable by VRU and flare control devices. For storage vessels below the throughput levelsdescribed above (“small throughput tanks”), for which we donot consider flares or VRU to be cost effective controls,we evaluated other measures to reduce VOC emissions.Standard practices for such tanks include requiring a coverthat is well designed, maintained in good condition andkept closed. Crude oil and condensate storage tanks in theoil and natural gas sector are designed to operate at orjust slightly above or below atmospheric pressure.Accordingly, they are provided with vents to prevent tankdestruction under rapid pressure increases due to flashemissions conditions. Studies by the Natural Gas STARprogram and by others have shown that working losses (i.e.,those emissions absent flash emission conditions) are verylow, approaching zero. During times of flash emissions,tanks are designed such that the flash emissions arereleased through a vent on the fixed roof of the tank whenpressure reaches just a few ounces to prevent pressurebuildup and resulting tank damage. At those times, vaporreadily escapes through the vent to protect the tank. Testshave shown that open hatches or leaking hatch gaskets have This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

156. Page 156 of 604little effect on emissions from uncontrolled tanks due tothe functioning roof vent. However, in the case ofcontrolled tanks, the control requirements includeprovisions for maintaining integrity of the closed ventsystem that conveys emissions to the control device,including hatches and other tank openings. As a result,hatches are required to be kept closed and gaskets kept ingood repair to meet control requirements of controlledstorage vessels. Because the measures we evaluated,including maintenance of hatch integrity, do not provideappreciable emission reductions for storage vessels withthroughputs under 1 barrel of condensate per day and 21barrels of crude oil per day, we believe that the controloptions we evaluated do not reflect BSER for the smallthroughput tanks and we are not proposing standards forthese tanks. As discussed in section VII of this package, we areproposing to amend the NESHAP for oil and natural gasproduction facilities at 40 CFR part 63, subpart HH torequire that all storage vessels at production facilitiesreduce HAP emissions by 95 percent. Because the controlsused to achieve the 95-percent HAP reduction are the sameas the proposed BSER for VOC reduction for storage vessels This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

157. Page 157 of 604(i.e., VRU and flare), sources that are achieving the 95-percent HAP reduction would also be meeting the proposedNSPS of 95-percent VOC reduction. In light of the above,and to avoid duplicate monitoring, recordkeeping andreporting, we propose that storage vessels subject to therequirements of subpart HH are exempt from the proposedNSPS for storage vessel in 40 CFR part 60, subpart OOOO.e. NSPS for VOC Equipment Leaks Equipment leaks are fugitive emissions emanating fromvalves, pump seals, flanges, compressor seals, pressurerelief valves, open-ended lines and other process andoperation components. There are several potential reasonsfor equipment leak emissions. Components such as pumps,valves, pressure relief valves, flanges, agitators andcompressors are potential sources that can leak due to sealfailure. Other sources, such as open-ended lines andsampling connections may leak for reasons other than faultyseals. In addition, corrosion of welded connections,flanges, and valves may also be a cause of equipment leakemissions. Because of the large number of valves, pumps andother components within an oil and gas production,processing and transmission facility, equipment leakvolatile emissions from these components can be This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

158. Page 158 of 604significant. Natural gas processing plants, especiallythose using refrigerated absorption and transmissionstations tend to have a large number of components.Equipment leaks from processing plants are addressed in ourreview of 40 CFR part 60, subpart KKK, which is discussedabove in section VI.B.1. In addition to gas processing plants, these types ofequipment also exist at oil and gas production sites andgas transmission and storage facilities. While the numberof components at individual transmission and storagefacilities is relatively smaller than at processing plants,collectively, there are many components that can result insignificant emissions. Therefore, we evaluated applying NSPS for equipmentleaks to facilities in the production segment of theindustry, which includes everything from the wellhead tothe point that the gas enters the processing plant,transmission pipeline or distribution pipeline. Productionfacilities can vary significantly in the operationsperformed and the processes, all of which impact the numberof components and potential emissions from leakingequipment and, thus, impact the annual costs related toimplementing a LDAR program. We used data collected by the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

159. Page 159 of 604Gas Research Institute to develop model productionfacilities. Baseline emissions, along with emissionreductions and costs of regulatory alternatives, wereestimated using these model production facilities. Weconsidered production facilities where separation, storage,compression and other processes occur. These facilities maynot have a wellhead on-site, but would be associated with awellhead. We also evaluated gathering and boostingfacilities, where gas and/or oil are collected from anumber of wells, then processed and transported downstreamto processing plants or transmission stations. We evaluatedthe impacts at these production facilities with varyingnumber of operations and equipment. We also developed amodel plant for the transmission and storage segment usingdata from the Gas Research Institute. Details of theseevaluations may be found in the TSD in the docket. For an average production site at or associated with awellhead, we estimated annual VOC emissions from equipmentleaks of around 2.6 tpy. For an average gathering/boostingfacility, we estimated the annual VOC emissions fromequipment leaks to be around 9.8 tpy. The averagetransmission and storage facility emits 2.7 tpy of VOC. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

160. Page 160 of 604 For facilities in each non-gas processing plantsegment, we evaluated the same four options as we did forgas processing plants in section VI.B.1 above. These fouroptions are as follows: (1) 40 CFR part 60, subpart VVa-level LDAR (which is based on conducting Method 21 monthly,defining “leak” at 500 ppm threshold, and adding connectorsto the VV list of components to be monitored); (2) monthlyoptical gas imaging with annual Method 21 check (thealternative work practice for monitoring equipment forleaks at 40 CFR 60.18(g)); (3) monthly optical gas imagingalone; and (4) annual optical gas imaging alone. For option 1, we evaluated subpart VVa-LDAR as awhole. We also analyzed separately the individual types ofcomponents (valves, connectors, pressure relief devices andopen-ended lines). Detailed discussions of these componentby component analyses are included in the TSD in thedocket. Based on our evaluation, subpart VVa-level LDAR(Option 1) results in more VOC reduction than the subpartVV-level LDAR currently required for gas processing plants,because more leaks are found based on the lower definitionof “leak” under subpart VVa(10,000 ppm for subpart VV and500 ppm for subpart VVa). In addition, our evaluation This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

161. Page 161 of 604shows that the cost per ton of VOC reduced for subpart VValevel controls is less than the cost per ton of VOC reducedfor the less stringent subpart VV level of control.Although the cost of repairing more leaks is higher, theincreased VOC control afforded by subpart VVa levelcontrols more than offsets the increased costs. For the subpart VVa level of control at the averageproduction site associated with a wellhead, averagefacility-wide cost-effectiveness would be $16,084 per tonof VOC. Component-specific cost-effectiveness ranged from$15,063 per ton of VOC (for valves) to $211,992 per ton ofVOC (for pressure relief devices), with connectors andopen-ended lines being $74,283 and $180,537 per ton of VOC,respectively. We also looked at component costs for amodified subpart VVa level of control with less frequentmonitoring for valves and connectors at production sitesassociated with a wellhead.12 The cost-effectiveness forvalves was calculated to be $17,828 per ton of VOC byreducing the monitoring frequency from monthly to annually.The cost-effectiveness for connectors was calculated to be$87,277 per ton of VOC by reducing the monitoring frequency This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

162. Page 162 of 604from every 4 years to every 8 years after the initialcompliance period. We performed a similar facility-wide and component-specific analysis of option 1 LDAR for gathering andboosting stations. For the subpart VVa level of control atthe average gathering and boosting station, facility-widecost-effectiveness was estimated to be $9,344 per ton ofVOC. Component-specific cost-effectiveness ranged from$6,079 per ton of VOC (for valves) to $77,310 per ton ofVOC (for open-ended lines), with connectors and pressurerelief devices being $23,603 and $72,523 per ton,respectively. For the modified subpart VVa level ofcontrol at gathering and boosting stations, cost-effectiveness ranged from $5,221 per ton of VOC (forvalves) to $77,310 per ton of VOC (for open-ended lines),with connectors and pressure relief devices being $27,274and $72,523 per ton, respectively. The modified subpart VValevel controls were more cost-effective than the subpartVVa level controls for valves, but not for connectors.This is due to the low cost of monitoring connectors andthe low VOC emissions from leaking connectors. We also performed a similar analysis of option 1subpart VVa-level LDAR for gas transmission and storage This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

163. Page 163 of 604facilities. For the subpart VVa level of control at theaverage transmission and storage facility, facility-widecost-effectiveness was $20,215. Component-specific cost-effectiveness ranged from $24,762 per ton of VOC (for open-ended lines) to $243,525 per ton of VOC (for pressurerelief devices), with connectors and valves being $36,527and $43,111 per ton of VOC, respectively. For the modifiedsubpart VVa level of control at transmission and storagefacilities, cost-effectiveness ranged from $24,762 per tonof VOC (for open-ended lines) to $243,525 per ton of VOC(for pressure relief devices), with connectors and valvesbeing $42,140 and $40,593 per ton of VOC, respectively.Again, the modified subpart VVa level controls were morecost-effective for valves and less cost effective forconnectors than the subpart VVa level controls. This is dueto the low cost of monitoring connectors and the low VOCemissions from leaking connectors. For each of the non-gas processing segments, we alsoevaluated monthly optical gas imaging with annual Method 21check (Option 2). As discussed in secton VI.B.1, we hadpreviously determined that the VOC reductions achievedunder this option would be the same as for option 1 subpartVVa-level LDAR. In our evaluation of Option 2, we estimated This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

164. Page 164 of 604that a single optical imaging instrument could be used for160 well sites and 13 gathering and boosting stations,which means that the cost of the purchase or rental of thecamera would be spread across 173 facilities. For production sites, gathering and boosting stations,and transmission and storage facilities, we estimated thatoption 2 monthly optical gas imaging with annual Method 21check would have cost-effectiveness of $16,123, $10,095,and $19,715 per ton of VOC, respectively.13 The annual costs for option 1 and option 2 leakdetection and repair programs for production sitesassociated with a wellhead, gathering and boosting stationsand transmission and storage facilities were higher thanthose estimated for natural gas processing plants becausenatural gas processing plant annual costs are based on theincremental cost of implementing subpart VVa-levelstandards, whereas the other facilities are not currentlyregulated under an LDAR program. The currently unregulatedsites would be required to set up a new LDAR program;perform initial monitoring, tagging, logging and repairing13 Because optical gas imaging is used to view several pieces ofequipment at a facility at once to survey for leaks, optionsinvolving imaging are not amenable to a component by componentanalysis. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

165. Page 165 of 604of components; as well as planning and training personnelto implement the new LDAR program. In addition to options 1 and 2, we evaluated a thirdoption that consisted of monthly optical gas imagingwithout an annual Method 21 check. Because we were unableto estimate the VOC emissions achieved by an opticalimaging program alone, we were unable to estimate the cost-effectiveness of this option. However, we estimated theannual cost of the monthly optical gas imaging LDAR programat production sites, gathering and boosting stations, andtransmission and storage facilities to be $37,049, $86,135,and $45,080, respectively, based on camera purchase, or$32,693, $81,780, and $40,629, respectively, based oncamera rental. Finally, we evaluated a fourth option similar to thethird option except that the optical gas imaging would beperformed annually rather than monthly. For this option, weestimated the annual cost for production sites, gatheringand boosting stations, and transmission and storagefacilities to be $30,740, $64,416, and $24,031,respectively, based on camera purchase, or $26,341,$60,017, and $19,493, respectively, based on camera rental. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

166. Page 166 of 604 We request comment on the applicability of a leakdetection and repair program based solely on the use ofoptical imaging or other technologies. Of most use to uswould be information on the effectiveness of advancedmeasurement technologies to detect and repair small leakson the same order or smaller as specified in the VVaequipment leak requirements and the effects of increasedfrequency of and associated leak detection, recording, andrepair practices. Based on the evaluation described above, we believethat neither option 1 nor option 2 is cost effective forreducing fugitive VOC emissions from equipment leaks atsites, gathering and boosting stations, and transmissionand storage facilities. For options 3 and 4, we were unableto estimate their cost effectiveness and, therefore, couldnot identify either of these two options as BSER foraddressing equipment leak of VOC at production facilitiesassociated with wellheads, at gathering and boostingstations or at gas transmission and storage facilities. Weare, therefore, not proposing NSPS for addressing VOCemissions from equipment leaks at these facilities.5. What are the SSM provisions? The EPA is proposing standards in this rule that apply This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

167. Page 167 of 604at all times, including during periods of startup orshutdown, and periods of malfunction. In proposing thestandards in this rule, the EPA has taken into accountstartup and shutdown periods. The General Provisions in 40 CFR part 60 requirefacilities to keep records of the occurrence and durationof any startup, shutdown or malfunction (40 CFR 60.7(b))and either report to the EPA any period of excess emissionsthat occurs during periods of SSM (40 CFR 60.7(c)(2)) orreport that no excess emissions occurred (40 CFR60.7(c)(4)). Thus, any comments that contend that sourcescannot meet the proposed standard during startup andshutdown periods should provide data and other specificssupporting their claim. Periods of startup, normal operations and shutdown areall predictable and routine aspects of a source’soperations. However, by contrast, malfunction is defined asa “sudden, infrequent, and not reasonably preventablefailure of air pollution control and monitoring equipment,process equipment or a process to operate in a normal orusual manner...” (40 CFR 60.2.) The EPA has determined thatmalfunctions should not be viewed as a distinct operatingmode and, therefore, any emissions that occur at such times This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

168. Page 168 of 604do not need to be factored into development of CAA section111 standards. Further, nothing in CAA section 111 or incase law requires that the EPA anticipate and account forthe innumerable types of potential malfunction events insetting emission standards. See, Weyerhaeuser v Costle, 590F.2d 1011, 1058 (D.C. Cir. 1978) (“In the nature of things,no general limit, individual permit, or even any upsetprovision can anticipate all upset situations. After acertain point, the transgression of regulatory limitscaused by ‘uncontrollable acts of third parties,’ such asstrikes, sabotage, operator intoxication or insanity, and avariety of other eventualities, must be a matter for theadministrative exercise of case-by-case enforcementdiscretion, not for specification in advance byregulation.”), and, therefore, any emissions that occur atsuch times do not need to be factored into development ofCAA section 111 standards. Further, it is reasonable to interpret CAA section 111as not requiring the EPA to account for malfunctions insetting emissions standards. For example, we note that CAAsection 111 provides that the EPA set standards ofperformance which reflect the degree of emission limitationachievable through “the application of the best system of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

169. Page 169 of 604emission reduction” that the EPA determines is adequatelydemonstrated. Applying the concept of “the application ofthe best system of emission reduction” to periods duringwhich a source is malfunctioning presents difficulties. The“application of the best system of emission reduction” ismore appropriately understood to include operating units insuch a way as to avoid malfunctions. Moreover, even if malfunctions were considered adistinct operating mode, we believe it would beimpracticable to take malfunctions into account in settingCAA section 111 standards for affected facilities under 40CFR part 60, subpart OOOO. As noted above, by definition,malfunctions are sudden and unexpected events and it wouldbe difficult to set a standard that takes into account themyriad different types of malfunctions that can occuracross all sources in the category. Moreover, malfunctionscan vary in frequency, degree and duration, furthercomplicating standard setting. In the event that a source fails to comply with theapplicable CAA section 111 standards as a result of amalfunction event, the EPA would determine an appropriateresponse based on, among other things, the good faithefforts of the source to minimize emissions during This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

170. Page 170 of 604malfunction periods, including preventative and correctiveactions, as well as root cause analyses to ascertain andrectify excess emissions. The EPA would also considerwhether the source’s failure to comply with the CAA section111 standard was, in fact, “sudden, infrequent, notreasonably preventable” and was not instead “caused in partby poor maintenance or careless operation.” 40 CFR 60.2(definition of malfunction). Finally, the EPA recognizes that even equipment thatis properly designed and maintained can sometimes fail.Such failure can sometimes cause an exceedance of therelevant emission standard (See, e.g., State ImplementationPlans: Policy Regarding Excessive Emissions DuringMalfunctions, Startup, and Shutdown (September 20, 1999);Policy on Excess Emissions During Startup, Shutdown,Maintenance, and Malfunctions (February 15, 1983)). The EPAis, therefore, proposing to add an affirmative defense tocivil penalties for exceedances of emission limits that arecaused by malfunctions. See 40 CFR 60.41Da (defining“affirmative defense” to mean, in the context of anenforcement proceeding, a response or defense put forwardby a defendant, regarding which the defendant has theburden of proof and the merits of which are independently This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

171. Page 171 of 604and objectively evaluated in a judicial or administrativeproceeding). We also are proposing other regulatoryprovisions to specify the elements that are necessary toestablish this affirmative defense; the source must proveby a preponderance of the evidence that it has met all ofthe elements set forth in 40 CFR 60.46Da. (See 40 CFR22.24). These criteria ensure that the affirmative defenseis available only where the event that causes an exceedanceof the emission limit meets the narrow definition ofmalfunction in 40 CFR 60.2 (sudden, infrequent, notreasonably preventable and not caused by poor maintenanceand or careless operation). For example, to successfullyassert the affirmative defense, the source must prove by apreponderance of the evidence that excess emissions “[w]erecaused by a sudden, infrequent, and unavoidable failure ofair pollution control and monitoring equipment, processequipment, or a process to operate in a normal or usualmanner...” The criteria also are designed to ensure thatsteps are taken to correct the malfunction, to minimizeemissions in accordance with 40 CFR 60.40Da and to preventfuture malfunctions. For example, the source would have toprove by a preponderance of the evidence that “[r]epairswere made as expeditiously as possible when the applicable This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

172. Page 172 of 604emission limitations were being exceeded...” and that“[a]ll possible steps were taken to minimize the impact ofthe excess emissions on ambient air quality, theenvironment and human health...” In any judicial oradministrative proceeding, the Administrator may challengethe assertion of the affirmative defense and, if therespondent has not met the burden of proving all of therequirements in the affirmative defense, appropriatepenalties may be assessed in accordance with CAA section113 (see also 40 CFR part 22.77).VII. Rationale for Proposed Action for NESHAPA. What data were used for the NESHAP analyses? To perform the technology review and residual riskanalysis for the two NESHAP, we created a comprehensivedataset (i.e., the MACT dataset). This dataset was based onthe EPA’s 2005 National Emissions Inventory (NEI). The NEIdatabase contains information about sources that emitcriteria air pollutants and their precursors and HAP. Thedatabase includes estimates of annual air pollutantemissions from point, nonpoint and mobile sources in the 50states, the District of Columbia, Puerto Rico and theVirgin Islands. The EPA collects information about sources This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

173. Page 173 of 604and releases an updated version of the NEI database every 3years. The NEI database is compiled from these primarysources: • Emissions inventories compiled by state and local environmental agencies • Databases related to the EPAs MACT programs • Toxics Release Inventory data • For electric generating units, the EPAs Emission Tracking System/CEM data and United States Department of Energy (DOE) fuel use data • For onroad sources, the United States Federal Highway Administration’s estimate of vehicle miles traveled and emission factors from the EPAs MOBILE computer model • For nonroad sources, the EPAs NONROAD computer model • Emissions inventories from previous years, if states do not submit current data To concentrate on only records pertaining to the oiland natural gas industry sector, data were extracted usingtwo criteria. First, we specified that all facilitiescontaining codes identifying the Oil and Natural GasProduction and the Natural Gas Transmission and Storage This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

174. Page 174 of 604MACT source categories (MACT codes 0501 and 0504,respectively). Second, we extracted facilities identifiedwith the following NAICS codes: 211*** (Oil and GasExtraction), 221210 (Natural Gas Distribution), 4861**(Pipeline Transportation of Crude Oil), and 4862**(Pipeline Transportation of Natural Gas). Once the datawere extracted, we reviewed the Source Classification Codes(SCC) to assess whether there were any records included inthe dataset that were clearly not a part of the oil andnatural gas sector. Our review of the SCC also includedassigning each SCC to an “Emission Process Group” thatrepresents emission point types within the oil and naturalgas sector. Since these MACT standards only apply to majorsources, only facilities designated as major sources in theNEI were extracted. In the NEI, sources are identified asmajor if the facility-wide emissions are greater than 10tpy for any single HAP or 25 tpy for any combination ofHAP. We believe that this may overestimate the number ofmajor sources in the oil and natural gas sector because itdoes not take into account the limitations set forth in theCAA regarding aggregation of emissions from wells andassociated equipment in determining major source status. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

175. Page 175 of 604 The final dataset contained a total of 1,311 majorsources in the oil and natural gas sector; 990 in Oil andNatural Gas Production, and 321 in Natural Gas Transmissionand Storage. To assess how representative this number offacilities was, we obtained information on the number ofsubject facilities for both MACT standards from theEnforcement and Compliance History Online (ECHO) database.The ECHO database is a web-based tool (http://www.epa-echo.gov/echo/index.html) that provides public access tocompliance and enforcement information for approximately800,000 EPA-regulated facilities. The ECHO database allowsusers to find permit, inspection, violation, enforcementaction and penalty information covering the past 3 years.The site includes facilities regulated as CAA stationarysources, as well as Clean Water Act direct dischargers, andResource Conservation and Recovery Act hazardous wastegenerators/handlers. The data in the ECHO database areupdated monthly. We performed a query on the ECHO database requestingrecords for major sources, with NAICS codes 211*, 221210,4861* and 4862*, with information for MACT. The ECHOdatabase query identified records for a total of 555facilities, 269 in the Oil and Natural Gas Production This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

176. Page 176 of 604source category (NAICS 211* and 221210) and 286 in theNatural Gas Transmission and Storage source category (NAICS4861* and 4862*). This comparison leads us to concludethat, for the Natural Gas Transmission and Storage segment,the NEI database is representative of the number of sourcessubject to the rule. For the Oil and Natural Gas Productionsource category, it confirms our assumption that the NEIdataset contains more facilities than are subject to therule. However, this provides a conservative overestimate ofthe number of sources, which we believe is appropriate forour risk analyses. We are requesting that the public provide a detailedreview of the information in this dataset and providecomments and updated information where appropriate. SectionX of this preamble provides an explanation of how toprovide updated information for these datasets.B. What are the proposed decisions regarding certainunregulated emissions sources? In addition to actions relative to the technologyreview and risk reviews discussed below, we are proposing,pursuant to CAA sections 112(d)(2) and (3), MACT standardsfor glycol dehydrators and storage vessels for whichstandards were not previously developed. We are also This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

177. Page 177 of 604proposing changes that affect the definition of “associatedequipment” which could apply these MACT standards topreviously unregulated sources.1. Glycol Dehydrators Once natural gas has been separated from any liquidmaterials or products (e.g., crude oil, condensate orproduced water), residual entrained water is removed fromthe natural gas by dehydration. Dehydration is necessarybecause water vapor may form hydrates, which are ice-likestructures, and can cause corrosion in or plug equipmentlines. The most widely used natural gas dehydrationprocesses are glycol dehydration and solid desiccantdehydration. Solid desiccant dehydration, which istypically only used for lower throughputs, uses adsorptionto remove water and is not a source of HAP emissions. Glycol dehydration is an absorption process in which aliquid absorbent, glycol, directly contacts the natural gasstream and absorbs any entrained water vapor in a contacttower or absorption column. The majority of glycoldehydration units use triethylene glycol as the absorbent,but ethylene glycol and diethylene glycol are also used.The rich glycol, which has absorbed water vapor from thenatural gas stream, leaves the bottom of the absorption This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

178. Page 178 of 604column and is directed either to (1) a gas condensateglycol (GCG) separator (flash tank) and then a reboiler or(2) directly to a reboiler where the water is boiled off ofthe rich glycol. The regenerated glycol (lean glycol) iscirculated, by pump, into the absorption tower. The vaporgenerated in the reboiler is directed to the reboiler vent. The reboiler vent is a source of HAP emissions. In theglycol contact tower, glycol not only absorbs water, butalso absorbs selected hydrocarbons, including BTEX and n-hexane. The hydrocarbons are boiled off along with thewater in the reboiler and vented to the atmosphere or to acontrol device. The most commonly used control device is acondenser. Condensers not only reduce emissions, but alsorecover condensable hydrocarbon vapors that can berecovered and sold. In addition, the dry non-condensableoff-gas from the condenser may be used as fuel or recycledinto the production process or directed to a flare,incinerator or other combustion device. If present, the GCG separator (flash tank) is also apotential source of HAP emissions. Some glycol dehydrationunits use flash tanks prior to the reboiler to separateentrained gases, primarily methane and ethane from theglycol. The flash tank off-gases are typically recovered as This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

179. Page 179 of 604fuel or recycled to the natural gas production header.However, the flash tank may also be vented directly to theatmosphere. Flash tanks typically enhance the reboilercondenser’s emission reduction efficiency by reducing theconcentration of non-condensable gases present in thestream prior to being introduced into the condenser. In the development of the MACT standards for the twooil and natural gas source categories, the EPA created twosubcategories of glycol dehydrators based on actual annualaverage natural gas flowrate and actual average benzeneemissions. Under 40 CFR part 63, subpart HH, (the Oil andNatural Gas Production NESHAP), the EPA established MACTstandards for glycol dehydration units with an actualannual average natural gas flowrate greater than or equalto 85,000 scmd and actual average benzene emissions greaterthan or equal to 0.90 Mg/yr (40 CFR 63.765(a)). The EPA didnot establish standards for the other subcategory, whichconsists of glycol dehydration units that are below theflowrate and emission thresholds specified in subpart HH.Similarly, under 40 CFR part 63, subpart HHH (the NaturalGas Transmission and Storage NESHAP), the EPA establishedMACT standards for the subcategory of glycol dehydrationunits with an actual annual average natural gas flowrate This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

180. Page 180 of 604greater than or equal to 283,000 scmd and actual averagebenzene emissions greater than or equal to 0.90 Mg/yr, butdid not establish standards for the other subcategory,which consists of glycol dehydration units that are belowthe flowrate and emission thresholds specified in subpartHHH. As mentioned above, we refer to these unregulateddehydration units in both subparts HH and HHH as “smalldehydrators” in this proposed rule. The EPA is proposing emission standards for thesesubcategories of small dehydrators (i.e., those dehydratorswith an actual annual average natural gas flowrate lessthan 85,000 scmd at production sites or 283,000 scmd atnatural gas transmission and storage sites, or actualaverage benzene emissions less than 0.9 Mg/yr). Because wedo not have any new emissions data concerning theseemission points, we evaluated the dataset collected fromindustry during the development of the original MACTstandards (legacy docket A-94-04, item II-B-01, disk 1 foroil and natural gas production facilities; and items IV-G-24, 26, 27, 30 and 31 for natural gas transmission andstorage facilities). We believe this dataset isrepresentative of currently operating glycol dehydratorsbecause it contains information for a varied group of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

181. Page 181 of 604sources (i.e., units owned by different companies, locatedin different states, representing a range of gascompositions and emission controls) and that the processeshave not changed significantly since the data werecollected. In the Oil and Natural Gas Production source category,there were 91 glycol dehydration units with throughput andemissions data identified that would be classified as smallglycol dehydration units. We evaluated the possibility ofestablishing a MACT floor as a Mg/yr limit. However, due tovariability of gas throughput and inlet gas composition, wecould not properly identify the best performing units byonly considering emissions. To allow us to normalize theemissions for a more accurate determination of the bestperforming sources, we created an emission factor in termsof grams BTEX/scm-ppmv for each facility. The emissionfactor reflects the facility’s emission level, taking intoconsideration its natural gas throughput and inlet naturalgas BTEX concentration. To determine the MACT floor for theexisting dehydrators, we ranked each unit from lowest tohighest, based on their emission factor, to determine thefacilities in the top 12 percent of the dataset. The MACTfloor was an emission factor of 1.10x10-4 grams BTEX/scm- This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

182. Page 182 of 604ppmv. To meet this level of emissions, we anticipate thatsources will use a variety of options, including, but notlimited to, routing emissions to a condenser or to acombustion device. We also considered beyond-the-floor options for theexisting sources, as required by section 112 (d)(2) of theCAA. To achieve further reductions beyond the MACT floorlevel of control, sources would have to install anadditional add-on control device, most likely a combustiondevice. Assuming the MACT floor control device is acombustion device, which generally achieves at least a 95-percent HAP reduction, then less than 5 percent of theinitial HAP emissions remain. Installing a second devicewould involve the same costs as the first control, butwould only achieve 1/20 of the reduction (i.e., reducingthe remaining 5 percent by another 95 percent represents a4.49-percent reduction of the initial, uncontrolledemissions, which is 1/20 of the 95-percent reductionachieved with the first control). Based on the $8,360/Mgcost effectiveness of the floor level of control, weestimate that the incremental cost effectiveness of thesecond control to be $167,200/Mg. We do not believe thiscost to be reasonable given the level of emission This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

183. Page 183 of 604reduction. We are, therefore, proposing an emissionstandard for existing small dehydrators that reflects theMACT floor. For new small glycol dehydrators in the Oil andNatural Gas Production source category, based on ourperformance ranking, the best performing source has anemission factor of 4.66x10-6 grams BTEX/scm-ppmv. To meetthis level of emissions, we anticipate that sources willuse a variety of options, including, but not limited to,routing emissions to a condenser or to a combustion device.The consideration of beyond-the-floor options for new smalldehydrators would be the same as for existing smalldehydrators, and, as stated above, we do not believe a costof $167,200/Mg to be reasonable given the level of emissionreduction. We are, therefore, proposing a MACT standard fornew small dehydrators that reflects the MACT floor level ofcontrol. Under our proposal, a small dehydrator’s actual MACTemission limit would be determined by multiplying the MACTfloor emission factor in g BTEX/scm-ppmv by its unit-specific incoming natural gas throughput and BTEXconcentration for the dehydrator. A formula is provided in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

184. Page 184 of 60440 CFR 63.765(b)(1)(iii) to calculate the MACT limit as anannual value. In the Natural Gas Transmission and Storage sourcecategory, there were 16 facilities for which throughput andemissions data were available that would be classified assmall glycol dehydration units. Since the number of unitswas less than 30, the MACT floor for existing sources wasbased on the top five performing units. Using the sameemission factor concept, we determined that the MACT floorfor existing sources is an emission factor equal to 6.42x10-5 grams BTEX/scm-ppmv. To meet this level of emissions, weanticipate that sources will use a variety of options,including, but not limited to, routing emissions to acondenser or to a combustion device. We also considered beyond-the-floor options for theexisting small dehydrators as required by section 112(d)(2) of the CAA. To achieve further reductions beyond theMACT floor level of control, sources would have to installan additional add-on control device, most likely acombustion device. Assuming the MACT floor control deviceis a combustion device, which generally achieves at least a95-percent HAP reduction, then less than 5 percent of theinitial HAP emissions remain. Installing a second device This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

185. Page 185 of 604would involve the same costs as the first control device,but would only achieve 1/20 of the reduction (i.e.,reducing the remaining 5 percent by another 95 percentrepresents a 4.49-percent reduction of the initial,uncontrolled emissions, which is 1/20 of the 95-percentreduction achieved with the first control). Based on the$1,650/Mg cost effectiveness of the floor level of control,we estimate that the incremental cost effectiveness of thesecond control to be $33,000/Mg. We do not believe thiscost to be reasonable given the level of emissionreduction. We are, therefore, proposing an emissionstandard for existing small dehydrators that reflects theMACT floor. For new small glycol dehydrators, based on ourperformance ranking, the best performing source has anemission factor of 1.10x10-5 grams BTEX/scm-ppmv. To meetthis level of emissions, we anticipate that sources willuse a variety of options, including, but not limited to,routing emissions to a condenser or to a combustion device.The consideration of beyond-the-floor options for new smalldehydrators would be the same as for existing smalldehydrators, and, as stated above, we do not believe a costof $33,000/Mg to be reasonable given the level of emission This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

186. Page 186 of 604reduction. We are, therefore, proposing an emissionstandard for new sources that reflects the MACT floor levelof control. Under our proposal, a source’s actual MACT emissionslimit would be determined by multiplying this emissionfactor by their unit-specific incoming natural gasthroughput and BTEX concentration for the dehydrator. Aformula is provided in 40 CFR 63.1275(b)(1)(iii) tocalculate the limit as an annual value. As discussed below, we are proposing that, with theremoval of the 1-ton alternative compliance option from theexisting standards for glycol dehydrators, the MACT forthese two source categories would provide an ample marginof safety to protect public health. We, therefore, maintainthat, after the implementation of the small dehydratorstandards discussed above, these MACT will continue toprovide an ample margin of safety to protect public health.Consequently, we do not believe it will be necessary toconduct another residual risk review under CAA section112(f) for these two source categories 8 years followingpromulgation of the small dehydrator standards merely dueto the addition of these new MACT requirements.2. Storage Vessels This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

187. Page 187 of 604 Crude oil, condensate and produced water are typicallystored in fixed-roof storage vessels. Some vessels used forstoring produced water may be open-top tanks. Thesevessels, which are operated at or near atmospheric pressureconditions, are typically located at tank batteries. A tankbattery refers to the collection of process components usedto separate, treat and store crude oil, condensate, naturalgas and produced water. The extracted products fromproductions wells enter the tank battery through theproduction header, which may collect product from manywells. Emissions from storage vessels are a result ofworking, breathing and flash losses. Working losses occurdue to the emptying and filling of storage tanks. Breathinglosses are the release of gas associated with dailytemperature fluctuations and other equilibrium effects.Flash losses occur when a liquid with entrained gases istransferred from a vessel with higher pressure to a vesselwith lower pressure, thus, allowing entrained gases or aportion of the liquid to vaporize or flash. In the oil andnatural gas production segment, flashing losses occur whenlive crude oils or condensates flow into a storage tankfrom a processing vessel operated at a higher pressure. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

188. Page 188 of 604Typically, the larger the pressure drop, the more flashingemission will occur in the storage stage. Temperature ofthe liquid may also influence the amount of flashemissions. In the Oil and Natural Gas Production NESHAP (40 CFRpart 63, subpart HH), the MACT standards for storagevessels apply only to those with the PFE. Storage vesselswith the PFE are defined as storage vessels that containhydrocarbon liquids that meet the following criteria: • A stock tank gas to oil ratio (GOR) greater than or equal to 0.31 cubic meters per liter (m3/liter); and • An American Petroleum Institute (API) gravity greater than or equal to 40 degrees; and • An actual annual average hydrocarbon liquid throughput greater than or equal to 79,500 liters per day (liter/day). Accordingly, there is no emission limit in theexisting MACT for storage vessels without the PFE. However,the MACT analysis performed at the time indicates that theMACT floor was based on all storage vessels, not just thosevessels with flash emissions. See, Recommendation of MACT This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

189. Page 189 of 604Floor Levels for HAP Emission Points at Major Sources inthe Oil and Natural Gas Production Source Category,(September 23, 1997, Docket A-94-04, Item II-A-07). We,therefore, propose to apply the existing MACT for storagevessels with PFE to all storage vessels (i.e., storagevessels with the PFE, as well as those without the PFE).3. Definition of Associated EquipmentCAA section 112(n)(4)(A) provides: “Notwithstanding the provisions of subsection (a), emissions from any oil or gas exploration or production well (with its associated equipment) and emission from any pipeline compressor or pump station shall not be aggregated with emissions from other similar units, whether or not such units are in contiguous area or under common control, to determine whether such units or stations are major sources.” As stated above, the CAA prevents aggregation of HAPemissions from wells and associated equipment in makingmajor source determinations. In the absence of clearguidance in the statute on what constitutes “associatedequipment,” the EPA sought to define “associated equipment”in a way that recognizes the need to implement relief forthis industry as Congress intended and that also allow for This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

190. Page 190 of 604the appropriate regulation of significant emission points.64 FR at 32619. Accordingly, in the existing Oil andNatural Gas Production NESHAP (1998 and 1999 NESHAP), theEPA defined “associated equipment” to exclude glycoldehydration units and storage vessels with PFE (thusallowing their emissions to be included in determiningmajor source status) because EPA identified these sourcesas substantial contributors to HAP emissions. Id. EPAexplained in that NESHAP that, because a single storagevessel with flash emissions may emit several Mg of HAP peryear and individual glycol dehydrators may emit above themajor source level, storage vessels with PFE and glycoldehydrators are large individual sources of HAP, 63 FR6288, 6301 (1998). The EPA therefore considered theseemission sources substantial contributors to HAP emissionsand excluded them from the definition of “associatedequipment.” 64 FR at 32619. We have recently examined HAPemissions from storage vessels without flash emissions andfound that these emissions are significant and comparableto those vessels with flash emissions. For example, onestorage vessel with an API gravity of 30 degrees and a GORof 2.09x10-3 m3/liter with a throughput of 79,500 liter/day This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

191. Page 191 of 604had HAP emissions of 9.91 Mg/yr, including 9.45 Mg/yr of n-hexane. Because storage vessels without the PFE can havesignificant emissions at levels that are comparable toemissions from storage vessels with the PFE, there is noappreciable difference between storage vessels with the PFEand those without the PFE for purposes of defining“associated equipment.” We are, therefore, proposing toamend the associated equipment definition to exclude allstorage vessels and not just storage vessels with the PFE.C. How did we perform the risk assessment and what are theresults and proposed decisions?1. How did we estimate risks posed by the sourcecategories? The EPA conducted risk assessments that providedestimates for each source in a category of the MIR posed bythe HAP emissions, the HI for chronic exposures to HAP withthe potential to cause noncancer health effects, and thehazard quotient (HQ) for acute exposures to HAP with thepotential to cause noncancer health effects. Theassessments also provided estimates of the distribution ofcancer risks within the exposed populations, cancerincidence and an evaluation of the potential for adverse This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

192. Page 192 of 604environmental effects for each source category. The riskassessments consisted of seven primary steps, as discussedbelow. The docket for this rulemaking contains thefollowing document which provides more information on therisk assessment inputs and models: Draft Residual RiskAssessment for the Oil and Gas Production and Natural GasTransmission and Storage Source Categories. The methodsused to assess risks (as described in the seven primarysteps below) are consistent with those peer-reviewed by apanel of the EPA’s Science Advisory Board (SAB) in 2009 anddescribed in their peer review report issued in 201014; theyare also consistent with the key recommendations containedin that report.a. Establishing the Nature and Magnitude of ActualEmissions and Identifying the Emissions ReleaseCharacteristics As discussed in section VII.A of this preamble, weused a dataset based on the 2005 NEI as the basis for therisk assessment. In addition to the quality assurance (QA)of the facilities contained in the dataset, we also checked14 U.S. EPA SAB. Risk and Technology Review (RTR) Risk AssessmentMethodologies: For Review by the EPA’s Science Advisory Boardwith Case Studies – MACT I Petroleum Refining Sources andPortland Cement Manufacturing, May 2010. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

193. Page 193 of 604the coordinates of every facility in the dataset throughvisual observations using tools such as GoogleEarth andArcView. Where coordinates were found to be incorrect, weidentified and corrected them to the extent possible. Wealso performed QA of the emissions data and releasecharacteristics to ensure there were no outliers.b. Establishing the Relationship Between Actual Emissionsand MACT-Allowable Emissions Levels The available emissions data in the MACT datasetrepresent the estimates of mass of emissions actuallyemitted during the specified annual time period. These“actual” emission levels are often lower than the emissionlevels that a facility might be allowed to emit and stillcomply with the MACT standards. The emissions level allowedto be emitted by the MACT standards is referred to as the“MACT-allowable” emissions level. This represents thehighest emissions level that could be emitted by thefacility without violating the MACT standards. We discussed the use of both MACT-allowable and actualemissions in the final Coke Oven Batteries residual riskrule (70 FR 19998–19999, April 15, 2005) and in theproposed and final Hazardous Organic NESHAP residual riskrules (71 FR 34428, June 14, 2006, and 71 FR 76609, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

194. Page 194 of 604December 21, 2006, respectively). In those previousactions, we noted that assessing the risks at the MACT-allowable level is inherently reasonable since these risksreflect the maximum level sources could emit and stillcomply with national emission standards. But we alsoexplained that it is reasonable to consider actualemissions, where such data are available, in both steps ofthe risk analysis, in accordance with the Benzene NESHAP.(54 FR 38044, September 14, 1989.) To estimate emissions at the MACT-allowable level, wedeveloped a ratio of MACT-allowable to actual emissions foreach emissions source type in each source category, basedon the level of control required by the MACT standardscompared to the level of reported actual emissions andavailable information on the level of control achieved bythe emissions controls in use.c. Conducting Dispersion Modeling, Determining InhalationExposures and Estimating Individual and PopulationInhalation Risks Both long-term and short-term inhalation exposureconcentrations and health risks from each source in thesource categories addressed in this proposal were estimatedusing the Human Exposure Model (HEM) (Community and Sector This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

195. Page 195 of 604HEM–3 version 1.1.0). The HEM–3 performs three primary riskassessment activities: (1) Conducting dispersion modelingto estimate the concentrations of HAP in ambient air, (2)estimating long-term and short-term inhalation exposures toindividuals residing within 50 km of the modeled sourcesand (3) estimating individual and population-levelinhalation risks using the exposure estimates andquantitative dose-response information. The dispersion model used by HEM–3 is AERMOD, which isone of the EPA’s preferred models for assessing pollutant 15concentrations from industrial facilities. To perform thedispersion modeling and to develop the preliminary riskestimates, HEM–3 draws on three data libraries. The firstis a library of meteorological data, which is used fordispersion calculations. This library includes 1 year ofhourly surface and upper air observations for more than 158meteorological stations, selected to provide coverage ofthe United States and Puerto Rico. A second library of 16United States Census Bureau census block internal point15 U.S. EPA. Revision to the Guideline on Air Quality Models:Adoption of a Preferred General Purpose (Flat and ComplexTerrain) Dispersion Model and Other Revisions (70 FR 68218,November 9, 2005).16 A census block is generally the smallest geographic area for This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

196. Page 196 of 604locations and populations provides the basis of humanexposure calculations (Census, 2000). In addition, for eachcensus block, the census library includes the elevation andcontrolling hill height, which are also used in dispersioncalculations. A third library of pollutant unit riskfactors and other health benchmarks is used to estimatehealth risks. These risk factors and health benchmarks arethe latest values recommended by the EPA for HAP and othertoxic air pollutants. These values are available athttp://www.epa.gov/ttn/atw/toxsource/summary.html and arediscussed in more detail later in this section. In developing the risk assessment for chronicexposures, we used the estimated annual average ambient airconcentration of each of the HAP emitted by each source forwhich we have emissions data in the source category. Theair concentrations at each nearby census block centroidwere used as a surrogate for the chronic inhalationexposure concentration for all the people who reside inthat census block. We calculated the MIR for each facilityas the cancer risk associated with a continuous lifetime(24 hours per day, 7 days per week, and 52 weeks per yearfor a 70-year period) exposure to the maximum concentrationwhich census statistics are tabulated. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

197. Page 197 of 604at the centroid of an inhabited census block. Individualcancer risks were calculated by multiplying the estimatedlifetime exposure to the ambient concentration of each ofthe HAP (in micrograms per cubic meter) by its unit riskestimate (URE), which is an upper bound estimate of anindividual’s probability of contracting cancer over alifetime of exposure to a concentration of 1 microgram ofthe pollutant per cubic meter of air. For residual riskassessments, we generally use URE values from the EPA’sIntegrated Risk Information System (IRIS). For carcinogenicpollutants without the EPA IRIS values, we look to otherreputable sources of cancer dose-response values, oftenusing California EPA (CalEPA) URE values, where available.In cases where new, scientifically credible dose-responsevalues have been developed in a manner consistent with theEPA guidelines and have undergone a peer review processsimilar to that used by the EPA, we may use such dose-response values in place of or in addition to other values,if appropriate. Formaldehyde is a unique case. In 2004, the EPAdetermined that the Chemical Industry Institute ofToxicology (CIIT) cancer dose-response value forformaldehyde (5.5 x 10-9 per μg/m3) was based on better This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

198. Page 198 of 604science than the IRIS cancer dose-response value (1.3 x 10-5 per μg/m3) and we switched from using the IRIS value tothe CIIT value in risk assessments supporting regulatoryactions. However, subsequent research published by the EPAsuggests that the CIIT model was not appropriate and in2010 the EPA returned to using the 1991 IRIS value, whichis more health protective.17 The EPA has been working onrevising the formaldehyde IRIS assessment and the NationalAcademy of Sciences (NAS) completed its review of the EPA’sdraft in May of 2011. EPA is reviewing the public commentsand the NAS independent scientific peer review, and thedraft IRIS assessment will be revised and the finalassessment will be posted on the IRIS database. In theinterim, we will present findings using the 1991 IRIS valueas a primary estimate, and may also consider otherinformation as the science evolves. In the case of benzene, the high end of the reportedcancer URE range was used in our assessments to provide aconservative estimate of potential cancer risks. Use of thehigh end of the range provides risk estimates that are17 For details on the justification for this decision, see thememorandum in the docket from Peter Preuss to Steve Pageentitled, Recommendation for Formaldehyde Inhalation Cancer RiskValues, January 22, 2010. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

199. Page 199 of 604approximately 3.5 times higher than use of the equally-plausible low end value. We also evaluated the impact ofusing the low end of the URE range on our risk results. We also note that polycyclic organic matter (POM), acarcinogenic HAP with a mutagenic mode of action, is 18emitted by some of the facilities in these two categories. 19For this compound group, the age-dependent adjustmentfactors (ADAF) described in the EPA’s Supplemental Guidancefor Assessing Susceptibility from Early-Life Exposure to 20Carcinogens were applied. This adjustment has the effectof increasing the estimated lifetime risks for POM by afactor of 1.6. In addition, although only a small fractionof the total POM emissions were not reported as individualcompounds, the EPA expresses carcinogenic potency forcompounds in this group in terms of benzo[a]pyreneequivalence, based on evidence that carcinogenic POM has18 U.S. EPA. Performing risk assessments that include carcinogensdescribed in the Supplemental Guidance as having a mutagenic modeof action. Science Policy Council Cancer GuidelinesImplementation Work Group Communication II: Memo from W.H.Farland, dated October 4, 2005.19 See the Risk Assessment for Source Categories documentavailable in the docket for a list of HAP with a mutagenic modeof action.20 U.S. EPA. Supplemental Guidance for Assessing Early-LifeExposure to Carcinogens. EPA/630/R-03/003F, 2005.http://www.epa.gov/ttn/atw/childrens_supplement_final.pdf. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

200. Page 200 of 604the same mutagenic mechanism of action as benzo[a]pyrene.For this reason, the EPA’s Science 21Policy Council recommends applying the SupplementalGuidance to all carcinogenic polycyclic aromatichydrocarbons for which risk estimates are based on relativepotency. Accordingly, we have applied the ADAF to thebenzo[a]pyrene equivalent portion of all POM mixtures. Incremental individual lifetime cancer risksassociated with emissions from the source category wereestimated as the sum of the risks for each of thecarcinogenic HAP (including those classified ascarcinogenic to humans, likely to be carcinogenic to humans 22and suggestive evidence of carcinogenic potential ) emittedby the modeled source. Cancer incidence and thedistribution of individual cancer risks for the population21 U.S. EPA. Science Policy Council Cancer GuidelinesImplementation Workgroup Communication II: Memo from W.H.Farland, dated June 14, 2006.22 These classifications also coincide with the terms "knowncarcinogen, probable carcinogen and possible carcinogen,"respectively, which are the terms advocated in the EPAs previousGuidelines for Carcinogen Risk Assessment, published in 1986 (51FR 33992, September 24, 1986). Summing the risks of theseindividual compounds to obtain the cumulative cancer risks is anapproach that was recommended by the EPAs SAB in their 2002 peerreview of EPAs NATA entitled, NATA - Evaluating the National-scale Air Toxics Assessment 1996 Data -- an SAB Advisory,available at:http://yosemite.epa.gov/sab/sabproduct.nsf/214C6E915BB04E14852570CA007A682C/$File/ecadv02001.pdf. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

201. Page 201 of 604within 50 km of any source were also estimated for thesource category as part of these assessments by summingindividual risks. A distance of 50 km is consistent withboth the analysis supporting the 1989 Benzene NESHAP (54 FR38044) and the limitations of Gaussian dispersion models,including AERMOD. To assess risk of noncancer health effects fromchronic exposures, we summed the HQ for each of the HAPthat affects a common target organ system to obtain the HIfor that target organ system (or target organ-specific HI,TOSHI). The HQ for chronic exposures is the estimatedchronic exposure divided by the chronic reference level,which is either the EPA reference concentration (RfC),defined as “an estimate (with uncertainty spanning perhapsan order of magnitude) of a continuous inhalation exposureto the human population (including sensitive subgroups)that is likely to be without an appreciable risk ofdeleterious effects during a lifetime,” or, in cases wherean RfC from the EPA’s IRIS database is not available, theEPA will utilize the following prioritized sources for ourchronic dose-response values: (1) The Agency for ToxicSubstances and Disease Registry Minimum Risk Level, whichis defined as “an estimate of daily human exposure to a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

202. Page 202 of 604substance that is likely to be without an appreciable riskof adverse effects (other than cancer) over a specifiedduration of exposure”; (2) the CalEPA Chronic ReferenceExposure Level (REL), which is defined as “theconcentration level at or below which no adverse healtheffects are anticipated for a specified exposure duration”;and (3), as noted above, in cases where scientificallycredible dose-response values have been developed in amanner consistent with the EPA guidelines and haveundergone a peer review process similar to that used by theEPA, we may use those dose-response values in place of orin concert with other values. Screening estimates of acute exposures and risks werealso evaluated for each of the HAP at the point of highestoff-site exposure for each facility (i.e., not just thecensus block centroids), assuming that a person is locatedat this spot at a time when both the peak (hourly) emissionrate and worst-case dispersion conditions (1991 calendaryear data) occur. The acute HQ is the estimated acuteexposure divided by the acute dose-response value. In eachcase, acute HQ values were calculated using best available,short-term dose-response values. These acute dose-responsevalues, which are described below, include the acute REL, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

203. Page 203 of 604acute exposure guideline levels (AEGL) and emergencyresponse planning guidelines (ERPG) for 1-hour exposuredurations. As discussed below, we used conservativeassumptions for emission rates, meteorology and exposurelocation for our acute analysis. As described in the CalEPA’s Air Toxics Hot SpotsProgram Risk Assessment Guidelines, Part I, TheDetermination of Acute Reference Exposure Levels forAirborne Toxicants, an acute REL value(http://www.oehha.ca.gov/air/pdf/acuterel.pdf) is definedas “the concentration level at or below which no adversehealth effects are anticipated for a specified exposureduration.” Acute REL values are based on the mostsensitive, relevant, adverse health effect reported in themedical and toxicological literature. Acute REL values aredesigned to protect the most sensitive individuals in thepopulation by the inclusion of margins of safety. Sincemargins of safety are incorporated to address data gaps anduncertainties, exceeding the acute REL does notautomatically indicate an adverse health impact. AEGL values were derived in response torecommendations from the National Research Council (NRC).As described in Standing Operating Procedures (SOP) of the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

204. Page 204 of 604National Advisory Committee on Acute Exposure GuidelineLevels for Hazardous Substances 23(http://www.epa.gov/opptintr/aegl/pubs/sop.pdf), “theNRC’s previous name for acute exposure levels—communityemergency exposure levels—was replaced by the term AEGL toreflect the broad application of these values to planning,response, and prevention in the community, the workplace,transportation, the military, and the remediation ofSuperfund sites.” This document also states that AEGLvalues “represent threshold exposure limits for the generalpublic and are applicable to emergency exposures rangingfrom 10 minutes to eight hours.” The document lays out thepurpose and objectives of AEGL by stating (page 21) that“the primary purpose of the AEGL program and the NationalAdvisory Committee for Acute Exposure Guideline Levels forHazardous Substances is to develop guideline levels foronce-in-a-lifetime, short-term exposures to airborneconcentrations of acutely toxic, high-priority chemicals.”In detailing the intended application of AEGL values, thedocument states (page 31) that “[i]t is anticipated thatthe AEGL values will be used for regulatory and23 NAS, 2001. Standing Operating Procedures for Developing AcuteExposure Levels for Hazardous Chemicals, page 2. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

205. Page 205 of 604nonregulatory purposes by U.S. Federal and state agenciesand possibly the international community in conjunctionwith chemical emergency response, planning, and preventionprograms. More specifically, the AEGL values will be usedfor conducting various risk assessments to aid in thedevelopment of emergency preparedness and prevention plans,as well as real-time emergency response actions, foraccidental chemical releases at fixed facilities and fromtransport carriers.” The AEGL–1 value is then specifically defined as “theairborne concentration of a substance above which it ispredicted that the general population, includingsusceptible individuals, could experience notablediscomfort, irritation, or certain asymptomatic nonsensoryeffects. However, the effects are not disabling and aretransient and reversible upon cessation of exposure.” Thedocument also notes (page 3) that, “Airborne concentrationsbelow AEGL–1 represent exposure levels that can producemild and progressively increasing but transient andnondisabling odor, taste, and sensory irritation or certainasymptomatic, nonsensory effects.” Similarly, the documentdefines AEGL–2 values as “the airborne concentration(expressed as ppm or mg/m3) of a substance above which it This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

206. Page 206 of 604is predicted that the general population, includingsusceptible individuals, could experience irreversible orother serious, long-lasting adverse health effects or animpaired ability to escape.” ERPG values are derived for use in emergency response,as described in the American Industrial HygieneAssociation’s document entitled, Emergency ResponsePlanning Guidelines (ERPG) Procedures and Responsibilities(http://www.aiha.org/1documents/committees/ERPSOPs2006.pdf)which states that, “Emergency Response Planning Guidelineswere developed for emergency planning and are intended ashealth based guideline concentrations for single exposures 24to chemicals.” The ERPG–1 value is defined as “the maximumairborne concentration below which it is believed thatnearly all individuals could be exposed for up to 1 hourwithout experiencing other than mild transient adversehealth effects or without perceiving a clearly defined,objectionable odor.” Similarly, the ERPG–2 value is definedas “the maximum airborne concentration below which it isbelieved that nearly all individuals could be exposed forup to 1 hour without experiencing or developing24 ERP Committee Procedures and Responsibilities. November 1,2006. American Industrial Hygiene Association. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

207. Page 207 of 604irreversible or other serious health effects or symptomswhich could impair an individual’s ability to takeprotective action.” As can be seen from the definitions above, the AEGLand ERPG values include the similarly-defined severitylevels 1 and 2. For many chemicals, a severity level 1value AEGL or ERPG has not been developed; in theseinstances, higher severity level AEGL–2 or ERPG–2 valuesare compared to our modeled exposure levels to screen forpotential acute concerns. Acute REL values for 1-hour exposure durations aretypically lower than their corresponding AEGL–1 and ERPG–1values. Even though their definitions are slightlydifferent, AEGL–1 values are often the same as thecorresponding ERPG–1 values, and AEGL–2 values are oftenequal to ERPG–2 values. Maximum HQ values from our acutescreening risk assessments typically result when basingthem on the acute REL value for a particular pollutant. Incases where our maximum acute HQ value exceeds 1, we alsoreport the HQ value based on the next highest acute dose-response value (usually the AEGL–1 and/or the ERPG–1value). To develop screening estimates of acute exposures, we This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

208. Page 208 of 604developed estimates of maximum hourly emission rates bymultiplying the average actual annual hourly emission ratesby a factor to cover routinely variable emissions. We chosethe factor based on process knowledge and engineeringjudgment and with awareness of a Texas study of short-termemissions variability, which showed that most peak emissionevents, in a heavily-industrialized 4-county area (Harris,Galveston, Chambers and Brazoria Counties, Texas) were lessthan twice the annual average hourly emission rate. Thehighest peak emission event was 74 times the annual averagehourly emission rate, and the 99th percentile ratio of peakhourly emission rate to the annual average hourly emission 25rate was 9. This analysis is provided in Appendix 4 of theDraft Residual Risk Assessment for the Oil and GasProduction and Natural Gas Transmission and Storage SourceCategories, which is available in the docket for thisaction. Considering this analysis, unless specific processknowledge or data are available to provide an alternatevalue, to account for more than 99 percent of the peakhourly emissions, we apply a conservative screening25 Seehttp://www.tceq.state.tx.us/compliance/field_ops/eer/index.htmlor docket to access the source of these data. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

209. Page 209 of 604multiplication factor of 10 to the average annual hourlyemission rate in these acute exposure screeningassessments. The factor of 10 was used for both the Oil andNatural Gas Production and the Natural Gas Transmission andStorage source categories. In cases where acute HQ values from the screening stepwere less than or equal to 1, acute impacts were deemednegligible and no further analysis was performed. In caseswhere an acute HQ from the screening step was greater than1, additional site-specific data were considered to developa more refined estimate of the potential for acute impactsof concern. The data refinements employed for these sourcecategories consisted of using the site-specific facilitylayout to distinguish facility property from an area wherethe public could be exposed. These refinements arediscussed in the draft risk assessment document, which isavailable in the docket for each of these sourcecategories. Ideally, we would prefer to have continuousmeasurements over time to see how the emissions vary byeach hour over an entire year. Having a frequencydistribution of hourly emission rates over a year wouldallow us to perform a probabilistic analysis to estimatepotential threshold exceedances and their frequency of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

210. Page 210 of 604 occurrence. Such an evaluation could include a more complete statistical treatment of the key parameters and elements adopted in this screening analysis. However, we recognize that having this level of data is rare, hence our use of the multiplier approach. To better characterize the potential health risks associated with estimated acute exposures to HAP, and in response to a key recommendation from the SAB’s peer review of the EPA’s RTR risk assessment methodologies,26 we generally examine a wider range of available acute health metrics than we do for our chronic risk assessments. This is in response to the SAB’s acknowledgement that there are generally more data gaps and inconsistencies in acute reference values than there are in chronic reference values. Comparisons of the estimated maximum off-site 1- hour exposure levels are not typically made to occupational levels for the purpose of characterizing public health risks in RTR assessments. This is because they are developed for working age adults and are not generally considered protective for the general public. We note that26 The SAB peer review of RTR Risk Assessment Methodologies is available at: http://yosemite.epa.gov/sab/sabproduct.nsf/4AB3966E263D943A852577 1F00668381/$File/EPA-SAB-10-007-unsigned.pdf. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

211. Page 211 of 604occupational ceiling values are, for most chemicals, set atlevels higher than a 1-hour AEGL-1. As discussed in section VII.C.2 of this preamble, themaximum estimated worst-case 1-hour exposure to benzeneoutside the facility fence line for a facility in eithersource category is 12 mg/m3. This estimated exposure exceedsthe 6-hour REL by a factor of 9 (HQREL = 9), but issignificantly below the 1-hour AEGL-1 (HQAEGL-1 = 0.07).Although this worst-case exposure estimate does not exceedthe AEGL-1, we note here that it slightly exceeds workplaceceiling level guidelines designed to protect the workerpopulation for short duration (<15 minute) increases inexposure to benzene, as discussed below. The occupationalshort-term exposure limit (STEL) standard for benzenedeveloped by the Occupational Safety and HealthAdministration is 16 mg/m3, “as averaged over any 15-minuteperiod.”27 Occupational guideline STEL for exposures tobenzene have also been developed by the American Conferenceof Governmental Industrial Hygienists (ACGIH)28 for less27 29 CFR 1910.1028, Benzene. Available online athttp://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&p_id=10042.28 ACGIH (2001) Benzene. In Documentation of the TLVs® and BEIs®with Other Worldwide Occupational Exposure Values. ACGIH, 1300 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

212. Page 212 of 604than 15 minutes29 (ACGIH threshold limit value (TLV)-STELvalue of 8.0 mg/m3), and by the National Institute forOccupational Safety and Health (NIOSH)30 “for any 15 minuteperiod in a work day” (NIOSH REL-STEL of 3.2 mg/m3). Theseshorter duration occupational values indicate potentialconcerns regarding health effects at exposure levels belowthe 1-hour AEGL-1 value. We solicit comment on the use ofthe occupational values described above in theinterpretation of these worst-case acute screening exposureestimates.d. Conducting Multi-Pathway Exposure and Risk Modeling The potential for significant human health risks dueto exposures via routes other than inhalation (i.e., multi-pathway exposures) and the potential for adverseenvironmental impacts were evaluated in a three-stepprocess. In the first step, we determined whether anyfacilities emitted any HAP known to be PB-HAP (HAP known tobe persistent and bio-accumulative) in the environment.Kemper Meadow Drive, Cincinnati, OH 45240 (ISBN: 978-1-882417-74-2) and available online at http://www.acgih.org.29 The ACGIH definition of a TLV-STEL states that “Exposures abovethe TLV-TWA up to the TLV-STEL should be less than 15 minutes,should occur no more than four times per day, and there should beat least 60 minutes between successive exposures in this range.”30 NIOSH. Occupational Safety and Health Guideline for Benzene;http://www.cdc.gov/niosh/74-137.html. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

213. Page 213 of 604There are 14 PB-HAP compounds or compound classesidentified for this screening in the EPA’s Air Toxics RiskAssessment Library (available athttp://www.epa.gov/ttn/fera/risk_atra_vol1.html). They arecadmium compounds, chlordane, chlorinated dibenzodioxinsand furans, dichlorodiphenyldichloroethylene, heptachlor,hexachlorobenzene, hexachlorocyclohexane, lead compounds,mercury compounds, methoxychlor, polychlorinated biphenyls,POM, toxaphene and trifluralin. Since one or more of these PB-HAP are emitted by atleast one facility in both source categories, we proceededto the second step of the evaluation. In this step, wedetermined whether the facility-specific emission rates ofeach of the emitted PB–HAP were large enough to create thepotential for significant non-inhalation human orenvironmental risks under reasonable worst-case conditions.To facilitate this step, we have developed emission ratethresholds for each PB–HAP using a hypothetical worst-casescreening exposure scenario developed for use inconjunction with the EPA’s TRIM.FaTE model. Thehypothetical screening scenario was subjected to asensitivity analysis to ensure that its key designparameters were established such that environmental media This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

214. Page 214 of 604concentrations were not underestimated (i.e., to minimizethe occurrence of false negatives or results that suggestthat risks might be acceptable when, in fact, actual risksare high) and to also minimize the occurrence of falsepositives for human health endpoints. We call thisapplication of the TRIM.FaTE model TRIM-Screen. Thefacility-specific emission rates of each of the PB–HAP ineach source category were compared to the TRIM-Screenemission threshold values for each of the PB–HAP identifiedin the source category datasets to assess the potential forsignificant human health risks or environmental risks vianon-inhalation pathways. There was only one facility in the Natural GasTransmission and Storage source category with reportedemissions of PB-HAP, and the emission rates were less thanthe emission threshold values. There were 29 facilities inthe Oil and Natural Gas Production source category withreported emissions of PB-HAP, and one of these had emissionrates greater than the emission threshold values. In thiscase, the emission threshold value for POM was exceeded bya factor of 6. For POM, dairy, vegetables and fruits werethe three most dominant exposure pathways driving humanexposures in the hypothetical screening exposure scenario. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

215. Page 215 of 604The single facility with emissions exceeding the emissionthreshold value for POM is located in a highlyindustrialized area. Therefore, since the exposure pathwayswhich would drive high human exposure are not locallyavailable, multi-pathway exposures and environmental riskswere deemed negligible, and no further analysis wasperformed. For further information on the multi-pathwayanalysis approach, see the residual risk documentation.e. Assessing Risks Considering Emissions Control Options In addition to assessing baseline inhalation risks andscreening for potential multi-pathway risks, whereappropriate, we also estimated risks considering thepotential emission reductions that would be achieved by theparticular control options under consideration. In thesecases, the expected emissions reductions were applied tothe specific HAP and emissions sources in the sourcecategory dataset to develop corresponding estimates of riskreductions.f. Conducting Other Risk-Related Analyses: Facility-WideAssessments To put the source category risks in context, we alsoexamined the risks from the entire “facility,” where thefacility includes all HAP-emitting operations within a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

216. Page 216 of 604contiguous area and under common control. In other words,for each facility that includes one or more sources fromone of the source categories under review, we examined theHAP emissions not only from the source category ofinterest, but also from all other emission sources at thefacility. The emissions data for generating these“facility-wide” risks were also obtained from the 2005 NEI.For every facility included in the MACT database, we alsoretrieved emissions data and release characteristics forall other emission sources at the same facility. Weestimated the risks due to the inhalation of HAP that areemitted “facility-wide” for the populations residing within50 km of each facility, consistent with the methods usedfor the source category analysis described above. For thesefacility-wide risk analyses, the modeled source categoryrisks were compared to the facility-wide risks to determinethe portion of facility-wide risks that could be attributedto the source categories addressed in this proposal. Wespecifically examined the facilities associated with thehighest estimates of risk and determined the percentage ofthat risk attributable to the source category of interest.The risk documentation available through the docket forthis action provides the methodology and the results of the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

217. Page 217 of 604facility-wide analyses for each source category.g. Conducting Other Analyses: Demographic Analysis To examine the potential for any environmental justice(EJ) issues that might be associated with each sourcecategory, we performed a demographic analysis of populationrisk. In this analysis, we evaluated the distributions ofHAP-related cancer and noncancer risks across differentsocial, demographic and economic groups within thepopulations living near the facilities where these sourcecategories are located. The development of demographicanalyses to inform the consideration of EJ issues in theEPA rulemakings is an evolving science. The EPA offers thedemographic analyses in this rulemaking to inform theconsideration of potential EJ issues and invites publiccomment on the approaches used and the interpretations madefrom the results, with the hope that this will support therefinement and improve the utility of such analyses forfuture rulemakings. For the demographic analyses, we focus on thepopulations within 50 km of any facility estimated to haveexposures to HAP which result in cancer risks of 1-in-1million or greater, or noncancer HI of 1 or greater (basedon the emissions of the source category or the facility, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

218. Page 218 of 604respectively). We examine the distributions of those risksacross various demographic groups, comparing thepercentages of particular demographic groups to the totalnumber of people in those demographic groups nationwide.The results, including other risk metrics, such as averagerisks for the exposed populations, are documented insource-category-specific technical reports in the docketfor both source categories covered in this proposal. The basis for the risk values used in these analyseswere the modeling results based on actual emissions levelsobtained from the HEM-3 model described above. The riskvalues for each census block were linked to a database ofinformation from the 2000 Decennial census that includesdata on race and ethnicity, age distributions, povertystatus, household incomes and education level. The CensusDepartment Landview® database was the source of the data onrace and ethnicity and the data on age distributions,poverty status, household incomes and education level wereobtained from the 2000 Census of Population and HousingSummary File 3 Long Form. While race and ethnicity censusdata are available at the census block level, the age andincome census data are only available at the census blockgroup level (which includes an average of 26 blocks or an This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

219. Page 219 of 604average of 1,350 people). Where census data are availableat the block group level, but not the block level, weassumed that all census blocks within the block group havethe same distribution of ages and incomes as the blockgroup. For each source category, we focused on those censusblocks where source category risk results show estimatedlifetime inhalation cancer risks above 1-in-1 million orchronic noncancer indices above 1 and determined therelative percentage of different racial and ethnic groups,different age groups, adults with and without a high schooldiploma, people living in households below the nationalmedian income and for people living below the poverty linewithin those census blocks. The specific census populationcategories studied include:• Total population• White• African American (or Black)• Native Americans• Other races and multiracial• Hispanic or Latino• Children 18 years of age and under• Adults 19 to 64 years of age This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

220. Page 220 of 604• Adults 65 years of age and over• Adults without a high school diploma• Households earning under the national median income• People living below the poverty line It should be noted that these categories overlap insome instances, resulting in some populations being countedin more than one category (e.g., other races andmultiracial and Hispanic). In addition, while not aspecific census population category, we also examined risksto “Minorities,” a classification which is defined forthese purposes as all race population categories exceptwhite. For further information about risks to the populationslocated near the facilities in these source categories, wealso evaluated the estimated distribution of inhalationcancer and chronic noncancer risks associated with the HAPemissions from all the emissions sources at the facility(i.e., facility-wide). This analysis used the facility-wideRTR modeling results and the census data described above. The methodology and the results of the demographicanalyses for each source category are included in a source-category-specific technical report for each of thecategories, which are available in the docket for this This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

221. Page 221 of 604action.h. Considering Uncertainties in Risk Assessment Uncertainty and the potential for bias are inherent inall risk assessments, including those performed for thesource categories addressed in this proposal. Althoughuncertainty exists, we believe that our approach, whichused conservative tools and assumptions, ensures that ourdecisions are health-protective. A brief discussion of theuncertainties in the emissions datasets, dispersionmodeling, inhalation exposure estimates and dose-responserelationships follows below. A more thorough discussion ofthese uncertainties is included in the risk assessmentdocumentation (referenced earlier) available in the docketfor this action.i. Uncertainties in the Emissions Datasets Although the development of the MACT dataset involvedQA/quality control processes, the accuracy of emissionsvalues will vary depending on the source of the data, thedegree to which data are incomplete or missing, the degreeto which assumptions made to complete the datasets areinaccurate, errors in estimating emissions values and otherfactors. The emission estimates considered in this analysisgenerally are annual totals for certain years that do not This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

222. Page 222 of 604reflect short-term fluctuations during the course of a yearor variations from year to year. The estimates of peak hourly emission rates for theacute effects screening assessment were based on amultiplication factor of 10 applied to the average annualhourly emission rate, which is intended to account foremission fluctuations due to normal facility operations.Additionally, although we believe that we have data formost facilities in these two source categories in our RTRdataset, our dataset may not include data for all existingfacilities. Moreover, there are uncertainties with regardto the identification of sources as major or area in theNEI for these source categories.ii. Uncertainties in Dispersion Modeling While the analysis employed the EPA’s recommendedregulatory dispersion model, AERMOD, we recognize thatthere is uncertainty in ambient concentration estimatesassociated with any model, including AERMOD. Incircumstances where we had to choose between various modeloptions, where possible, model options (e.g., rural/urban,plume depletion, chemistry) were selected to provide anoverestimate of ambient air concentrations of the HAPrather than underestimates. However, because of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

223. Page 223 of 604practicality and data limitation reasons, some factors(e.g., meteorology, building downwash) have the potentialin some situations to overestimate or underestimate ambientimpacts. For example, meteorological data were taken from asingle year (1991) and facility locations can be asignificant distance from the site where these data weretaken. Despite these uncertainties, we believe that at off-site locations and census block centroids, the approachconsidered in the dispersion modeling analysis shouldgenerally yield overestimates of ambient HAPconcentrations.iii. Uncertainties in inhalation exposure The effects of human mobility on exposures were notincluded in the assessment. Specifically, short-termmobility and long-term mobility between census blocks in 31the modeling domain were not considered. The assumption ofnot considering short or long-term population mobility doesnot bias the estimate of the theoretical MIR, nor does itaffect the estimate of cancer incidence since the totalpopulation number remains the same. It does, however,31 Short-term mobility is movement from one micro-environment toanother over the course of hours or days. Long-term mobility ismovement from one residence to another over the course of alifetime. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

224. Page 224 of 604affect the shape of the distribution of individual risksacross the affected population, shifting it toward higherestimated individual risks at the upper end and reducingthe number of people estimated to be at lower risks,thereby increasing the estimated number of people atspecific risk levels. In addition, the assessment predicted the chronicexposures at the centroid of each populated census block assurrogates for the exposure concentrations for all peopleliving in that block. Using the census block centroid topredict chronic exposures tends to over-predict exposuresfor people in the census block who live further from thefacility, and under-predict exposures for people in thecensus block who live closer to the facility. Thus, usingthe census block centroid to predict chronic exposures maylead to a potential understatement or overstatement of thetrue maximum impact, but is an unbiased estimate of averagerisk and incidence. The assessments evaluate the cancer inhalation risksassociated with continuous pollutant exposures over a 70-year period, which is the assumed lifetime of anindividual. In reality, both the length of time thatmodeled emissions sources at facilities actually operate This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

225. Page 225 of 604(i.e., more or less than 70 years), and the domestic growthor decline of the modeled industry (i.e., the increase ordecrease in the number or size of United Statesfacilities), will influence the risks posed by a givensource category. Depending on the characteristics of theindustry, these factors will, in most cases, result in anoverestimate both in individual risk levels and in thetotal estimated number of cancer cases. However, in rarecases, where a facility maintains or increases its emissionlevels beyond 70 years, residents live beyond 70 years atthe same location, and the residents spend most of theirdays at that location, then the risks could potentially beunderestimated. Annual cancer incidence estimates fromexposures to emissions from these sources would not beaffected by uncertainty in the length of time emissionssources operate. The exposure estimates used in these analyses assumechronic exposures to ambient levels of pollutants. Becausemost people spend the majority of their time indoors,actual exposures may not be as high, depending on thecharacteristics of the pollutants modeled. For many of theHAP, indoor levels are roughly equivalent to ambientlevels, but for very reactive pollutants or larger This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

226. Page 226 of 604particles, these levels are typically lower. This factorhas the potential to result in an overstatement of 25 to 30 32percent of exposures. In addition to the uncertainties highlighted above,there are several factors specific to the acute exposureassessment that should be highlighted. The accuracy of anacute inhalation exposure assessment depends on thesimultaneous occurrence of independent factors that mayvary greatly, such as hourly emissions rates, meteorology,and human activity patterns. In this assessment, we assumethat individuals remain for 1 hour at the point of maximumambient concentration as determined by the co-occurrence ofpeak emissions and worst-case meteorological conditions.These assumptions would tend to overestimate actualexposures since it is unlikely that a person would belocated at the point of maximum exposure during the time ofworst-case impact.iv. Uncertainties in Dose-Response Relationships There are uncertainties inherent in the development ofthe dose-response values used in our risk assessments forcancer effects from chronic exposures and noncancer effects32 U.S. EPA. National-Scale Air Toxics Assessment for 1996. (EPA453/R–01–003; January 2001; page 85.) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

227. Page 227 of 604from both chronic and acute exposures. Some uncertaintiesmay be considered quantitatively, and others generally areexpressed in qualitative terms. We note as a preface tothis discussion a point on dose-response uncertainty thatis brought out in the EPA 2005 Cancer Guidelines; namely,that “the primary goal of the EPA actions is protection ofhuman health; accordingly, as an Agency policy, riskassessment procedures, including default options that areused in the absence of scientific data to the contrary,should be health protective.” (EPA 2005 Cancer Guidelines,pages 1–7.) This is the approach followed here assummarized in the next several paragraphs. A completedetailed discussion of uncertainties and variability indose-response relationships is given in the residual riskdocumentation, which is available in the docket for thisaction. Cancer URE values used in our risk assessments arethose that have been developed to generally provide anupper bound estimate of risk. That is, they represent a“plausible upper limit to the true value of a quantity”(although this is usually not a true statistical confidence This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

228. Page 228 of 604 33limit). In some circumstances, the true risk could be aslow as zero; however, in other circumstances, the risk 34could also be greater. When developing an upper boundestimate of risk and to provide risk values that do notunderestimate risk, health-protective default approachesare generally used. To err on the side of ensuring adequatehealth-protection, the EPA typically uses the upper boundestimates rather than lower bound or central tendencyestimates in our risk assessments, an approach that mayhave limitations for other uses (e.g., priority-setting orexpected benefits analysis). Chronic noncancer reference (RfC and reference dose(RfD)) values represent chronic exposure levels that areintended to be health-protective levels. Specifically,these values provide an estimate (with uncertainty spanningperhaps an order of magnitude) of daily oral exposure (RfD)or of a continuous inhalation exposure (RfC) to the humanpopulation (including sensitive subgroups) that is likelyto be without an appreciable risk of deleterious effects33 IRIS glossary (http://www.epa.gov/NCEA/iris/help_gloss.htm).34 An exception to this is the URE for benzene, which isconsidered to cover a range of values, each end of which isconsidered to be equally plausible and which is based on maximumlikelihood estimates. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

229. Page 229 of 604during a lifetime. To derive values that are intended to be“without appreciable risk,” the methodology relies upon anuncertainty factor (UF) approach (U.S. EPA, 1993, 1994)which includes consideration of both uncertainty andvariability. When there are gaps in the availableinformation, UF are applied to derive reference values thatare intended to protect against appreciable risk of 35deleterious effects. The UF are commonly default values,e.g., factors of 10 or 3, used in the absence of compound-specific data; where data are available, UF may also bedeveloped using compound-specific information. When dataare limited, more assumptions are needed and more UF areused. Thus, there may be a greater tendency to overestimate35 According to the NRC report, Science and Judgment in RiskAssessment (NRC, 1994) “[Default] options are generic approaches,based on general scientific knowledge and policy judgment, thatare applied to various elements of the risk assessment processwhen the correct scientific model is unknown or uncertain.” The1983 NRC report, Risk Assessment in the Federal Government:Managing the Process, defined default option as “the optionchosen on the basis of risk assessment policy that appears to bethe best choice in the absence of data to the contrary” (NRC,1983a, p. 63). Therefore, default options are not rules that bindthe Agency; rather, the Agency may depart from them in evaluatingthe risks posed by a specific substance when it believes this tobe appropriate. In keeping with EPA’s goal of protecting publichealth and the environment, default assumptions are used toensure that risk to chemicals is not underestimated (althoughdefaults are not intended to overtly overestimate risk). See EPA,2004, An Examination of EPA Risk Assessment Principles andPractices, EPA/100/B–04/001 available at:http://www.epa.gov/osa/pdfs/ratf-final.pdf. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

230. Page 230 of 604risk in the sense that further study might supportdevelopment of reference values that are higher (i.e., lesspotent) because fewer default assumptions are needed.However, for some pollutants, it is possible that risks maybe underestimated. While collectively termed “uncertaintyfactor,” these factors account for a number of differentquantitative considerations when using observed animal(usually rodent) or human toxicity data in the developmentof the RfC. The UF are intended to account for: (1)Variation in susceptibility among the members of the humanpopulation (i.e., inter-individual variability); (2)uncertainty in extrapolating from experimental animal datato humans (i.e., interspecies differences); (3) uncertaintyin extrapolating from data obtained in a study with less-than-lifetime exposure (i.e., extrapolating from sub-chronic to chronic exposure); (4) uncertainty inextrapolating the observed data to obtain an estimate ofthe exposure associated with no adverse effects; and (5)uncertainty when the database is incomplete or there areproblems with the applicability of available studies. Manyof the UF used to account for variability and uncertaintyin the development of acute reference values are quitesimilar to those developed for chronic durations, but they This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

231. Page 231 of 604more often use individual UF values that may be less than10. UF are applied based on chemical-specific or healtheffect-specific information (e.g., simple irritationeffects do not vary appreciably between human individuals,hence a value of 3 is typically used), or based on thepurpose for the reference value (see the followingparagraph). The UF applied in acute reference valuederivation include: (1) Heterogeneity among humans; (2)uncertainty in extrapolating from animals to humans; (3)uncertainty in lowest observed adverse effect (exposure)level to no observed adverse effect (exposure) leveladjustments; and (4) uncertainty in accounting for anincomplete database on toxic effects of potential concern.Additional adjustments are often applied to account foruncertainty in extrapolation from observations at oneexposure duration (e.g., 4 hours) to derive an acutereference value at another exposure duration (e.g., 1hour). Not all acute reference values are developed for thesame purpose and care must be taken when interpreting theresults of an acute assessment of human health effectsrelative to the reference value or values being exceeded.Where relevant to the estimated exposures, the lack of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

232. Page 232 of 604short-term dose-response values at different levels ofseverity should be factored into the risk characterizationas potential uncertainties. Although every effort is made to identify peer-reviewed reference values for cancer and noncancer effectsfor all pollutants emitted by the sources included in thisassessment, some HAP continue to have no reference valuesfor cancer or chronic noncancer or acute effects. Sinceexposures to these pollutants cannot be included in aquantitative risk estimate, an understatement of risk forthese pollutants at environmental exposure levels ispossible. For a group of compounds that are eitherunspeciated or do not have reference values for everyindividual compound (e.g., glycol ethers), weconservatively use the most protective reference value toestimate risk from individual compounds in the group ofcompounds. Additionally, chronic reference values for several ofthe compounds included in this assessment are currentlyunder the EPA IRIS review and revised assessments maydetermine that these pollutants are more or less potentthan the current value. We may re-evaluate residual risksfor the final rulemaking if these reviews are completed This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

233. Page 233 of 604prior to our taking final action for these sourcecategories and a dose-response metric changes enough toindicate that the risk assessment supporting this noticemay significantly understate human health risk.v. Uncertainties in the Multi-Pathway and EnvironmentalEffects Assessment We generally assume that when exposure levels are notanticipated to adversely affect human health, they also arenot anticipated to adversely affect the environment. Foreach source category, we generally rely on the site-specific levels of PB-HAP emissions to determine whether afull assessment of the multi-pathway and environmentaleffects is necessary. As discussed above, we conclude thatthe potential for these types of impacts is low for thesesource categories.vi. Uncertainties in the Facility-Wide Risk Assessment Given that the same general analytical approach andthe same models were used to generate facility-wide riskresults as were used to generate the source category riskresults, the same types of uncertainties discussed abovefor our source category risk assessments apply to thefacility-wide risk assessments. Additionally, the degree ofuncertainty associated with facility-wide emissions and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

234. Page 234 of 604risks is likely greater because we generally have notconducted a thorough engineering review of emissions datafor source categories not currently undergoing an RTRreview.vii. Uncertainties in the Demographic Analysis Our analysis of the distribution of risks acrossvarious demographic groups is subject to the typicaluncertainties associated with census data (e.g., errors infilling out and transcribing census forms), as well as theadditional uncertainties associated with the extrapolationof census-block group data (e.g., income level andeducation level) down to the census block level.2. What are the results and proposed decisions from therisk review for the Oil and Natural Gas Production sourcecategory?a. Results of the Risk Assessments and Analyses We conducted an inhalation risk assessment for the Oiland Natural Gas Production source category. We alsoconducted an assessment of facility-wide risk. Details ofthe risk assessments and analyses can be found in theresidual risk documentation, which is available in thedocket for this action. For informational purposes and toexamine the potential for any EJ issues that might be This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

235. Page 235 of 604associated with each source category, we performed ademographic analysis of population risks.i. Inhalation Risk Assessment Results Table 2 provides an overall summary of the results ofthe inhalation risk assessment. TABLE 2. OIL AND NATURAL GAS PRODUCTION INHALATION RISK ASSESSMENT RESULTS Maximum Maximum Individual Estimated Maximum Off- Cancer Risk Estimated Annual Chronic Noncancer Site (in 1 million)2 Population Cancer TOSHI4 Acute Number of Actual Allowable at Risk ≥ Incidence Actual Allowable NoncancFacilities Emissions Emissions 1-in-1 (cases Emissions Emissions er HQ5 1 Level Level Million per year) Level Level HQREL = 9 (benzen 100 - 0.007 - e) 990 40 160,0003 0.1 0.7 4003 0.023 HQAEGL-1 = 0.07 (benzen e)1 Number of facilities evaluated in the risk analysis.2 Estimated maximum individual excess lifetime cancer risk.3 The EPA IRIS assessment for benzene provides a range ofequally-plausible URE (2.2E-06 to 7.8E-06 per ug/m3),giving rise to ranges for the estimates of cancer MIR andcancer incidence. Estimated population values are notscalable with benzene URE range, but would be lower usingthe lower end of the URE range.4 Maximum TOSHI. The target organ with the highest TOSHIfor the Oil and Natural Gas Production source category isthe respiratory system.5 The maximum estimated acute exposure concentration wasdivided by available short-term dose-response values todevelop an array of HQ values. As shown in Table 2, the results of the inhalationrisk assessment performed using actual emissions dataindicate the maximum lifetime individual cancer risk could This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

236. Page 236 of 604be as high as 40-in-1 million, with POM driving the highestrisk, and benzene driving risks overall. The totalestimated cancer incidence from this source category is0.02 excess cancer cases per year (0.007 excess cancercases per year based on the lower end of the benzene URErange), or one case in every 50 years. Approximately160,000 people are estimated to have cancer risks at orabove 1-in-1 million as a result of the emissions from 89facilities (use of the lower end of the benzene URE rangewould further reduce this population estimate). The maximumchronic non-cancer TOSHI value for the source categorycould be up to 0.1 from emissions of naphthalene,indicating no significant potential for chronic noncancerimpacts. As explained above, our analysis of potentialdifferences between actual emission levels and emissionsallowable under the oil and natural gas production MACTstandard indicate that MACT-allowable emission levels maybe up to 50 times greater than actual emission levels.Considering this difference, the risk results from theinhalation risk assessment indicate the maximum lifetimeindividual cancer risk could be as high as 400-in-1 million(100-in-1 million based on the lower end of the benzene URE This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

237. Page 237 of 604range) and the maximum chronic noncancer TOSHI value couldbe as high as 0.7 at the MACT-allowable emissions level.ii. Facility-Wide Risk Assessment Results A facility-wide risk analysis was also conducted basedon actual emissions levels. Table 3 displays the results ofthe facility-wide risk assessment. For detailed facility-specific results, see Table 2 of Appendix 6 of the riskdocument in the docket for this rulemaking. TABLE 3. OIL AND NATURAL GAS PRODUCTION FACILITY-WIDE RISK ASSESSMENT RESULTSNumber of facilities analyzed 990 Estimated maximum facility-wide 100 individual cancer risk (in 1 million) Number of facilities with estimated facility-wide individual cancer risk 1 of 100-in-1 million or more Number of facilities at which the Oil and Natural Gas Production source category contributes 50 percent or 0 more to the facility-wide individualCancer cancer risks of 100-in-1 million orRisk more Number of facilities with facility- wide individual cancer risk of 1-in-1 140 million or more Number of facilities at which the Oil and Natural Gas Production source category contributes 50 percent or 85 more to the facility-wide individual cancer risk of 1-in-1 million or more Maximum facility-wide chronicChronic 9 noncancer TOSHINoncancerRisk Number of facilities with facility- 10 wide maximum noncancer TOSHI greater This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

238. Page 238 of 604 than 1 Number of facilities at which the Oil and Natural Gas Production source category contributes 50 percent or 0 more to the facility-wide maximum noncancer TOSHI of 1 or more The facility-wide MIR from all HAP emissions at afacility that contains sources subject to the oil andnatural gas production MACT standards is estimated to be100-in-1 million, based on actual emissions. Of the 990facilities included in this analysis, only one has afacility-wide MIR of 100-in-1 million. At this facility,oil and natural gas production accounts for less than 2percent of the total facility-wide risk. Nickel emissionsfrom oil-fired boilers and formaldehyde emissions fromreciprocating internal combustion engines (RICE) contributeessentially all the facility-wide risks at this facility,with over 80 percent of the risk attributed to the nickelemissions.36 There are 140 facilities with facility-wide MIRof 1-in-1 million or greater. Of these facilities, 85 haveoil and natural gas production operations that contribute36 We note that there is an ongoing IRIS reassessment forformaldehyde, and that future RTR risk assessments will use thecancer potency for formaldehyde that results from thatreassessment. As a result, the current results may not matchthose of future assessments. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

239. Page 239 of 604greater than 50 percent to the facility-wide risks. Asdiscussed above, we are proposing MACT standards for BTEXemissions from small glycol dehydrators in this action.These standards would reduce the risk from benzeneemissions at facilities with oil and gas production.Formaldehyde emissions will be assessed under future RTRfor RICE. The facility-wide maximum individual chronic noncancerTOSHI is estimated to be 9 based on actual emissions. Ofthe 990 facilities included in this analysis, 10 havefacility-wide maximum chronic noncancer TOSHI valuesgreater than 1. Of these facilities, none had oil andnatural gas production operations that contributed greaterthan 50 percent to these facility-wide risks. The chronicnoncancer risks at these 10 facilities are primarily drivenby acrolein emissions from RICE.iii. Demographic Risk Analysis Results The results of the demographic analyses performed toinvestigate the distribution of cancer risks at or above 1-in-1 million among the surrounding population aresummarized in Table 4 below. These results, for variousdemographic groups, are based on actual emissions levelsfor the population living within 50 km of the facilities. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

240. Page 240 of 604 TABLE 4. OIL AND NATURAL GAS PRODUCTION DEMOGRAPHIC RISK ANALYSIS RESULTS Population with Cancer Risk at or Above 1-in-1 Million Due to…. Source Category Facility-Wide Nationwide HAP Emissions HAP EmissionsTotal 285,000,000 160,000 597,000Population Race by PercentWhite 75 62 61All Other 25 38 39Races Race by PercentWhite 75 62 61African 12 8 12AmericanNative 0.7 1.3 0.9AmericanOther and 25 30 12Multiracial Ethnicity by PercentHispanic 14 22 34Non-Hispanic 86 78 66 Income by PercentBelow Poverty 13 14 19LevelAbove Poverty 87 86 81Level Education by PercentOver 25 andwithout High 13 10 16SchoolDiplomaOver 25 andwith a High 87 90 84SchoolDiploma The results of the Oil and Natural Gas Productionsource category demographic analysis indicate that there This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

241. Page 241 of 604are approximately 160,000 people exposed to a cancer riskat or above 1-in-1 million due to emissions from the sourcecategory, including an estimated 38 percent that areclassified as minority (listed as “All Other Races” in thetable above). Of the 160,000 people with estimated cancerrisks at or above 1-in-1 million from the source category,25 percent are in the “Other and Multiracial” demographicgroup, 22 percent are in the “Hispanic or Latino”demographic group, and 14 percent are in the “Below PovertyLevel” demographic group, results which are 13, 8 and 1percentage points higher, respectively, than the respectivepercentages for these demographic groups across the UnitedStates. The percentages for the other demographic groupsare lower than their respective nationwide percentages. Thetable also shows that there are approximately 597,000people exposed to an estimated cancer risk at or above 1-in-1 million due to facility-wide emissions, including 30percent in the “Other and Multiracial” demographic group,34 percent in the “Hispanic or Latino” demographic group,1.3 percent in the “Native American” demographic group and16 percent in the “Over 25 and without High School Diploma”demographic group, results which are 18, 2, 0.4 and 3percentage points higher than the percentages for these This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

242. Page 242 of 604demographic groups across the United States, respectively.The percentages for the other demographic groups are lowerthan their respective nationwide percentages.b. What are the proposed risk decisions for the Oil andNatural Gas Production source category?i. Risk Acceptability In the risk analysis we performed for this sourcecategory, pursuant to CAA section 112(f)(2), we consideredthe available health information--the MIR; the numbers ofpersons in various risk ranges; cancer incidence; themaximum noncancer HI; the maximum acute noncancer hazard;the extent of noncancer risks; the potential for adverseenvironmental effects; and distribution of risks in theexposed population; and risk estimation uncertainty (54 FR38044, September 14, 1989). For the Oil and Natural Gas Production sourcecategory, the risk analysis we performed indicates that thecancer risks to the individual most exposed could be ashigh as 40-in-1 million due to actual emissions and as highas 400-in-1 million due to MACT-allowable emissions (100-in-1 million, based on the lower end of the benzene URErange). While the 40-in-1 million risk due to actualemissions is considerably less than 100-in-1 million, which This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

243. Page 243 of 604is the presumptive limit of acceptability, the 400-in-1million risk due to allowable emissions is considerablyhigher and is considered unacceptable. We do note, however,that the risk analysis shows low cancer incidence (1 casein every 50 years), low potential for adverse environmentaleffects or human health multi-pathway effects and thatchronic noncancer health impacts are unlikely. We also conclude that acute noncancer health impactsare unlikely. As discussed above, screening estimates ofacute exposures and risks were evaluated for each of theHAP at the point of highest off-site exposure for eachfacility (i.e., not just the census block centroids)assuming that a person is located at this spot at a timewhen both the peak emission rate and worst-case dispersionconditions occur. Under these worst-case conditions, weestimate benzene acute HQ values (based on the REL) couldbe as high as 9. Although the REL (which indicates thelevel below which adverse effects are not anticipated) isexceeded in this case, we believe the potential for acuteeffects is low for several reasons. First, the acutemodeling scenario is worst-case because of the confluenceof peak emission rates and worst-case dispersionconditions. Second, the benzene REL is based on a 6-hour This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

244. Page 244 of 604exposure duration because a 1-hour exposure duration valuewas unavailable. An REL based on a 6-hour exposure durationis generally lower than an REL based on a 1-hour exposureduration and, consequently, easier to exceed. Also,although there are exceedances of the REL, the highestestimated 1-hour exposure is less than 10 percent of theAEGL-1 value, which is a level at which effects could beexperienced. Finally, the generally sparse populations nearthese facilities make it less likely that a person would benear the plant to be exposed. For example, in the two caseswhere the acute HQ value is as high as 9, there are only 30people associated with the census blocks within 2 miles ofthe two facilities. While our additional analysis of facility-wide risksshowed that there is one facility with maximum facility-wide cancer risk of 100-in-1 million or greater and 10facilities with a maximum chronic noncancer TOSHI greaterthan 1, it also showed that oil and natural gas productionoperations did not drive these risks. In determining whether risk is acceptable, weconsidered the available health information, as describedabove. In this case, although a number of factors weconsidered indicate relatively low risk concern, we are This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

245. Page 245 of 604proposing to determine that the risks are unacceptable, inlarge part, because the MIR is 400-in-1 million due toMACT-allowable emissions, which greatly exceeds the“presumptive limit on maximum individual lifetime risk ofapproximately 1-in-10 thousand [100-in-1 million]recognized in the Benzene NESHAP (54 FR 38045).” The MIR,based on MACT-allowable emissions, is driven by theallowable emissions of 0.9 Mg/yr benzene under the MACT asa compliance option. We are, therefore, proposing toeliminate the alternative compliance option of 0.9 Mg/yrbenzene from the existing glycol dehydrator MACTrequirements. With this change, the source category MIR,based on MACT-allowable emissions, would be reduced to 40-in-1 million, which we find acceptable in light of all theother factors considered. Thus, we are proposing that therisks from the Oil and Natural Gas Production sourcecategory are acceptable, with the removal of thealternative compliance option of 0.9 Mg/yr benzene limitfrom the current glycol dehydrator MACT requirements. Pursuant to CAA section 112(f)(4), we are proposingthat this change (i.e., removal of the 0.9 Mg/yr compliancealternative) apply 90 days after its effective date. We arerequesting comment on whether or not this is sufficient This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

246. Page 246 of 604time for the large dehydrators that have been relying onthis compliance alternative to come into compliance withthe 95-percent control requirement or if additional time isneeded. See CAA section 112(f)(4)(A). We recognize that our proposal to remove the 0.9 Mg/yrcompliance alternative for the 95-percent control glycoldehydrator MACT standard could have negative impacts onsome sources that have come to rely on the flexibility thisalternative provides. We solicit comment on any suchimpacts and whether such impacts warrant adding a differentcompliance alternative that would result in less risk thanthe 0.9 Mg/yr benzene limit compliance option. If acommenter suggests a different compliance alternative, thecommenter should explain, in detail, what that alternativewould be, how it would work and how it would reduce risk.ii. Ample Margin of Safety We next considered whether this revised standard(existing MACT plus removal of 0.9 Mg/yr benzene complianceoption) provides an ample margin of safety. In thisanalysis, we investigated available emissions controloptions that might reduce the risk associated withemissions from the source category and considered thisinformation along with all of the health risks and other This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

247. Page 247 of 604health information considered in the risk acceptabilitydetermination. For glycol dehydrators, we considered the addition ofa second control device in the same manner that wasdiscussed in the floor evaluation in section VII.B.1 above.The cost effectiveness associated with that option would be$167,200/Mg, which we believe is too high to requireadditional controls on glycol dehydrators. Similarly, we considered the addition of a secondcontrol device to the required MACT floor control device(cost effectiveness of $18,300/Mg). Similar to ourdiscussion of beyond-the-MACT-floor controls for glycoldehydrators in section VII.B.1 of this preamble, theincremental cost to add a second control device for storagevessels would be approximately 20 times higher than theMACT floor cost effectiveness, or $366,000/Mg. We do notbelieve this cost effectiveness is reasonable. For leak detection, we considered implementation ofLDAR programs that are more stringent than the currentstandards. An assessment performed for various LDAR optionsunder the NSPS in section VI.B.4.b of this preamble yieldedthe lowest cost effectiveness of $5,170/Mg ($4,700/ton) forcontrol of VOC for the options evaluated. A LDAR program to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

248. Page 248 of 604control HAP would involve similar costs for equipment,labor, etc., to those considered in the NSPS assessment,but since there is approximately 20 times less HAP than VOCpresent in material handled in regulated equipment, thecost effectiveness to control HAP would be approximately 20times greater (i.e., $100,000/Mg) for HAP, which we believeis not reasonable. In accordance with the approach established in theBenzene NESHAP, the EPA weighed all health risk measuresand information considered in the risk acceptabilitydetermination, along with the costs and economic impacts ofemissions controls, technological feasibility,uncertainties and other relevant factors in making ourample margin of safety determination. Considering thehealth risk information and the high cost effectiveness ofthe options identified, we propose that the existing MACTstandards, with the removal of the 1 tpy benzene limitcompliance option from the glycol dehydrator standards,provide an ample margin of safety to protect public health. While we are proposing that the oil and natural gasproduction MACT standards (with the removal of thealternative compliance option of 1 tpy benzene limit)provide an ample margin of safety to protect public health, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

249. Page 249 of 604we are concerned about the estimated facility-wide risksidentified through these screening analyses. As describedpreviously, the highest estimated facility-wide cancerrisks are mostly due to emissions from oil fired boilersand RICE. Both of these sources are regulated under othersource categories and we anticipate that emissionreductions from those sources will occur as standards forthose source categories are implemented.3. What are the results and proposed decisions from therisk review for the Natural Gas Transmission and Storagesource category?a. Results of the Risk Assessments and Analyses We conducted an inhalation risk assessment for theNatural Gas Transmission and Storage source category. Wealso conducted an assessment of facility-wide risk andperformed a demographic analysis of population risks.Details of the risk assessments and analyses can be foundin the residual risk documentation, which is available inthe docket for this action.i. Inhalation Risk Assessment Results Table 5 provides an overall summary of the results ofthe inhalation risk assessment. For informational purposesand to examine the potential for any EJ issues that might This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

250. Page 250 of 604be associated with each source category, we performed ademographic analysis of population risks. TABLE 5. NATURAL GAS TRANSMISSION AND STORAGE INHALATION RISK ASSESSMENT RESULTS Maximum Individual Estimated Maximum Maximum Cancer Risk Estimated Annual Chronic Noncancer Off- (in 1 million)2 Population Cancer TOSHI4 Site Number of Actual Allowable at Risk ≥ Incidence Actual Allowable AcuteFacilities Emissions Emissions 1-in-1 (cases Emissions Emissions Noncanc 1 Level Level Million per year) Level Level er HQ5 HQREL = 5 (benzen 0.0003 - e) 321 30 - 903 30 - 903 2,5003 0.4 0.8 HQAEGL-1 0.0013 = 0.2 (chloro benzene )1 Number of facilities evaluated in the risk analysis.2 Estimated maximum individual excess lifetime cancer risk.3 The EPA IRIS assessment for benzene provides a range ofequally-plausible URE (2.2E-06 to 7.8E-06 per ug/m3),giving rise to ranges for the estimates of cancer MIR andcancer incidence. Estimated population values are notscalable with benzene URE range, but would be lower usingthe lower end of the URE range.4 Maximum TOSHI. The target organ with the highest TOSHIfor the Natural Gas Transmission and Storage sourcecategory is the immune system.5 The maximum estimated acute exposure concentration wasdivided by available short-term dose-response values todevelop an array of HQ values. As shown in Table 5 above, the results of theinhalation risk assessment performed using actual emissionsdata indicate the maximum lifetime individual cancer riskcould be as high as 90-in-1 million, (30-in-1 million basedon the lower end of the benzene URE range), with benzene as This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

251. Page 251 of 604the major contributor to the risk. The total estimatedcancer incidence from the source category is 0.001 excesscancer cases per year (0.0003 excess cancer cases per yearbased on the lower end of the benzene URE range), or onecase in every polycyclic organic matter 1,000 years.Approximately 2,500 people are estimated to have cancerrisks at or above 1-in-1 million as a result of theemissions from 15 facilities (use of the lower end of thebenzene URE range would further reduce this populationestimate). The maximum chronic noncancer TOSHI value forthe source category could be up to 0.4 from emissions ofbenzene, indicating no significant potential for chronicnoncancer impacts. As explained above in section VII.C.1.b, our analysisof potential differences between actual emission levels andemissions allowable under the natural gas transmission andstorage MACT standard indicate that MACT-allowable emissionlevels may be up to 50 times greater than actual emissionlevels at some sources. However, because some sources areemitting at the level allowed under the current NESHAP, therisk results from the inhalation risk assessment indicatethe maximum lifetime individual cancer risk would still be90-in-1 million (30-in-1 million based on the lower end of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

252. Page 252 of 604the benzene URE range), based on both actual and allowableemission levels, and the maximum chronic noncancer TOSHIvalue could be as high as 0.8 at the MACT-allowableemissions level.ii. Facility-Wide Risk Assessment Results A facility-wide risk analysis was also conducted basedon actual emissions levels. Table 6 below displays theresults of the facility-wide risk assessment. For detailedfacility-specific results, see Table 2 of Appendix 6 of therisk document in the docket for this rulemaking. TABLE 6. NATURAL GAS TRANSMISSION AND STORAGE FACILITY-WIDE RISK ASSESSMENT RESULTSNumber of Facilities Analyzed 321 Estimated maximum facility-wide 2001 individual cancer risk (in 1 million) Number of facilities with estimated facility-wide individual cancer risk 3 of 100-in-1 million or more Number of facilities at which the Natural Gas Transmission and Storage source category contributes 50 percent 1 or more to the facility-wideCancer individual cancer risks of 100-in-1Risk million or more Number of facilities with facility- wide individual cancer risk of 1-in-1 74 million or more Number of facilities at which the Natural Gas Transmission and Storage source category contributes 50 percent 10 or more to the facility-wide individual cancer risk of 1-in-1 million or more This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

253. Page 253 of 604 Maximum facility-wide chronic 80 noncancer TOSHI Number of facilities with facility-Chronic wide maximum noncancer TOSHI greater 30Noncancer than 1Risk Number of facilities at which the Natural Gas Transmission and Storage source category contributes 50 percent 0 or more to the facility-wide maximum noncancer TOSHI of 1 or more1 We note that the MIR would be 100-in-1 million if the CIIT UREfor formaldehyde were used instead of the IRIS URE. The facility-wide MIR from all HAP emissions at anyfacility that contains sources subject to the natural gastransmission and storage MACT standards is estimated to be200-in-1 million, based on actual emissions. Of the 321facilities included in this analysis, three have facility-wide MIR of 100-in-1 million or greater. The facility-wideMIR is 200-in-1 million at two of these facilities, drivenby formaldehyde from RICE.37 Another facility has afacility-wide risk of 100-in-1 million, with 90 percent ofthe risk attributed to natural gas transmission andstorage. There are 74 facilities with facility-wide MIR of1-in-1 million or greater. Of these facilities, 10 havenatural gas transmission and storage operations that37 We note that there is an ongoing IRIS reassessment forformaldehyde, and that future RTR risk assessments will use thecancer potency for formaldehyde that results from thatreassessment. As a result, the current results may not matchthose of future assessments. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

254. Page 254 of 604contribute greater than 50 percent to the facility-widerisks. As discussed above, we are proposing MACT standardsfor benzene emissions from small glycol dehydrators in thisaction. These standards would reduce the risk from benzeneemissions at facilities with natural gas transmission andstorage operations. The facility-wide cancer risks at thefacilities with risks of 1-in-1 million or more areprimarily driven by formaldehyde emissions from RICE, whichwill be assessed in a future RTR for that category. The facility-wide maximum individual chronic noncancerTOSHI is estimated to be 80, based on actual emissions. Ofthe 321 facilities included in this analysis, 30 havefacility-wide maximum chronic noncancer TOSHI valuesgreater than 1. Of these facilities, none had natural gastransmission and storage operations that contributedgreater than 50 percent to these facility-wide risks. Thechronic noncancer risks at these facilities are primarilydriven by acrolein emissions from RICE.iii. Demographic Risk Analysis Results The results of the demographic analyses performed toinvestigate the distribution of cancer risks at or above 1-in-1 million among the surrounding population aresummarized in Table 7 below. These results, for various This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

255. Page 255 of 604demographic groups, are based on actual emissions levelsfor the population living within 50 km of the facilities. TABLE 7. NATURAL GAS TRANSMISSION AND STORAGE DEMOGRAPHIC RISK ANALYSIS RESULTS Population with Cancer Risk at or Above 1-in-1 Million Due to…. Source Category Facility-Wide Nationwide HAP Emissions HAP EmissionsTotal 285,000,000 2,500 99,000Population Race by PercentWhite 75 92 58All Other 25 8 42Races Race by PercentWhite 75 92 58African 6 40 12AmericanNative 0.1 0.2 0.9AmericanOther and 1 2 12Multiracial Ethnicity by PercentHispanic 14 1 2Non-Hispanic 86 99 98 Income by PercentBelow Poverty 13 17 20LevelAbove poverty 87 83 80level Education by PercentOver 25 andwithout High 13 20 15SchoolDiplomaOver 25 andwith a High 87 80 85SchoolDiploma This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

256. Page 256 of 604 The results of the Natural Gas Transmission andStorage source category demographic analysis indicate thatthere are approximately 2,500 people exposed to a cancerrisk at or above 1-in-1 million due to emissions from thesource category, including an estimated 8 percent that areclassified as minority (listed as “All Other Races” inTable 7 above). Of the 2,500 people with estimated cancerrisks at or above 1-in-1 million from the source category,17 percent are in the “Below Poverty Level” demographicgroup, and 20 percent are in the “Over 25 and without HighSchool Diploma” demographic group, results which are 4 and7 percentage points higher, respectively, than thepercentages for these demographic groups across the UnitedStates. The percentages for the other demographic groupsare lower than their respective nationwide percentages. Thetable also shows that there are approximately 99,000 peopleexposed to an estimated cancer risk at or above 1-in-1million due to facility-wide emissions, including anestimated 42 percent that are classified as minority (“AllOther Races” in Table 7 above). Of the 99,000 people withestimated cancer risk at or above 1-in-1 million fromfacility-wide emissions, 40 percent are in the “AfricanAmerican” demographic group, 20 percent are in the “Below This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

257. Page 257 of 604Poverty Level” demographic group, and 15 percent are in the“Over 25 and without High School Diploma” demographicgroup, results which are 28, 7 and 2 percentage pointshigher, respectively, than the percentages for thesedemographic groups across the United States. Thepercentages for the other demographic groups are equal toor lower than their respective nationwide percentages.b. What are the proposed risk decisions for the Natural GasTransmission and Storage source category?i. Risk Acceptability In the risk analysis we performed for this sourcecategory, pursuant to CAA section 112(f)(2), we consideredthe available health information--the MIR; the numbers ofpersons in various risk ranges; cancer incidence; themaximum noncancer HI; the maximum acute noncancer hazard;the extent of noncancer risks; the potential for adverseenvironmental effects; distribution of risks in the exposedpopulation; and risk estimation uncertainty (54 FR 38044,September 14, 1989). For the Natural Gas Transmission and Storage sourcecategory, the risk analysis we performed indicates that thecancer risks to the individual most exposed could be ashigh as 90-in-1 million due to actual and allowable This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

258. Page 258 of 604emissions (30-in-1 million, based on the lower end of thebenzene URE range). These risks are near 100-in-1 million,which is the presumptive limit of acceptability. On theother hand, the risk analysis shows low cancer incidence (1case in every 1,000 years), low potential for adverseenvironmental effects or human health multi-pathway effectsand that chronic and acute noncancer health impacts areunlikely. We conclude that acute noncancer health impactsare unlikely for reasons similar to those described insection VII.C.2.b.i of this preamble. Our additional analysis of facility-wide risks showedthat, among three facilities with maximum facility-widecancer risk of 100-in-1 million or greater, one facilityhas a facility-wide cancer risk of 100-in-1 million, with90 percent of the risk attributed to natural gas andtransmission and storage. There are 30 facilities with amaximum chronic noncancer TOSHI greater than 1, but naturalgas transmission and storage operations did not drive thisrisk. In determining whether risk is acceptable, weconsidered the available health information, as describedabove. In this case, because the MIR is approaching, butstill less than 100-in-1 million risk, and because a number This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

259. Page 259 of 604of other factors indicate relatively low risk concern(e.g., low cancer incidence, low potential for adverseenvironmental effects or human health multi-pathwayeffects, chronic and acute noncancer health impactsunlikely), we are proposing to determine that the risks areacceptable.ii. Ample Margin of Safety We next considered whether the existing MACT standardprovides an ample margin of safety. In this analysis, weinvestigated available emissions control options that mightreduce the risk associated with emissions from the sourcecategory and considered this information, along with all ofthe health risks and other health information considered inthe risk acceptability determination. The estimated MIR of90-in-1 million discussed above is driven by the 0.9Mg/year benzene limit compliance alternative for the glycoldehydrator MACT standard in the current NESHAP. Removal ofthis compliance alternative would lower the MIR for thesource category to 20-in-1 million. We, therefore,considered removing this compliance alternative as anoption for reducing risk and assessed the cost of suchalternative. Without the compliance alternative, affectedglycol dehydrators (i.e., those units with annual average This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

260. Page 260 of 604benzene emissions of 0.9 Mg/yr or greater and an annualaverage natural gas throughput of 283,000 scmd or greater)must demonstrate compliance with the 95-percent controlrequirement, which we believe can be shown with theirexisting control devices in most cases, although, in someinstances, installation of a different or an additionalcontrol may be necessary. In section VII.B.1 above, we discuss the costs forrequiring controls on currently unregulated “small glycoldehydrators,” which are similar, in operation and type ofemission controls, to the dehydrators subject to thecurrent MACT (“large dehydrators”). The HAP costeffectiveness determined for small dehydrators at the floorlevel of control was $1,650/Mg. Although controlmethodologies are similar for large and small dehydrators,we expect that the costs for controls on large units couldbe as much as twice as high as for small units because ofthe large gas flow being processed. However, we also expectthat the amount of HAP emission reduction for the largedehydrators, in general, to be as much as, or more than,the amount achieved by small dehydrators. In light of theabove, we do not expect the cost effectiveness of thecontrol device needed to meet the 95-percent control This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

261. Page 261 of 604requirement for large dehydrators to exceed $3,300/Mg(i.e., twice the cost effectiveness for small dehydrators),which we consider to be reasonable. In accordance with the approach established in theBenzene NESHAP, the EPA weighed all health risk measuresand information considered in the risk acceptabilitydetermination, along with the costs and economic impacts ofemissions controls, technological feasibility,uncertainties and other relevant factors in making ourample margin of safety determination. Considering thehealth risk information and the reasonable costeffectiveness of the option identified, we propose that theexisting MACT standards, with the removal of the 0.9 Mgbenzene limit compliance option from the glycol dehydratorstandards, provide an ample margin of safety to protectpublic health. Pursuant to CAA section 112(f)(4), we are proposingthat this change (i.e., removal of the 0.9 Mg/yr compliancealternative) apply 90 days after its effective date. We arerequesting comment on whether or not there is sufficienttime for the large dehydrators that have been relying onthis compliance alternative to come into compliance withthe 95-percent control requirement or if additional time is This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

262. Page 262 of 604needed. See CAA section 112(f)(4)(A). We recognize that our proposal to remove the one-toncompliance alternative for the 95-percent control glycoldehydrator MACT standard could have negative impacts onsome sources that have come to rely on the flexibility thisalternative provides. We solicit comment on any suchimpacts and whether such impacts warrant adding a differentcompliance alternative that would result in less risk thanthe 0.9 Mg/yr benzene limit compliance option. If acommenter suggests a different compliance alternative, thecommenter should explain, in detail, what that alternativewould be, how it would work, and how it would reduce risk. As described above, we are proposing that the naturalgas transmission and storage MACT standards (with theremoval of the 0.9 Mg/yr benzene limit compliance option)provide an ample margin of safety to protect public health.We recognize that one facility has a facility-wide cancerrisk of 100-in-1 million, with 90 percent of the riskattributed to natural gas transmission and storage. Thisrisk is driven by benzene emissions from glycol dehydratorsand is being addressed by our proposed revision to theNatural Gas Transmission and Storage NESHAP (removal of the0.9 Mg/yr benzene limit compliance option). As previously This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

263. Page 263 of 604mentioned, two facilities have facility-wide MIR of 200-in-1 million, driven by formaldehyde from RICE. Emissions fromRICE are regulated under another source category and willbe assessed under a future RTR for that category.D. How did we perform the technology review and what arethe results and proposed decisions?1. What was the methodology for the technology review? Our technology review is focused on the identificationand evaluation of “developments in practices, processes,and control technologies” since the promulgation of theMACT standards for the two oil and gas source categories.If a review of available information identifies suchdevelopments, then we conduct an analysis of the technicalfeasibility of requiring the implementation of thesedevelopments, along with the impacts (costs, emissionreductions, risk reductions, etc.). We then make a decisionon whether it is necessary to amend the regulation torequire these developments. Based on specific knowledge of each source category,we began by identifying known developments in practices,processes and control technologies. For the purpose of thisexercise, we considered any of the following to be a“development”: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

264. Page 264 of 604• Any add-on control technology or other equipment that was not identified and considered during MACT development;• Any improvements in add-on control technology or other equipment (that was identified and considered during MACT development) that could result in significant additional emission reduction;• Any work practice or operational procedure that was not identified and considered during MACT development; and• Any process change or pollution prevention alternative that could be broadly applied that was not identified and considered during MACT development. In addition to looking back at practices, processes orcontrol technologies reviewed at the time we developed theMACT standards, we reviewed a variety of sources of data toaid in our evaluation of whether there were additionalpractices, processes or controls to consider. One of thesesources of data was subsequent air toxics rules. Since thepromulgation of the MACT standards for the sourcecategories addressed in this proposal, the EPA hasdeveloped air toxics regulations for a number of additionalsource categories. We reviewed the regulatory requirements This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

265. Page 265 of 604and/or technical analyses associated with these subsequentregulatory actions to identify any practices, processes andcontrol technologies considered in these efforts that couldpossibly be applied to emission sources in the sourcecategories under this current RTR review. We also consulted the EPA’s RBLC. The terms "RACT,""BACT," and "LAER" are acronyms for different programrequirements under the CAA provisions addressing the NAAQS.Control technologies classified as RACT, BACT or LAER applyto stationary sources depending on whether the sourceexists or is new and on the size, age and location of thefacility. The BACT and LAER (and sometimes RACT) aredetermined on a case-by-case basis, usually by state orlocal permitting agencies. The EPA established the RBLC toprovide a central database of air pollution technologyinformation (including technologies required in source-specific permits) to promote the sharing of informationamong permitting agencies and to aid in identifying futurepossible control technology options that might applybroadly to numerous sources within a category or apply onlyon a source-by-source basis. The RBLC contains over 5,000air pollution control permit determinations that can helpidentify appropriate technologies to mitigate many air This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

266. Page 266 of 604pollutant emission streams. We searched this database todetermine whether any practices, processes or controltechnologies are included for the types of processes usedfor emission sources (e.g., spray booths) in the sourcecategories under consideration in this proposal. We also consulted information from the Natural GasSTAR program. The Natural Gas STAR program is a flexible,voluntary partnership that encourages oil and natural gascompanies to adopt cost effective technologies andpractices that improve operational efficiency and reducepollutant emissions. The program provides the oil and gasindustry with information on new techniques anddevelopments to reduce pollutant emissions from the variousprocesses.2. What are the results and proposed decisions from thetechnology review? There are three types of emission sources covered bythe two oil and gas NESHAP. These sources and the controltechnologies (including add-on control devices and processmodifications) considered during the development of theMACT standards are: Glycol dehydrators (combustion devices,recovery devices, process modifications), storage vesselswith the PFE (combustion devices, recovery devices) and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

267. Page 267 of 604equipment leaks (LDAR programs, specific equipmentmodifications). Dehydrators are addressed by both 40 CFRpart 63, subpart HH and 40 CFR part 63, subpart HHH, whileequipment leaks and storage vessels with the PFE are onlycovered by subpart HH. Since the promulgation of 40 CFR part 63, subpart HH,which established MACT standards to address HAP emissionsfrom equipment leaks at gas processing plants, the EPA hasdeveloped LDAR programs that are more stringent than whatis required in subpart HH. The most prevalent differencesbetween these more stringent programs and subpart HH relateto the frequency of monitoring and the concentration whichconstitutes a “leak.” We do consider these programs torepresent a development in practices and evaluated whetherto revise the MACT standards for equipment leaks at naturalgas processing plants under subpart HH in light of thisdevelopment. An analysis was performed above in section VI.B.1 toassess the VOC reduction, costs and other impactsassociated with these more stringent LDAR program optionsat natural gas processing plants. One option considered wasto require compliance with 40 CFR part 60, subpart VVainstead of 40 CFR part 60, subpart VV (the current NSPS This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

268. Page 268 of 604requirement for equipment leaks of VOC at natural gasprocessing plants), which changes the leak definition(based on methane) from 10,000 ppm to 500 ppm and requiresmonitoring of connectors. Because the current leakdefinition under NESHAP 40 CFR part 63, subpart HH is thesame as that in NSPS subpart VV, and the ratio of VOC toHAP is approximately 20 to 1, we expect that the HAPreduction would be 1/20th of the VOC reduction undersubpart VVa. The estimated incremental cost for that optionwas determined to be $3,340 per ton of VOC. Based on the20-to-1 ratio, we estimate the incremental cost to controlHAP at the subpart VVa level would be approximately $66,800per ton of HAP ($73,480/Mg). Other options considered insection VI.B.1 of this preamble (and the incremental costof each option for reducing HAP) are as follows: The use ofan optical gas imaging camera monthly with an annual EPAMethod 21 check ($129,000 per ton of HAP/$143,600 per Mg,if purchasing the camera; $93,000 per ton of HAP/$103,300per Mg, if renting the camera); monthly optical gasimagining alone; and annual optical gas imaging.38 In light38 As stated above in section VI.B.1, emissions for the twooptions using the optical gas imaging camera alone cannot bequantified and, therefore, no cost effectiveness values weredetermined. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

269. Page 269 of 604of the above, we do not believe that the additional costsof these programs are justified. In addition to the plant-wide evaluations, a componentanalysis was also evaluated at gas processing plants forthe 40 CFR part 60, subpart VVa-level of control (option 1considered in section VI.B.1).39 That assessment shows thatthe subpart VVa-level of control for connectors has anincremental cost effectiveness of $4,360 per ton for VOCfor connectors and $144 per ton for VOC for valves. Thismeans the incremental cost to control HAP would beapproximately $87,200 per ton ($96,900/Mg) for connectorsand $2,880 per ton ($3,200/Mg) for valves. We do notbelieve the additional cost for the more stringentrequirement for connectors is justified, but the additionalcost for valves is justified. Therefore, we are proposingto revise the equipment leak requirements in 40 CFR part63, subpart HH to lower the leak definition for valves toan instrument reading of at least 500 ppm as a result ofour technology review.39 Because optical gas imaging is used to view several pieces ofequipment at a facility at once to survey for leaks, optionsinvolving imaging are not amenable to a component by componentanalysis. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

270. Page 270 of 604 Some of the practices, processes or controltechnologies listed by the Natural Gas STAR programapplicable to the emission sources in these categories werenot identified and evaluated during the original MACTdevelopment. While the Natural Gas STAR program doescontain information regarding new innovative techniquesthat are available to reduce HAP emissions, they are notconsidered to have emission reductions higher than what isset by the original MACT. One control technology identifiedin the Natural Gas STAR program that would result in no HAPemissions from glycol dehydration units would be thereplacement of a glycol dehydration unit with a desiccantdehydrator. This technology cannot be used for natural gasoperations with gas streams having high temperature, highvolume, and low pressure. Due to the limitations posed bythese conditions, we do not consider desiccant dehydratorsas MACT. For storage vessels, the applicable technologiesidentified by the Gas STAR program, which are evaluatedabove for proposal under NSPS in section VI.B.4, aresimilar to the cover and control technologies currentlyrequired for storage vessels under the existing MACT.Therefore, these technologies would not result in any This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

271. Page 271 of 604further emissions reductions than what is achieved by theoriginal MACT. Our review of the RBLC did not identify any practices,processes and control technologies applicable to theemission sources in these categories that were notidentified and evaluated during the original MACTdevelopment. In light of the above, we are not proposingany revisions to the existing MACT standards for storagevessels pursuant to section 112(d)(6) of the CAA.E. What other actions are we proposing?1. Combustion Control Device Testing As explained below in section VII.E.2, under ourproposal, performance testing would be required initiallyand every 5 years for non-condenser control devices.However, for certain enclosed combustion control devices,we are proposing to allow, as an alternative to on-sitetesting, a performance test conducted by a control devicemanufacturer in accordance with the procedures provided inthis proposal. We propose to allow a unit whose model meetsthe proposed performance criteria to claim a BTEX or HAPdestruction efficiency of 98 percent at the facility. Thisvalue is lower than the 99.9-percent destruction efficiencyrequired in the manufacturers’ test due to variations This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

272. Page 272 of 604between the test fuel specified and the gas streamscombusted at the actual facility. A source subject to thesmall dehydrator BTEX limit would use the 98-percentdestruction efficiency to calculate their dehydrator’s BTEXemissions for the purpose of demonstrating compliance. Forthe 95-percent control MACT standard, a control devicematching the tested model would be considered to meet thatrequirement. Once a device has been demonstrated to meetthe proposed performance criteria (and, therefore, isassigned a 98-percent destruction efficiency), installationof a unit matching the tested model at a facility wouldrequire no further performance testing (i.e., periodictests would not be required every 5 years). We are proposing this alternative to minimize issuesassociated with performance testing of certain combustioncontrol devices. We believe that testing units that are notconfigured with a distinct combustion chamber presentseveral technical issues that are more optimally addressedthrough manufacturer testing, and once these units areinstalled at a facility, through periodic inspection andmaintenance in accordance with manufacturers’recommendations. One issue is that an extension abovecertain existing combustion control device enclosures will This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

273. Page 273 of 604be necessary to get adequate clearance above the flamezone. Such extensions can more easily be configured by themanufacturer of the control device rather than having tomodify an extension in the field to fit devices at everysite. Issues related to transporting, installing andsupporting the extension in the field are also eliminatedthrough manufacturer testing. Another concern is that thepitot tube used to measure flow can be altered by radiantheat from the flame such that gas flow rates are notaccurate. This issue is best overcome by having themanufacturer select and use the pitot tube best suited totheir specific unit. For these reasons, we believe themanufacturers’ test is appropriate for these controldevices with ongoing performance ensured by periodicinspection and maintenance. This proposed alternative does not apply to flares, asdefined in 40 CFR 63.761 and 40 CFR 63.1271, which mustdemonstrate compliance by meeting the design and operationrequirements in 40 CFR 63.11(b), 40 CFR 63.772(e)(2) and 40CFR 63.1282(d)(2). It also would not apply to thermaloxidizers having a combustion chamber/firebox wherecombustion temperature and residence time can be measuredduring an on-site performance test and are valid indicators This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

274. Page 274 of 604of performance. These thermal oxidizers do not present theissues described above relative to on-site performancetesting and, therefore, do not need an alternative testingoption. The proposed alternative would, therefore, apply toenclosed combustion control devices except for thesethermal oxidizers. In conjunction with the proposed manufacturer testingalternative, we are proposing to add a definition for flareto clarify that flares, as referenced in the NESHAP (and towhich the proposed testing alternative does not apply),refers to a thermal oxidation system with an open flame(i.e., without enclosure). Accordingly, any thermaloxidation system that does not meet the proposed flaredefinition would be considered an enclosed combustioncontrol device. We estimate that there are many existing facilitiescurrently using enclosed combustion control devices thatwould be required to either conduct an on-site performancetest or install and operate a control device tested by themanufacturer under our proposal. Given the estimated numberof these combustion control devices in use, the timerequired for manufacturers to test and manufacture suchunits, we are proposing that existing sources have up to 3 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

275. Page 275 of 604years from the date of the final rules’ publication date tocomply with the initial performance testing requirements.2. Monitoring, Recordkeeping and Reporting We are proposing to make changes to the monitoringrequirements described below to address issues we haveidentified through a monitoring sufficiency reviewperformed during the RTR process. First, we are includingcalibration procedures associated with parametricmonitoring requirements in the existing NESHAP. The NESHAPrequire parametric monitoring of control device parameters(e.g., temperatures or flowrate monitoring), but did notinclude information on calibration or included inadequateinformation on calibration of monitoring devices.Therefore, we are specifying the calibration requirementsfor temperature and flow monitors that the NESHAP currentlylacks. In addition, under the current NESHAP, a designanalysis can be used in lieu of performance testing todemonstrate compliance and establish operating parameterlimits. We are proposing to allow the use of the designevaluation alternative only when the control device beingused is a condenser. The design evaluation option isappropriate for condensers because their emissions can be This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

276. Page 276 of 604accurately predicted using readily available physicalproperty information (e.g., vapor pressure data andcondensation calculations). In those cases, one would notneed to conduct emissions testing to determine actualemissions to demonstrate compliance with the MACT standard.For example, a requirement that "the temperature at theoutlet of the condenser shall be maintained at 50oFahrenheit below the condensation temperature calculatedfor the compound of interest using the reference equation”(e.g., National Institute of Standards and TechnologyChemistry WebBook at http://webbook.nist.gov/chemistry/) isadequate to assure proper operation of the condenser and,therefore, compliance with the required emission standard. For other types of control technologies, such ascarbon adsorption systems and enclosed combustion devices,40the ability to predict emissions depends on data developedby the vendor and such data may not reliably result in anaccurate prediction of emissions from a specific facility.There are variables (e.g., air to fuel ratios and wasteconstituents for combustion; varying organicconcentrations, constituents and capacity issues, including40 The design analysis alternative in the existing MACT does notapply to flares. As previously mentioned, the existing MACTprovides separate design and operation requirements for flares. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

277. Page 277 of 604break-through for carbon adsorption) that make theoreticalpredictions less reliable. The effects of these site-specific variables on emissions are not easily predictableand establishing monitoring conditions (e.g., combustiontemperature, vacuum regeneration) based on vendor data willlikely not account for those variables. Therefore, wepropose to eliminate the design evaluation alternative fornon-condenser controls. For non-condenser controls (and condensers not usingthe design analysis option), in addition to the initialcompliance testing, we are proposing that performance testsbe conducted at least once every 5 years and wheneversources desire to establish new operating limits. Under thecurrent NESHAP, a performance test is only conducted in twoinstances: (1) As an alternative to a design analysis fortheir compliance demonstration and identification ofoperating parameter ranges and (2) as a requirement toresolve a disagreement between the EPA and the owner oroperator regarding the design analysis. The current NESHAPdo not require additional performance testing beyond thesetwo cases (i.e., there is no periodic testing requirement).As mentioned above, we are proposing to remove the designevaluation option for non-condenser controls. For non- This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

278. Page 278 of 604condenser controls (and condensers not using the designanalysis option), the proposed periodic testing wouldensure compliance with the emission standards by verifyingthat the control device is meeting the necessary HAPdestruction efficiency determined in the initialperformance test. As discussed above in section VII.E.1, weare proposing that combustion control devices tested underthe manufacturers’ procedure are not required to conductperiodic testing. In addition, we are also proposing thatcombustion control devices that can demonstrate a uniformcombustion zone temperature meeting the required controlefficiency during the initial performance test are exemptfrom periodic testing. The requirement for continuousmonitoring of combustion zone temperature is an accurateindicator of control device performance and eliminates theneed for future testing. The current NESHAP (40 CFR 63.771(d) and 40 CFR63.1281(d)) require operating an enclosed combustion deviceat a minimum residence time of 0.5 seconds at a minimumtemperature of 760 degrees Celsius. We are proposing toremove the residence time requirement. The residence timerequirement is not needed because the compliancedemonstration made during the performance test is This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

279. Page 279 of 604sufficient to ensure that the combustion device hasadequate residence time to ensure the needed destructionefficiency. Therefore, we are proposing to remove theresidence time requirement. We are also clarifying at 40 CFR 63.773(d)(3)(i) and40 CFR 63.1283(d)(3)(i) for thermal vapor incinerators,boilers and process heaters, that the temperature sensorshall be installed at a location representative of thecombustion zone temperature. Currently, the regulationrequires that the temperature sensor be installed at alocation “downstream of the combustion zone” because we hadthought that the temperature downstream would berepresentative of combustion zone temperature. We have nowlearned that may or may not be the case. We are, therefore,proposing to amend this provision to more accuratelyreflect the intended requirement. Next, consistent with revisions for SSM, we’ve revised40 CFR 63.771(d)(4)(i) and 40 CFR 63.1281(d)(4)(i), exceptwhen maintenance or repair on a unit cannot be completedwithout a shutdown of the control device. Also, we’ve updated the criteria for prior performancetest results that can be used to demonstrate compliance inlieu of conducting a performance test. These updates ensure This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

280. Page 280 of 604that data for determining compliance are accurate, up-to-date, and truly representative of actual operatingconditions. In addition, we are proposing to revise thetemperature monitoring device minimum accuracy criteria in40 CFR 63.773(d)(3)(i) to better reflect the level ofperformance that is required of the temperature monitoringdevices. We believe that temperature monitoring devicescurrently used to meet the requirements of the NESHAP canmeet the proposed revised criteria without modification. Also, we are proposing to revise the calibration gasconcentration for the no detectable emissions procedureapplicable to closed vent systems in 40 CFR63.772(c)(4)(ii) from 10,000 ppmv to 500 ppmv methane to beconsistent with the leak threshold of 500 ppmv in 40 CFRpart 63, subpart HH. The current calibration level isinconsistent with achieving accurate readings at the levelnecessary to demonstrate there are no detectable emissions. Also, we are proposing recordkeeping and reportingrequirements for carbon adsorption systems. The currentNESHAP require the replacement of all carbon in the carbonadsorption system with fresh carbon on a regular,predetermined time interval that is no longer than the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

281. Page 281 of 604carbon service life established for the carbon system, butprovide no recordkeeping or reporting requirement todocument and assure compliance with this standard. Webelieve that maintaining some sort of log book is areasonable alternative combined with a requirement toreport instances when specified practices are not followed.Therefore, the proposed rule adds reporting andrecordkeeping requirements for establishing a schedule andmaintaining logs of carbon replacement. Finally, as noted above in section VII.B.1, we areproposing a BTEX emissions limit for small glycoldehydration unit process vents. For the compliancedemonstration, we propose that parametric monitoring of thecontrol device be performed. We believe that parametricmonitoring is adequate for glycol dehydrators in these twosource categories because temperature monitoring, whetherit be to verify proper condenser or combustion deviceoperation, is a reliable indicator of performance forreducing organic HAP emissions. We also considered the useof a continuous emissions monitoring system (CEMS) tomonitor compliance. However, for glycol dehydrators in theoil and natural gas sector, the necessary electricity,weather-protective enclosures and daily staffing are not This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

282. Page 282 of 604usually available. We, therefore, question the technicalfeasibility of operating a CEMS correctly in this sector.We request comment on the practicality of includingprovisions in the final rule for a CEMS to monitor BTEXemissions for small glycol dehydration units.3. Startup, Shutdown, Malfunction The United States Court of Appeals for the District ofColumbia Circuit vacated portions of two provisions in theEPA’s CAA section 112 regulations governing the emissionsof HAP during periods of SSM. Sierra Club v. EPA, 551 F.3d1019 (D.C. Cir. 2008), cert. denied, 130 S. Ct. 1735 (U.S.2010). Specifically, the Court vacated the SSM exemptioncontained in 40 CFR 63.6(f)(1) and 40 CFR 63.6(h)(1), thatis part of a regulation, commonly referred to as theGeneral Provisions Rule, that the EPA promulgated undersection 112 of the CAA. When incorporated into CAA section112(d) regulations for specific source categories, thesetwo provisions exempt sources from the requirement tocomply with the otherwise applicable CAA section 112(d)emission standard during periods of SSM. We are proposing the elimination of the SSM exemptionin the two oil and gas NESHAP. Consistent with Sierra Clubv. EPA, the EPA is proposing to apply the standards in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

283. Page 283 of 604these NESHAP at all times. In addition, we are proposing torevise 40 CFR 63.771(d)(4)(i) and 40 CFR 63.1281(d)(4)(i)to remove the provision allowing shutdown of the controldevice during maintenance or repair. We are also proposingseveral revisions to the General Provisions applicabilitytable for the MACT standard. For example, we are proposingto eliminate the incorporation of the General Provisions’requirement that the source develop a SSM plan. We are alsoproposing to eliminate or revise certain recordkeeping andreporting requirements related to the SSM exemption. TheEPA has attempted to ensure that we have not included inthe proposed regulatory language any provisions that areinappropriate, unnecessary or redundant in the absence ofthe SSM exemption. We are specifically seeking comment onwhether there are any such provisions that we haveinadvertently incorporated or overlooked. In proposing the MACT standards in these rules, theEPA has taken into account startup and shutdown periods. Webelieve that operations and emissions do not differ fromnormal operations during these periods such that itwarrants a separate standard. Therefore, we have notproposed different standards for these periods. Periods of startup, normal operations and shutdown are This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

284. Page 284 of 604all predictable and routine aspects of a source’soperations. However, by contrast, malfunction is defined asa “sudden, infrequent and not reasonably preventablefailure of air pollution control and monitoring equipment,process equipment or a process to operate in a normal orusual manner * * * ” (40 CFR 63.2). The EPA has determinedthat malfunctions should not be viewed as a distinctoperating mode and, therefore, any emissions that occur atsuch times do not need to be factored into development ofCAA section 112(d) standards, which, once promulgated,apply at all times. In Mossville Environmental Action Nowv. EPA, 370 F.3d 1232, 1242 (D.C. Cir. 2004), the Courtupheld as reasonable, standards that had factored invariability of emissions under all operating conditions.However, nothing in CAA section 112(d) or in case lawrequires that the EPA anticipate and account for theinnumerable types of potential malfunction events insetting emission standards. See Weyerhaeuser v. Costle, 590F.2d 1011, 1058 (D.C. Cir. 1978), (“In the nature ofthings, no general limit, individual permit, or even anyupset provision can anticipate all upset situations. Aftera certain point, the transgression of regulatory limitscaused by "uncontrollable acts of third parties," such as This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

285. Page 285 of 604strikes, sabotage, operator intoxication or insanity, and avariety of other eventualities, must be a matter for theadministrative exercise of case-by-case enforcementdiscretion, not for specification in advance byregulation.”). Further, it is reasonable to interpret CAA section112(d) as not requiring the EPA to account for malfunctionsin setting emissions standards. For example, we note thatCAA section 112 uses the concept of “best performing”sources in defining MACT, the level of stringency thatmajor source standards must meet. Applying the concept of“best performing” to a source that is malfunctioningpresents significant difficulties. The goal of bestperforming sources is to operate in such a way as to avoidmalfunctions of their units. Moreover, even if malfunctions were considered adistinct operating mode, we believe it would beimpracticable to take malfunctions into account in settingCAA section 112(d) standards for oil and natural gasproduction facility and natural gas transmission andstorage operations. As noted above, by definition,malfunctions are sudden and unexpected events, and it wouldbe difficult to set a standard that takes into account the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

286. Page 286 of 604myriad different types of malfunctions that can occuracross all sources in each source category. Moreover,malfunctions can also vary in frequency, degree andduration, further complicating standard setting. In the event that a source fails to comply with theapplicable CAA section 112(d) standards as a result of amalfunction event, the EPA would determine an appropriateresponse based on, among other things, the good faithefforts of the source to minimize emissions duringmalfunction periods, including preventative and correctiveactions, as well as root cause analyses to ascertain andrectify excess emissions. The EPA would also considerwhether the sources failure to comply with the CAA section112(d) standard was, in fact, “sudden, infrequent, notreasonably preventable” and was not instead “caused in partby poor maintenance or careless operation.” 40 CFR 63.2(definition of malfunction). Finally, the EPA recognizes that even equipment thatis properly designed and maintained can sometimes fail andthat such failure can sometimes cause or contribute to anexceedance of the relevant emission standard. (See, e.g.,State Implementation Plans: Policy Regarding ExcessiveEmissions During Malfunctions, Startup, and Shutdown This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

287. Page 287 of 604(September 20, 1999); Policy on Excess Emissions DuringStartup, Shutdown, Maintenance, and Malfunctions (February15, 1983)). The EPA is, therefore, proposing to add to thefinal rule an affirmative defense to civil penalties forexceedances of emission limits that are caused bymalfunctions in both of the MACT standards addressed inthis proposal. See 40 CFR 63.761 for sources subject to theoil and natural gas production MACT standards, or 40 CFR63.1271 for sources subject to the natural gas transmissionand storage MACT standards (defining “affirmative defense”to mean, in the context of an enforcement proceeding, aresponse or defense put forward by a defendant, regardingwhich the defendant has the burden of proof and the meritsof which are independently and objectively evaluated in ajudicial or administrative proceeding). We also areproposing other regulatory provisions to specify theelements that are necessary to establish this affirmativedefense; a source subject to the oil and natural gasproduction facilities or natural gas transmission MACTstandards must prove by a preponderance of the evidencethat it has met all of the elements set forth in 40 CFR63.762 and a source subject to the natural gas transmissionand storage facilities MACT standards must prove by a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

288. Page 288 of 604preponderance of the evidence that it has met all of theelements set forth in 40 CFR 63.1272. (See 40 CFR 22.24.)The criteria ensure that the affirmative defense isavailable only where the event that causes an exceedance ofthe emission limit meets the narrow definition ofmalfunction in 40 CFR 63.2 (sudden, infrequent, notreasonably preventable and not caused by poor maintenanceand or careless operation). For example, to successfullyassert the affirmative defense, the source must prove by apreponderance of evidence that excess emissions “[w]erecaused by a sudden, infrequent, and unavoidable failure ofair pollution control and monitoring equipment, processequipment, or a process to operate in a normal or usualmanner ….” The criteria also are designed to ensure thatsteps are taken to correct the malfunction, to minimizeemissions in accordance with 40 CFR 63.762 for sourcessubject to the oil and natural gas production facilitiesMACT standards or 40 CFR 63.1272 for sources subject to thenatural gas transmission and storage facilities MACTstandards and to prevent future malfunctions. For example,the source must prove by a preponderance of evidence that“[r]epairs were made as expeditiously as possible when theapplicable emission limitations were being exceeded…” and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

289. Page 289 of 604that “[a]ll possible steps were taken to minimize theimpact of the excess emissions on ambient air quality, theenvironment and human health….” In any judicial oradministrative proceeding, the Administrator may challengethe assertion of the affirmative defense and, if therespondent has not met its burden of proving all of therequirements in the affirmative defense, appropriatepenalties may be assessed in accordance with section 113 ofthe CAA (see also 40 CFR 22.77).4. Applicability and Compliancea. Calculating Potential to Emit (PTE) We are proposing to amend section 40 CFR63.760(a)(1)(iii) to clarify that sources must use a glycolcirculation rate consistent with the definition of PTE in40 CFR 63.2 in calculating emissions for purposes ofdetermining PTE. Affected parties have misinterpreted thecurrent language concerning measured values or annualaverage to apply to a broader range of parameters than wasintended. Those qualifiers were meant to apply to gascharacteristics that are measured, such as inlet gascomposition, pressure and temperature rather than processequipment settings. That means that the circulation rateused in PTE determinations shall be the maximum under its This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

290. Page 290 of 604physical and operational design. In addition to the proposed changes described above,we are seeking comment on several PTE related issues.According to the data available to the Administrator, when40 CFR part 63, subpart HH was promulgated, the level ofHAP emissions was predominantly driven by natural gasthroughput (i.e., HAP emissions went up or down in concertwith natural gas throughput). Since promulgation, we havelearned that there is not always a direct correlationbetween HAP emissions and natural gas throughput. We havereceived information suggesting that, in some cases, HAPemissions can increase despite decreasing natural gasthroughput due to changes in gas composition. We are askingfor comment regarding the likelihood of this occurrence anddata demonstrating the circumstances where it occurs. Inlight of the potential issue, we are asking for commentregarding the addition of provisions in the NESHAP torequire area sources to recalculate their PTE to confirmthat they are indeed area sources and whether thatcalculation should be performed on an annual or biannualbasis to verify that changes in gas composition have notincreased their emissions.b. Definition of Facility and Applicability Criteria This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

291. Page 291 of 604 Subpart HH of 40 CFR part 63 (section 63.760(a)(2))currently defines facilities as those where hydrocarbonliquids are processed, upgraded or stored prior to thepoint of custody transfer or where natural gas isprocessed, upgraded or stored prior to entering the NaturalGas Transmission and Storage source category. We areproposing to remove the references to “point of custodytransfer” and “transmission and storage source categories”from the definition because the operations performed at asite sufficiently define a facility and the scope of thesubpart is specified already under 40 CFR 63.760. Inaddition, we are removing the custody transfer referencefrom the applicability criteria in 40 CFR 63.760(a)(2).Since hydrocarbon liquids can pass through several custodytransfer points between the well and the final destination,the custody transfer criteria is not clear enough. We are,therefore, proposing to replace the reference to “point ofcustody transfer” with a more specific description of thepoint up to which the subpart applies (i.e., the pointwhere hydrocarbon liquids enter either the organic liquidsdistribution or petroleum refineries source categories) andexclude custody transfer from that criteria. We believethis change eliminates ambiguity and is consistent with the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

292. Page 292 of 604oil and natural gas production-specific provisions in theorganic liquids distribution MACT.5. Other Proposed Changes to Clarify These Rules The following lists additional changes to the NESHAPwe are proposing. This list includes proposed rule changesthat address editorial corrections and plain languagerevisions: • Revise 40 CFR 63.769(b) to clarify that the equipment leak provisions in 40 CFR part 63, subpart HH do not apply to a source if that source is required to control equipment leaks under either 40 CFR part 63, subpart H or 40 CFR part 60, subpart KKK. The current 40 CFR 63.769(b), which states that subpart HH does not apply if a source meets the requirements in either of the subparts mentioned above, does not clearly express our intent that such source must be implementing the LDAR provisions in the other 40 CFR part 60 or 40 CFR part 63 subparts to qualify for the exemption. • Revise 40 CFR 63.760(a)(1) to clarify that an existing area source that increases its emissions to major source levels has up to the first substantive compliance date to either reduce its emissions below This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

293. Page 293 of 604 major source levels by obtaining a practically enforceable permit or comply with the applicable major source provisions of 40 CFR part 63, subpart HH. We have revised the second to last sentence in 40 CFR 63.760(a)(1) by removing the parenthetical statement because it simply reiterates the last sentence of this section and is, therefore, unnecessary. • Revise 40 CFR 63.771(d)(1)(ii) and 40 CFR 63.1281(d)(1)(ii) to clarify that the vapor recovery device and “other control device” described in those provisions refer to non-destructive control devices only. • Revise the last sentence of 40 CFR 63.764(i) and 40 CFR 63.1274(g) to clarify the requirements following an unsuccessful attempt to repair a leak. • Updated the email and physical address for area source reporting in 40 CFR 63.775(c)(1).VIII. What are the cost, environmental, energy and economicimpacts of the proposed 40 CFR part 60, subpart OOOO andamendments to subparts HH and HHH of 40 CFR part 63? We are presenting a combined discussion of theestimates of the impacts for the proposed 40 CFR part 60, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

294. Page 294 of 604subpart OOOO and proposed amendments to 40 CFR part 63,subpart HH and 40 CFR part 63, subpart HHH. The cost,environmental and economic impacts presented in thissection are expressed as incremental differences betweenthe impacts of an oil and natural gas facility complyingwith the amendments to subparts HH and HHH and newstandards under 40 CFR 60, subpart OOOO and the baseline,i.e., the standards before these amendments. The impactsare presented for the year 2015, which will be the yearthat all existing oil and natural gas facilities will haveto be in compliance, and also the year that will representapproximately 5 years of construction of new oil andnatural gas facilities subject to the NSPS emissionslimits. The analyses and the documents referenced below canbe found in Docket ID Numbers EPA-HQ-OAR-2007-0877 and EPA-HQ-OAR-2002-0051.A. What are the affected sources? We expect that by 2015, the year when all existingsources will be required to come into compliance in theUnited States, there will be 97 oil and natural gasproduction facilities and 15 natural gas transmission andstorage facilities with one or more existing glycoldehydration units. We also estimate that there will be an This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

295. Page 295 of 604additional 329 (there are 47 facilities that already havean affected glycol dehydration unit) existing oil andnatural gas production facilities with existing storagevessels that we expect to be affected by these finalamendments. These facilities operate approximately 134glycol dehydration units (115 in production and 19 intransmission and storage) and 1,970 storage vessels.Approximately 10 oil and natural gas production and twotransmission and storage facilities would have new glycoldehydration units and 38 production facilities would havenew dehydration units. We expect new production facilitieswould operate approximately 12 production glycoldehydration units and 197 storage vessels and newtransmission and storage would operate approximately twoglycol dehydration units. Based on data provided by the United States EnergyInformation Administration, we anticipate that by 2015there will be approximately 21,800 gas wellhead facilities,790 reciprocating compressors, 30 centrifugal compressors,14,000 pneumatic devices and 300 storage vessels subject tothe new NSPS for VOC. Some of these affected facilitieswill be built at existing facilities and some at newgreenfield facilities. Based on data limitations, we assume This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

296. Page 296 of 604impacts are equal regardless of location. There are about 21 glycol dehydration units with highenough HAP emissions that we believe cannot meet theemissions limit without using more than one controltechnique. In developing the cost impacts, we assume thatthey would require multiple controls. The controls forwhich we have detailed cost data are condensers and VRU, sowe developed costs for both controls to develop what weconsider to be a reasonable cost estimate for thesefacilities. This does not imply that we believe thesefacilities will specifically use a combination of acondenser and vapor recovery limit, but we do believe thecombination of these control results is a reasonableestimate of cost.B. How are the impacts for this proposal evaluated? For these proposed Oil and Natural Gas Production andNatural Gas Transmission and Storage NESHAP amendments andNSPS, the EPA used two models to evaluate the impacts ofthe regulation on the industry and the economy. Typically,in a regulatory analysis, the EPA determines the regulatoryoptions suitable to meet statutory obligations under theCAA. Based on the stringency of those options, the EPA thendetermines the control technologies and monitoring This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

297. Page 297 of 604requirements that sources might rationally select to complywith the regulation. This analysis is documented in anengineering analysis. The selected control technologies andmonitoring requirements are then evaluated in a cost modelto determine the total annualized control costs. Theannualized control costs serve as inputs to an EconomicImpact Analysis model that evaluates the impacts of thosecosts on the industry and society as a whole. The Economic Impact Analysis used the National EnergyModeling System (NEMS) to estimate the impacts of theproposed NSPS on the United States energy system. The NEMSis a publically-available model of the United States energyeconomy developed and maintained by the Energy InformationAdministration of the United States DOE and is used toproduce the Annual Energy Outlook, a reference publicationthat provides detailed forecasts of the energy economy fromthe current year to 2035. The impacts we estimated includedchanges in drilling activity, price and quantity changes inthe production and consumption of crude oil and natural gasand changes in international trade of crude oil and naturalgas. We evaluated whether and to what extent the increasedproduction costs imposed by the NSPS might alter the mix offuels consumed at a national level. Additionally, we This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

298. Page 298 of 604combined estimated emissions co-reductions of methane fromthe engineering analysis with NEMS analysis to estimate thenet change in CO2e GHG from energy-related sources.C. What are the air quality impacts? For the oil and natural gas sector NESHAP and NSPS, weestimated the emission reductions that will occur due tothe implementation of the final emission limits. The EPAestimated emission reductions based on the controltechnologies selected by the engineering analysis. Theseemission reductions associated with the proposed amendmentsto 40 CFR part 63, subpart HH and 40 CFR part 63, subpartHHH are based on the estimated population in 2008. Underthe proposed limits for glycol dehydration units andstorage vessels, we have estimated that the HAP emissionsreductions will be 1,400 tpy for existing units subject tothe proposed emissions limits. For the NSPS, we estimated the emission reductionsthat will occur due to the implementation of the finalemission limits. The EPA estimated emission reductionsbased on the control technologies selected by theengineering analysis. These emission reductions are basedon the estimated population in 2015. Under the proposedNSPS, we have estimated that the emissions reductions will This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

299. Page 299 of 604be 540,000 tpy VOC for affected facilities subject to theNSPS. The control strategies likely adopted to meet theproposed NESHAP amendments and the proposed NSPS willresult in concurrent control of HAP, methane and VOCemissions. We estimate that direct reductions in HAP,methane and VOC for the proposed rules combined total about38,000 tpy, 3.4 million tpy and 540,000 tpy, respectively. Under the final standards, new monitoring requirementsare being added.D. What are the water quality and solid waste impacts? We estimated minimal water quality impacts for theproposed amendments and proposed NSPS. For the proposedamendments to the NESHAP, we anticipate that the waterimpacts associated with the installation of a condensersystem for the glycol dehydration unit process vent wouldbe minimal. This is because the condensed water collectedwith the hydrocarbon condensate can be directed back intothe system for reprocessing with the hydrocarbon condensateor, if separated, combined with produced water fordisposal, usually by reinjection. Similarly, the water impacts associated withinstallation of a vapor control system either on a glycol This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

300. Page 300 of 604dehydration unit or a storage vessel would be minimal. Thisis because the water vapor collected along with thehydrocarbon vapors in the vapor collection and redirectsystem can be directed back into the system forreprocessing with the hydrocarbon condensate or, ifseparated, combined with the produced water for disposalfor reinjection. There would be no water impacts expected forfacilities subject to the proposed NSPS. Further, we do notanticipate any adverse solid waste impacts from theimplementation of the proposed NESHAP amendments and theproposed NSPS.E. What are the secondary impacts? Indirect or secondary air quality impacts includeimpacts that will result from the increased electricityusage associated with the operation of control devices, aswell as water quality and solid waste impacts (which werejust discussed) that might occur as a result of theseproposed actions. We estimate the proposed amendments to 40CFR part 63, subpart HH and 40 CFR part 63, subpart HHHwill increase emissions of criteria pollutants due to thepotential use of flares for the control of storage vessels.We do not estimate an increased energy demand associated This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

301. Page 301 of 604with the installation of condensers, VRU or flares. Theincreases in criteria pollutant emissions associated withthe use of flares to control storage vessels subject toexisting source standards are estimated to be 5,500 tpy ofCO2, 16 tpy of carbon monoxide (CO), 3 tpy of NOx, less than1 tpy of particulate matter (PM) and 6 tpy totalhydrocarbons. For storage vessels subject to new sourcestandards, increases in secondary air pollutants areestimated to be less than 900 tpy of CO2, 3 tpy of CO, 1tpy of NOx, 1 tpy of PM and 1 tpy total hydrocarbons. In addition, we estimate that the secondary impactsassociated with the pneumatic controller requirements tocomply with the proposed NSPS would be about 22 tpy of CO2,1 tpy of NOx and 3 tpy PM. For gas wellhead affectedfacilities, we estimate that the use of flares would resultin increases in criteria pollutant emissions of about990,000 tons of CO2, 2,800 tpy of CO, 500 tpy of NOx, 5 tpyof PM and 1,000 tpy total hydrocarbons.F. What are the energy impacts? Energy impacts in this section are those energyrequirements associated with the operation of emissioncontrol devices. Potential impacts on the national energy This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

302. Page 302 of 604economy from the rule are discussed in the economic impactssection. There would be little national energy demandincrease from the operation of any of the control optionsanalyzed under the proposed NESHAP amendments and proposedNSPS. The proposed NESHAP amendments and proposed NSPSencourage the use of emission controls that recoverhydrocarbon products, such as methane and condensate thatcan be used on-site as fuel or reprocessed within theproduction process for sale. We estimated that the proposedstandards will result in a net cost savings due to therecovery of salable natural gas and condensate. Thus, thefinal standards have a positive impact associated with therecovery of non-renewable energy resources.G. What are the cost impacts? The estimated total capital cost to comply with theproposed amendments to 40 CFR part 63, subpart HH for majorsources in the Oil and Natural Gas Production sourcecategory is approximately $51.5 million. The total capitalcost for the proposed amendments to 40 CFR part 63, subpartHHH for major sources in the Natural Gas Transmission andStorage source category is estimated to be approximately$370 thousand. All costs are in 2008 dollars. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

303. Page 303 of 604 The total estimated net annual cost to industry tocomply with the proposed amendments to 40 CFR part 63,subpart HH for major sources in the Oil and Natural GasProduction source category is approximately $16 million.The total net annual cost for proposed amendments to 40 CFRpart 63, subpart HHH for major sources in the Natural GasTransmission and Storage source category is estimated to beapproximately $360,000. These estimated annual costsinclude: (1) The cost of capital, (2) operating andmaintenance costs, (3) the cost of monitoring, inspection,recordkeeping and reporting (MIRR) and (4) any associatedproduct recovery credits. All costs are in 2008 dollars. The estimated total capital cost to comply with theproposed NSPS is approximately $740 million in 2008dollars. The total estimated net annual cost to industry tocomply with the proposed NSPS is approximately $740 millionin 2008 dollars. This annual cost estimate includes: (1)The cost of capital, (2) operating and maintenance costsand (3) the cost of MIRR. This estimated annual cost doesnot take into account any producer revenues associated withthe recovery of salable natural gas and hydrocarboncondensates. When revenues from additional product recovery are This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

304. Page 304 of 604considered, the proposed NSPS is estimated to result in anet annual engineering cost savings overall. When includingthe additional natural gas recovery in the engineering costanalysis, we assume that producers are paid $4 per thousandcubic feet (Mcf) for the recovered gas at the wellhead.The engineering analysis cost analysis assumes the value ofrecovered condensate is $70 per barrel. Based on theengineering analysis, about 180,000,000 Mcf (180 billioncubic feet) of natural gas and 730,000 barrels ofcondensate are estimated to be recovered by controlrequirements in 2015. Using the price assumptions, theestimated revenues from natural gas product recovery areapproximately $780 million in 2008 dollars. This savings isestimated at $45 million in 2008 dollars. Using the engineering cost estimates, estimatednatural gas product recovery, and natural gas product priceassumptions, the net annual engineering cost savings isestimated for the proposed NSPS at about $45 million in2008 dollars. Totals may not sum due to independentrounding. As the price assumption is very influential onestimated annualized engineering costs, we performed asimple sensitivity analysis of the influence of the assumed This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

305. Page 305 of 604wellhead price paid to natural gas producers on the overallengineering annualized costs estimate of the proposed NSPS.At $4.22/Mcf, the price forecast reported in the 2011Annual Energy Outlook in 2008 dollars, the annualized costsare estimated at about -$90 million, which wouldapproximately double the estimate of net cost savings ofthe proposed NSPS. As indicated by this difference, EPA haschosen a relatively conservative assumption (leading to anestimate of few savings and higher net costs) for theengineering costs analysis. The natural gas price at whichthe proposed NSPS breaks-even from an estimated engineeringcosts perspective is around $3.77/Mcf. A $1/Mcf change inthe wellhead natural gas price leads to about a $180million change in the annualized engineering costs of theproposed NSPS. Consequently, annualized engineering costsestimates would increase to about $140 million under a$3/Mcf price or decrease to about -$230 million under a$5/Mcf price. For further details on this sensitivityanalysis, please refer the regulatory impact analysis (RIA)for this rulemaking located in the docket.H. What are the economic impacts? The NEMS analysis of energy system impacts for theproposed NSPS option estimates that domestic natural gas This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

306. Page 306 of 604production is likely to increase slightly (about 20 billioncubic feet or 0.1 percent) and average natural gas pricesto decrease slightly ($0.04 per Mcf in 2008 dollars or 0.9percent at the wellhead for onshore producers in the lower48 states) for 2015, the year of analysis. This increase inproduction and decrease in wellhead price is largely aresult of the increased natural gas and condensate recoveryas a result of complying with the NSPS. Domestic crude oilproduction is not expected to change, while average crudeoil prices are estimated to decrease slightly ($0.02/barrelin 2008 dollars or less than 0.1 percent at the wellheadfor onshore producers in the lower 48 states) in the yearof analysis, 2015. The NEMS-based analysis estimates in theyear of analysis, 2015, that net imports of natural gas andcrude will not change significantly. Total CO2e emissions from energy-related sources areexpected to increase about 2.0 million metric tons CO2e or0.04 percent under the proposed NSPS, according to the NEMSanalysis. This increase is attributable largely to naturalgas consumption increases. This estimate does not includeCO2e reductions from the implementation of the controls;these reductions are discussed in more detail in the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

307. Page 307 of 604benefits section that follows. We did not estimate the energy economy impacts of theproposed NESHAP amendments using NEMS, as the expectedcosts of the rule are not likely to have estimable impactson the national energy economy.I. What are the benefits? The proposed Oil and Natural Gas NSPS and NESHAPamendments are expected to result in significant reductionsin existing emissions and prevent new emissions fromexpansions of the industry. These proposed rules combinedare anticipated to reduce 38,000 tons of HAP, 540,000 tonsof VOC and 3.4 million tons of methane. These pollutantsare associated with substantial health effects, welfareeffects and climate effects. With the data available, weare not able to provide credible health benefit estimatesfor the reduction in exposure to HAP, ozone and PM (2.5microns and less) (PM2.5) for these rules, due to thedifferences in the locations of oil and natural gasemission points relative to existing information and thehighly localized nature of air quality responses associatedwith HAP and VOC reductions. This is not to imply that there are no benefits of therules; rather, it is a reflection of the difficulties in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

308. Page 308 of 604modeling the direct and indirect impacts of the reductionsin emissions for this industrial sector with the datacurrently available. In addition to health improvements,there will be improvements in visibility effects, ecosystemeffects and climate effects, as well as additional productrecovery. Although we do not have sufficient information ormodeling available to provide quantitative estimates forthis rulemaking, we include a qualitative assessment of thehealth effects associated with exposure to HAP, ozone andPM2.5 in the RIA for this rule. These qualitative effectsare briefly summarized below, but for more detailedinformation, please refer to the RIA, which is available inthe docket. One of the HAP of concern from the oil andnatural gas sector is benzene, which is a known humancarcinogen, and formaldehyde, which is a probable humancarcinogen. VOC emissions are precursors to both PM2.5 andozone formation. As documented in previousanalyses (U.S. EPA, 200641 and U.S. EPA, 201042), exposure to41 U.S. EPA. RIA. National Ambient Air Quality Standards forParticulate Matter, Chapter 5. Office of Air Quality Planning andStandards, Research Triangle Park, NC. October 2006. Available onthe Internet at<http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%205--Benefits.pdf>. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

309. Page 309 of 604PM2.5 and ozone is associated with significant public healtheffects. PM2.5 is associated with health effects such aspremature mortality for adults and infants, cardiovascularmorbidity, such as heart attacks, hospital admissions andrespiratory morbidity such as asthma attacks, acute andchronic bronchitis, hospital and emergency room visits,work loss days, restricted activity days and respiratorysymptoms, as well as visibility impairment.43 Ozone isassociated with health effects such as respiratorymorbidity such as asthma attacks, hospital and emergencydepartment visits, school loss days and prematuremortality, as well as injury to vegetation and climateeffects.44 In addition to the improvements in air quality andresulting benefits to human health and non-climate welfare42 U.S. EPA. RIA. National Ambient Air Quality Standards forOzone. Office of Air Quality Planning and Standards, ResearchTriangle Park, NC. January 2010. Available on the Internet at<http://www.epa.gov/ttn/ecas/regdata/RIAs/s1-supplemental_analysis_full.pdf>.43 U.S. EPA. Integrated Science Assessment for Particulate Matter(Final Report). EPA-600-R-08-139F. National Center forEnvironmental Assessment—RTP Division. December 2009. Availableat <http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546>.44 U.S. EPA. Air Quality Criteria for Ozone and RelatedPhotochemical Oxidants (Final). EPA/600/R-05/004aF-cF.Washington, DC: U.S. EPA. February 2006. Available on theInternet athttp://cfpub.epa.gov/ncea/CFM/recordisplay.cfm?deid=149923. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

310. Page 310 of 604effects previously discussed, this proposed rule isexpected to result in significant climate co-benefits dueto anticipated methane reductions. Methane is a potent GHGthat, once emitted into the atmosphere, absorbs terrestrialinfrared radiation, which contributes to increased globalwarming and continuing climate change. Methane reacts inthe atmosphere to form ozone and ozone also impacts globaltemperatures. According to the Intergovernmental Panel onClimate Change (IPCC) 4th Assessment Report (2007), methaneis the second leading long-lived climate forcer after CO2globally. Total methane emissions from the oil and gasindustry represent about 40 percent of the totalmethane emissions from all sources and account for about 5percent of all CO2e emissions in the United States, withnatural gas systems being the single largest contributor toUnited States anthropogenic methane emissions.45 Methane, inaddition to other GHG emissions, contributes to warming ofthe atmosphere, which, over time, leads to increased airand ocean temperatures, changes in precipitation patterns,melting and thawing of global glaciers and ice,45 U.S. EPA (2011), 2011 U.S. Greenhouse Gas Inventory ReportExecutive Summary available on the internet atwww.epa.gov/climateexchange/emissions/downloads11/US-GHG-Inventory-2011-Executive Summary.pdf. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

311. Page 311 of 604increasingly severe weather events, such as hurricanes ofgreater intensity and sea level rise, among other impacts. This rulemaking proposes emission control technologiesand regulatory alternatives that will significantlydecrease methane emissions from the oil and natural gassector in the United States. The regulatory alternativesproposed for the NESHAP and the NSPS are expected to reducemethane emissions annually by about 3.4 million short tonsor 65 million metric tons CO2e. After considering thesecondary impacts of this proposal previously discussed,such as increased CO2 emissions from well completioncombustion and decreased CO2e emissions because of fuel-switching by consumers, the methane reductions become about62 million metric tons CO2e. These reductions representabout 26 percent of the baseline methane emissions for thissector reported in the EPA’s U.S. Greenhouse Gas InventoryReport for 2009 (251.55 million metric tons CO2e whenpetroleum refineries and petroleum transportation areexcluded because these sources are not examined in thisproposal). After considering the secondary impacts of thisproposal, such as increased CO2 emissions from wellcompletion combustion and decreased CO2 emissions because This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

312. Page 312 of 604of fuel-switching by consumers, the CO2e GHG reductions arereduced to about 62 million metric tons CO2e. However, itis important to note that the emission reductions are basedupon predicted activities in 2015; the EPA did not forecastsector-level emissions in 2015 for this rulemaking. Theseemission reductions equate to the climate benefits oftaking approximately 11 million typical passenger cars offthe road or eliminating electricity use from about 7million typical homes each year.46 The EPA recognizes that the methane reductions proposedin this rule will provide for significant economic climatebenefits to society just described. However, there is nointeragency-accepted methodology to place monetary valueson these benefits. A ‘global warming potential (GWP)approach’ of converting methane to CO2e using the GWP ofmethane provides an approximation method for estimating themonetized value of the methane reductions anticipated fromthis rule. This calculation uses the GWP of the non-CO2 gasto estimate CO2 equivalents and then multiplies these CO2equivalent emission reductions by the social cost of carbon46 U.S. EPA. Greenhouse Gas Equivalency Calculator available at:http://www.epa.gov/cleanenergy/energy-resources/calculator.htmlaccessed 07/19/11. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

313. Page 313 of 604developed by the Interagency Social Cost of Carbon WorkGroup to generate monetized estimates of the benefits. The social cost of carbon is an estimate of the netpresent value of the flow of monetized damages from a 1-metric ton increase in CO2 emissions in a given year (orfrom the alternative perspective, the benefit to society ofreducing CO2 emissions by 1 ton). For more informationabout the social cost of carbon, see the Support Document:Social Cost of Carbon for Regulatory Impact Analysis UnderExecutive Order 1286647 and RIA for the Light-Duty VehicleGHG rule.48 Applying this approach to the methane reductionsestimated for the proposed NESHAP and NSPS of the oil andgas rule, the 2015 climate co-benefits vary by discountrate and range from about $370 million toapproximately $4.7 billion; the mean social cost of carbonat the 3-percent discount rate results in an estimate ofabout $1.6 billion in 2015.47 Interagency Working Group on Social Cost of Carbon (IWGSC).2010. Technical Support Document: Social Cost of Carbon forRegulatory Impact Analysis Under Executive Order 12866. Docket IDEPA-HQ-OAR-2009-0472-114577.<http://www.epa.gov/otaq/climate/regulations/scc-tsd.pdf>Accessed March 30, 2011.48 U.S. EPA. Final Rulemaking: Light-Duty Vehicle Greenhouse GasEmissions Standards and Corporate Average Fuel Economy Standards.May 2010. Available on the Internet athttp://www.epa.gov/otaq/climate/regulations.htm#finalR. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

314. Page 314 of 604 The ratio of domestic to global benefits of emissionreductions varies with key parameter assumptions. Forexample, with a 2.5 or 3 percent discount rate, the U.S.benefit is about 7-10 percent of the global benefit, onaverage, across the scenarios analyzed. Alternatively, ifthe fraction of GDP lost due to climate change is assumedto be similar across countries, the domestic benefit wouldbe proportional to the U.S. share of global GDP, which iscurrently about 23 percent. On the basis of this evidence,values from 7 to 23 percent should be used to adjust theglobal SCC to calculate domestic effects. It is recognizedthat these values are approximate, provisional and highlyspeculative. There is no a priori reason why domesticbenefits should be a constant fraction of net globaldamages over time.49 These co-benefits equate to a range of approximately$110 to $1,400 per short ton of methane reduced, dependingupon the discount rate assumed with a per ton estimate of$480 at the 3-percent discount rate. Methane climate co-benefit estimates for additional regulatory alternativesare included in the RIA for this proposed rule. These49 Interagency Working Group on Social Cost of Carbon (IWGSC).2010. Technical Support Document: Social Cost of Carbon forRegulatory Impact Analysis Under Executive Order 12866. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

315. Page 315 of 604social cost of methane benefit estimates are not the sameas would be derived from direct computations (using theintegrated assessment models employed to develop theInteragency Social Cost of Carbon estimates) for a varietyof reasons, including the shorter atmospheric lifetime ofmethane relative to CO2 (about 12 years compared to CO2whose concentrations in the atmosphere decay on timescalesof decades to millennia). The climate impacts also differbetween the pollutants for reasons other than the radiativeforcing profiles and atmospheric lifetimes of these gases. Methane is a precursor to ozone and ozone is a short-lived climate forcer that contributes to global warming.The use of the IPCC Second Assessment Report GWP toapproximate co-benefits may underestimate the directradiative forcing benefits of reduced ozone levels and doesnot capture any secondary climate co-benefits involved withozone-ecosystem interactions. In addition, a recent EPANational Center of Environmental Economics workingpaper suggests that this quick ‘GWP approach’ to benefitsestimation will likely understate the climate benefits ofmethane reductions in most cases.50 This conclusion is50 Marten and Newbold (2011), Estimating the Social Cost of Non-CO2 GHG Emissions: Methane and Nitrous Oxide, NCEE Working Paper This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

316. Page 316 of 604reached using the 100-year GWP for methane of 25 as putforth in the IPCC Fourth Assessment Report (AR 4), asopposed to the lower value of 21 used in this analysis.Using the higher GWP estimate of 25 would increase thesereported methane climate co-benefit estimates by about 19percent. Although the IPCC Assessment Report (AR4)suggested a GWP of 25 for methane, the EPA has used GWP of21 to estimate the methane climate co-benefits for this oiland gas proposal in order to provide estimates moreconsistent with global GHG inventories, which currently useGWP from the IPCC Second Assessment Report. Due to the uncertainties involved with the ‘GWPapproach’ estimates presented and methane climate co-benefits estimates available in the literature, the EPAchooses not to compare these co-benefit estimates to thecosts of the rule for this proposal. Rather, the EPApresents the ‘GWP approach’ climate co-benefit estimates asan interim method to produce these estimates until theInteragency Social Cost of Carbon Work Group developsvalues for non-CO2 GHG. The EPA requests comments frominterested parties and the public about this interimSeries #11-01.http://yosemite.epa.gov/EE/epa/eed.nsf/WPNumber/2011-01?OpenDocument. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

317. Page 317 of 604approach specifically and more broadly about appropriatemethods to monetize the climate benefits of methanereductions. In particular, the EPA seeks public comments tothis proposed rulemaking regarding social cost of methaneestimates that may be used to value the co-benefits ofmethane emission reductions anticipated for the oil and gasindustry from this rule. Comments specific to whether GWPis an acceptable method for generating a placeholder valuefor the social cost of methane until interagency-modeledestimates become available are welcome. Public comments maybe provided in the official docket for this proposedrulemaking in accordance with the process outlined earlierin this notice. These comments will be considered indeveloping the final rule for this rulemaking. For the proposed NESHAP amendments, a break-evenanalysis suggests that HAP emissions would need to bevalued at $12,000 per ton for the benefits to exceed thecosts if the health, ecosystem and climate benefits fromthe reductions in VOC and methane emissions are assumed tobe zero. Even though emission reductions of VOC and methaneare co-benefits for the proposed NESHAP amendments, theyare legitimate components of the total benefit-costcomparison. If we assume the health benefits from HAP This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

318. Page 318 of 604emission reductions are zero, the VOC emissions would needto be valued at $1,700 per ton or the methane emissionswould need to be valued at $3,300 per ton for the co-benefits to exceed the costs. All estimates are in 2008dollars. For the proposed NSPS, the revenue from additionalproduct recovery exceeds the costs, which renders a break-even analysis unnecessary when these revenues are includedin the analysis. Based on the methodology from Fann,Fulcher, and Hubbell (2009),51 ranges of benefit-per-tonestimates for emissions of VOC indicate that on average inthe United States, VOC emissions are valued from $1,200 to$3,000 per ton as a PM2.5 precursor, but emission reductionsin specific areas are valued from $280 to $7,000 per ton in2008 dollars. As a result, even if VOC emissions from oiland natural gas operations result in monetized benefitsthat are substantially below the national average, there isa reasonable chance that the benefits of the rule wouldexceed the costs, especially if we were able to monetizeall of the additional benefits associated with ozoneformation, visibility, HAP and methane.51 Fann, N., C.M. Fulcher, B.J. Hubbell. The influence oflocation, source, and emission type in estimates of the humanhealth benefits of reducing a ton of air pollution. Air QualAtmos Health (2009) 2:169-176. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

319. Page 319 of 604IX. Request for Comments We are soliciting comments on all aspects of thisproposed action. All comments received during the commentperiod will be considered. In addition to general commentson the proposed actions, we are also interested in anyadditional data that may help to reduce the uncertaintiesinherent in the risk assessments. We are specificallyinterested in receiving corrections to the datasets usedfor MACT analyses and risk modeling. Such data shouldinclude supporting documentation in sufficient detail toallow characterization of the quality andrepresentativeness of the data or information. Please seethe following section for more information on submittingdata.X. Submitting Data Corrections The facility-specific data used in the source categoryrisk analyses, facility-wide analyses and demographicanalyses for each source category subject to this actionare available for download on the RTR Web page athttp://www.epa.gov/ttn/atw/rrisk/rtrpg.html. These datafiles include detailed information for each HAP emissionsrelease point at each facility included in the sourcecategory and all other HAP emissions sources at these This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

320. Page 320 of 604facilities (facility-wide emissions sources). However, itis important to note that the source category risk analysisincluded only those emissions tagged with the MACT codeassociated with the source category subject to the riskanalysis. If you believe the data are not representative or areinaccurate, please identify the data in question, provideyour reason for concern and provide any “improved” datathat you have, if available. When you submit data, werequest that you provide documentation of the basis for therevised values to support your suggested changes. To submitcomments on the data downloaded from the RTR Web page,complete the following steps: 1. Within this downloaded file, enter suggestedrevisions to the data fields appropriate for thatinformation. The data fields that may be revised includethe following: Data Element Definition Are control measures in place?Control Measure (yes or no) Select control measure fromControl Measure Comment list provided and briefly describe the control measure Indicate here if the facilityDelete or record should be deleted Describes the reason forDelete Comment deletion This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

321. Page 321 of 604 Data Element Definition Code description of the methodEmission Calculation used to derive emissions. ForMethod Code For Revised example, CEM, materialEmissions balance, stack test, etc. Enter the general type of emission process associatedEmission Process Group with the specified emission point Enter release angle (clockwise from true North); orientation of the y-dimension relative toFugitive Angle true North, measured positive for clockwise starting at 0 degrees (maximum 89 degrees) Enter dimension of the source in the east-west (x-)Fugitive Length direction, commonly referred to as length (ft) Enter dimension of the source in the north-south (y-)Fugitive Width direction, commonly referred to as width (ft) Enter total annual emissionsMalfunction Emissions due to malfunctions (TPY) Enter maximum hourlyMalfunction Emissions Max malfunction emissions hereHourly (lb/hr) Enter datum for latitude/longitude coordinatesNorth American Datum (NAD27 or NAD83); if left blank, NAD83 is assumed Enter general comments aboutProcess Comment process sources of emissions Enter revised physical streetREVISED Address address for MACT facility hereREVISED City Enter revised city name hereREVISED County Name Enter revised county name hereREVISED Emission Release Enter revised Emission ReleasePoint Type Point Type hereREVISED End Date Enter revised End Date here Enter revised Exit GasREVISED Exit Gas Flow Rate Flowrate here (ft3/sec) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

322. Page 322 of 604 Data Element DefinitionREVISED Exit Gas Enter revised Exit GasTemperature Temperature here (OF) Enter revised Exit GasREVISED Exit Gas Velocity Velocity here (ft/sec) Enter revised FacilityREVISED Facility Category Category Code here, whichCode indicates whether facility is a major or area source Enter revised Facility NameREVISED Facility Name here Enter revised FacilityREVISED Facility Registry Registry Identifier here,Identifier which is an ID assigned by the EPA Facility Registry SystemREVISED HAP Emissions Enter revised HAP EmissionsPerformance Level Code Performance Level here Enter revised Latitude hereREVISED Latitude (decimal degrees) Enter revised Longitude hereREVISED Longitude (decimal degrees)REVISED MACT Code Enter revised MACT Code here Enter revised Pollutant CodeREVISED Pollutant Code here Enter revised routineREVISED Routine Emissions emissions value here (TPY)REVISED SCC Code Enter revised SCC Code here Enter revised Stack DiameterREVISED Stack Diameter here (ft) Enter revised Stack HeightREVISED Stack Height here (Ft)REVISED Start Date Enter revised Start Date hereREVISED State Enter revised state hereREVISED Tribal Code Enter revised Tribal Code hereREVISED Zip Code Enter revised Zip Code here Enter total annual emissionsShutdown Emissions due to shutdown events (TPY)Shutdown Emissions Max Enter maximum hourly shutdownHourly emissions here (lb/hr) Enter general comments aboutStack Comment emission release points Enter total annual emissionsStartup Emissions due to startup events (TPY) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

323. Page 323 of 604 Data Element DefinitionStartup Emissions Max Enter maximum hourly startupHourly emissions here (lb/hr) Enter date facility stoppedYear Closed operations 2. Fill in the commenter information fields for eachsuggested revision (i.e., commenter name, commenterorganization, commenter email address, commenter phonenumber and revision comments). 3. Gather documentation for any suggested emissionsrevisions (e.g., performance test reports, material balancecalculations, etc.). 4. Send the entire downloaded file with suggestedrevisions in Microsoft® Access format and all accompanyingdocumentation to Docket ID Number EPA-HQ-OAR-2010-0505(through one of the methods described in the ADDRESSESsection of this preamble). To expedite review of therevisions, it would also be helpful if you submitted a copyof your revisions to the EPA directly at RTR@epa.gov inaddition to submitting them to the docket. 5. If you are providing comments on a facility withmultiple source categories, you need only submit one filefor that facility, which should contain all suggestedchanges for all source categories at that facility. We This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

324. Page 324 of 604request that all data revision comments be submitted in theform of updated Microsoft® Access files, which are providedon the http://www.epa.gov/ttn/atw/rrisk/rtrpg.html Webpage.XI. Statutory and Executive Order ReviewsA. Executive Order 12866: Regulatory Planning and Reviewand Executive Order 13563: Improving Regulation andRegulatory Review Under Executive Order 12866 (58 FR 51735, October 4,1993), this action is an “economically significantregulatory action” because it is likely to have an annualeffect on the economy of $100 million or more. Accordingly,the EPA submitted this action to OMB for review underExecutive Order 12866 and Executive Order 13563 (76 FR3821, January 21, 2011) and any changes made in response toOMB recommendations have been documented in the docket forthis action. In addition, the EPA prepared a RIA of the potentialcosts and benefits associated with this action. The RIAavailable in the docket describes in detail the empiricalbasis for the EPA’s assumptions and characterizes thevarious sources of uncertainties affecting the estimatesbelow. Table 8 shows the results of the cost and benefits This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

325. Page 325 of 604analysis for these proposed rules. For more information onthe benefit and cost analysis, as well as details on theregulatory options considered, please refer to the RIA forthis rulemaking, which is available in the docket. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

326. Page 326 of 604Table 8. Summary of the Monetized Benefits, Costs and NetBenefits for the Proposed Oil and Natural Gas NSPS andNEHSAP Amendments in 2015 (millions of 2008$)1 Proposed NSPS and Proposed NESHAP Proposed NSPS NESHAP Amendments Amendments CombinedTotalMonetized N/A N/A N/ABenefits2Total Costs3 -$45 million $16 million -$29 millionNet Benefits N/A N/A N/A 37,000 tons of 1,400 tons of 38,000 tons of HAP HAP HAP 540,000 tons of 540,000 tons of 9,200 tons of VOC VOC VOC 3.4 million tons 4,900 tons of 3.4 million tonsNon-monetized of methane methane of methane Benefits4,5 Health effects of HAP exposure Health effects of PM2.5 and ozone exposure Visibility impairment Vegetation effects Climate effects1 All estimates are for the implementation year (2015).2 While we expect that these avoided emissions will result inimprovements in air quality and reductions in health effects associatedwith HAP, ozone and PM, as well as climate effects associated withmethane, we have determined that quantification of those benefitscannot be accomplished for this rule in a defensible way. This is notto imply that there are no benefits of the rules; rather, it is areflection of the difficulties in modeling the direct and indirectimpacts of the reductions in emissions for this industrial sector withthe data currently available.3 The engineering compliance costs are annualized using a 7-percentdiscount rate. The negative cost for the proposed NSPS reflects theinclusion of revenues from additional natural gas and hydrocarboncondensate recovery that are estimated as a result of the proposedNSPS.4 For the NSPS, reduced exposure to HAP and climate effects are co-benefits. For the NESHAP, reduced VOC emissions, PM2.5 and ozoneexposure, visibility and vegetation effects and climate effects are co-benefits.5 The specific control technologies for these proposed rules areanticipated to have minor secondary disbenefits. The net CO2-equivalentemission reductions are 93,000 metric tons for the NESHAP and 62million metric tons for the NSPS.B. Paperwork Reduction Act This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

327. Page 327 of 604 The information collection requirements in thisproposed action have been submitted for approval to OMBunder the Paperwork Reduction Act, 44 U.S.C. 3501, et seq.The ICR document prepared by the EPA has been assigned EPAICR Numbers 1716.07 (40 CFR part 60, subpart OOOO), 1788.10(40 CFR part 63, subpart HH), 1789.07 (40 CFR part 63,subpart HHH) and 1086.10 (40 CFR part 60, subparts KKK andsubpart LLL). The information to be collected for the proposed NSPSand the proposed NESHAP amendments are based onnotification, recordkeeping and reporting requirements inthe NESHAP General Provisions (40 CFR part 63, subpart A),which are mandatory for all operators subject to nationalemission standards. These recordkeeping and reportingrequirements are specifically authorized by section 114 ofthe CAA (42 U.S.C. 7414). All information submitted to theEPA pursuant to the recordkeeping and reportingrequirements for which a claim of confidentiality is madeis safeguarded according to Agency policies set forth in 40CFR part 2, subpart B. These proposed rules would require maintenanceinspections of the control devices, but would not requireany notifications or reports beyond those required by the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

328. Page 328 of 604General Provisions. The recordkeeping requirements requireonly the specific information needed to determinecompliance. For sources subject to the proposed NSPS, burdenchanges associated with these amendments result from therespondents’ annual reporting and recordkeeping burdenassociated with this proposed rule for this collection(averaged over the first 3 years after the effective dateof the standards). The burden is estimated to be 560,000labor hours at a cost of $18 million per year. Thisincludes the burden previously estimated for sourcessubject to 40 CFR part 60, subpart KKK (which is beingincorporated into 40 CFR part 60, subpart OOOO). Theaverage hours and cost per regulated entity subject to theNSPS for oil and natural gas production and natural gastransmissions and distribution facilities would be110 hours per response and $3,693 per response, based on anaverage of 1,459 operators responding per year and 16responses per year. The estimated recordkeeping and reporting burden afterthe effective date of the proposed amendments is estimatedfor all affected major and area sources subject to the Oiland Natural Gas Production NESHAP to be approximately This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

329. Page 329 of 60463,000 labor hours per year at a cost of $2.1 million peryear. For the Natural Gas Transmission and Storage NESHAP,the recordkeeping and reporting burden is estimated to be2,500 labor hours per year at a cost of $86,800 per year.This estimate includes the cost of reporting, includingreading instructions and information gathering.Recordkeeping cost estimates include reading instructions,planning activities and conducting compliance monitoring.The average hours and cost per regulated entity subject tothe Oil and Natural Gas Production NESHAP would be 72 hoursper year and $2,500 per year, based on an average of846 facilities per year and three responses per facility.For the Natural Gas Transmission and Storage NESHAP, theaverage hours and cost per regulated entity would be 50hours per year and $1,600 per year, based on an average of53 facilities per year and three responses per facility.Burden is defined at 5 CFR 1320.3(b). An agency may not conduct or sponsor, and a person isnot required to respond to, a collection of informationunless it displays a currently valid OMB control number.The OMB control numbers for the EPAs regulations in 40 CFRare listed in 40 CFR part 9. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

330. Page 330 of 604 To comment on the Agencys need for this information,the accuracy of the provided burden estimates and anysuggested methods for minimizing respondent burden, the EPAhas established a public docket for this rule, whichincludes this ICR, under Docket ID Number EPA-HQ-OAR-2010-0505. Submit any comments related to the ICR to the EPA andOMB. See ADDRESSES section at the beginning of this noticefor where to submit comments to the EPA. Send comments toOMB at the Office of Information and Regulatory Affairs,Office of Management and Budget, 725 17th Street, NW,Washington, DC 20503, Attention: Desk Office for the EPA.Since OMB is required to make a decision concerning the ICRbetween 30 and 60 days after [INSERT DATE OF PUBLICATION INTHE FEDERAL REGISTER], a comment to OMB is best assured ofhaving its full effect if OMB receives it by [INSERT DATE30 DAYS FROM DATE OF PUBLICATION IN THE FEDERAL REGISTER].The final rule will respond to any OMB or public commentson the information collection requirements contained inthis proposal.C. Regulatory Flexibility Act The Regulatory Flexibility Act generally requires anagency to prepare a regulatory flexibility analysis of anyrule subject to notice and comment rulemaking requirements This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

331. Page 331 of 604under the Administrative Procedure Act or any otherstatute, unless the agency certifies that the rule will nothave a significant economic impact on a substantial numberof small entities (SISNOSE). Small entities include smallbusinesses, small organizations, and small governmentaljurisdictions. For purposes of assessing the impact of thisrule on small entities, a small entity is defined as: (1) Asmall business whose parent company has no more than 500employees (or revenues of less than $7 million for firmsthat transport natural gas via pipeline); (2) a smallgovernmental jurisdiction that is a government of a city,county, town, school district, or special district with apopulation of less than 50,000; and (3) a smallorganization that is any not-for-profit enterprise which isindependently owned and operated and is not dominant in itsfield.Proposed NSPS After considering the economic impact of the proposedNSPS on small entities, I certify that this action will nothave a SISNOSE. The EPA performed a screening analysis forimpacts on a sample of expected affected small entities bycomparing compliance costs to entity revenues. Based uponthe analysis in the RIA, which is in the Docket, EPA This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

332. Page 332 of 604concludes the number of impacted small businesses isunlikely to be sufficiently large to declare a SISNOSE.Our judgment in this determination is informed by the factthat many affected firms are expected to receive revenuesfrom the additional natural gas and condensate recoveryengendered by the implementation of the controls evaluatedin this RIA. As much of the additional natural gas recoveryis estimated to arise from completion-related activities,we expect the impact on well-related compliance costs to besignificantly mitigated. This conclusion is enhancedbecause the returns to REC activities occur without asignificant time lag between implementing the control andobtaining the recovered product, unlike many controloptions where the emissions reductions accumulate over longperiods of time; the reduced emission completions andrecompletions occur over a short span of time, during whichthe additional product recovery is also accomplished.Proposed NESHAP Amendments After considering the economic impact of the proposedNESHAP amendments on small entities, I certify that thisaction will not have a SISNOSE. Based upon the analysis inthe RIA, which is in the Docket, we estimate that 62 of the118 firms (53 percent) that own potentially affected This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

333. Page 333 of 604facilities are small entities. The EPA performed ascreening analysis for impacts on all expected affectedsmall entities by comparing compliance costs to entityrevenues. Among the small firms, 52 of the 62 (84 percent)are likely to have impacts of less than 1 percent in termsof the ratio of annualized compliance costs to revenues.Meanwhile, 10 firms (16 percent) are likely to have impactsgreater than 1 percent. Four of these 10 firms are likelyto have impacts greater than 3 percent. While these 10firms might receive significant impacts from the proposedNESHAP amendments, they represent a very small slice of theoil and gas industry in its entirety, less than 0.2 percentof the estimated 6,427 small firms in NAICS 211. Althoughthis final rule will not impact a substantial number ofsmall entities, the EPA, nonetheless, has tried to reducethe impact of this rule on small entities by setting thefinal emissions limits at the MACT floor, the leaststringent level allowed by law. We continue to be interested in the potential impactsof the proposed rule on small entities and welcome commentson issues related to such impacts.D. Unfunded Mandates Reform Act This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

334. Page 334 of 604 This action contains no Federal mandates under theprovisions of title II of the Unfunded Mandates Reform Actof 1995 (UMRA), 2 U.S.C. 1531-1538 for state, local ortribal governments or the private sector. This proposedrule does not contain a Federal mandate that may result inexpenditures of $100 million or more for state, local andtribal governments, in the aggregate, or to the privatesector in any one year. Thus, this proposed rule is notsubject to the requirements of sections 202 or 205 of UMRA.This proposed rule is also not subject to the requirementsof section 203 of UMRA because it contains no regulatoryrequirements that might significantly or uniquely affectsmall governments. This action contains no requirementsthat apply to such governments nor does it imposeobligations upon them. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

335. Page 335 of 604E. Executive Order 13132: Federalism This proposed rule does not have federalismimplications. It will not have substantial direct effectson the states, on the relationship between the nationalgovernment and the states, or on the distribution of powerand responsibilities among the various levels ofgovernment, as specified in Executive Order 13132. Thus,Executive Order 13132 does not apply to this proposed rule.In the spirit of Executive Order 13132 and consistent withthe EPA policy to promote communications between the EPAand state and local governments, the EPA specificallysolicits comment on this proposed rule from state and localofficials.F. Executive Order 13175: Consultation and Coordinationwith Indian Tribal Governments This action does not have tribal implications, asspecified in Executive Order 13175 (65 FR 67249, November9, 2000). It will not have substantial direct effect ontribal governments, on the relationship between the Federalgovernment and Indian tribes or on the distribution ofpower and responsibilities between the Federal governmentand Indian tribes, as specified in Executive Order 13175.Thus, Executive Order 13175 does not apply to this action. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

336. Page 336 of 604 The EPA specifically solicits additional comment onthis proposed action from tribal officials.G. Executive Order 13045: Protection of Children fromEnvironmental Health Risks and Safety Risks This proposed rule is not subject to Executive Order13045 (62 FR 19885, April 23, 1997) because the Agency doesnot believe the environmental health risks or safety risksaddressed by this action present a disproportionate risk tochildren. This actions’ health and risk assessments arecontained in section VII.C of this preamble.The public is invited to submit comments or identify peer-reviewed studies and data that assess effects of early lifeexposure to HAP from oil and natural gas sector activities.H. Executive Order 13211: Actions Concerning RegulationsThat Significantly Affect Energy Supply, Distribution orUse Executive Order 13211, (66 FR 28,355, May 22, 2001),provides that agencies shall prepare and submit to theAdministrator of the Office of Information and RegulatoryAffairs, OMB, a Statement of Energy Effects for certainactions identified as significant energy actions. Section4(b) of Executive Order 13211 defines “significant energyactions” as “any action by an agency (normally published in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

337. Page 337 of 604the Federal Register) that promulgates or is expected tolead to the promulgation of a final rule or regulation,including notices of inquiry, advance notices of proposedrulemaking, and notices of proposed rulemaking: (1)(i) Thatis a significant regulatory action under Executive Order12866 or any successor order and (ii) is likely to have asignificant adverse effect on the supply, distribution, oruse of energy; or (2) that is designated by theAdministrator of the Office of Information and RegulatoryAffairs as a significant energy action.” The proposed rules will result in the addition ofcontrol equipment and monitoring systems for existing andnew sources within the oil and natural gas industry. Theproposed NESHAP amendments are unlikely to have asignificant adverse effect on the supply, distribution oruse of energy. As such, the proposed NESHAP amendments arenot “significant energy actions” as defined in ExecutiveOrder 13211 (66 FR 28355, May 22, 2001). The proposed NSPS is also unlikely to have asignificant effect on the supply, distribution or use ofenergy. As such, the proposed NSPS is not a “significantenergy action” as defined in Executive Order 13211 (66 FR This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

338. Page 338 of 60428355, May 22, 2001). The basis for the determination is asfollows. As discussed in the impacts section of the Preamble,we use the NEMS to estimate the impacts of the proposedNSPS on the United States energy system. The NEMS is apublically available model of the United States energyeconomy developed and maintained by the Energy InformationAdministration of the United States DOE and is used toproduce the Annual Energy Outlook, a reference publicationthat provides detailed forecasts of the United Statesenergy economy. Proposed emission controls for the NSPS capture VOCemissions that otherwise would be vented to the atmosphere.Since methane is co-emitted with VOC, a large proportion ofthe averted methane emissions can be directed into naturalgas production streams and sold. One pollution controlrequirement of the proposed NSPS also captures saleablecondensates. The revenues from additional natural gas andcondensate recovery are expected to offset the costs ofimplementing the proposed NSPS. The analysis of energy impacts for the proposed NSPSthat includes the additional product recovery shows thatdomestic natural gas production is estimated to increase This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

339. Page 339 of 604(20 billion cubic feet or 0.1 percent) and natural gasprices to decrease ($0.04/Mcf or 0.9 percent at thewellhead for producers in the lower 48 states) in 2015, theyear of analysis. Domestic crude oil production is notestimated to change, while crude oil prices are estimatedto decrease slightly ($0.02/barrel or less than 0.1 percentat the wellhead for producers in the lower 48 states) in2015, the year of analysis. All prices are in 2008 dollars.Additionally, the NSPS establishes several performancestandards that give regulated entities flexibility indetermining how to best comply with the regulation. In anindustry that is geographically and economicallyheterogeneous, this flexibility is an important factor inreducing regulatory burden. For more information on the estimated energy effects,please refer to the economic impact analysis for thisproposed rule. The analysis is available in the RIA, whichis in the public docket.I. National Technology Transfer and Advancement Act Section 12(d) of the National Technology Transfer andAdvancement Act of 1995 (NTTAA), Public Law No. 104-113 (15U.S.C. 272 note) directs the EPA to use voluntary consensusstandards (VCS) in its regulatory activities unless to do This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

340. Page 340 of 604so would be inconsistent with applicable law or otherwiseimpractical. VCS are technical standards (e.g., materialsspecifications, test methods, sampling procedures, andbusiness practices) that are developed or adopted by VCSbodies. NTTAA directs the EPA to provide Congress, throughOMB, explanations when the Agency decides not to useavailable and applicable VCS. The proposed rule involves technical standards.Therefore, the requirements of the NTTAA apply to thisaction. We are proposing to revise 40 CFR part 63, subpartHH and 40 CFR part 63, subpart HHH to allow ANSI/ASME PTC19.10–1981, Flue and Exhaust Gas Analyses (Part 10,Instruments and Apparatus) to be used in lieu of EPAMethods 3B, 6 and 16A. This standard is available from theAmerican Society of Mechanical Engineers (ASME), Three ParkAvenue, New York, NY 10016-5990. Also, we are proposing torevise subpart HHH to allow ASTM D6420-99 (2004), TestMethod for Determination of Gaseous Organic Compounds byDirect Interface Gas Chromatography/Mass Spectrometry, tobe used in lieu of EPA Method 18. For a detailed discussionof this VCS, and its appropriateness as a substitute forMethod 18, see the final Oil and Natural Gas ProductionNESHAP (Area Sources) (72 FR 36, January 3, 2007). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

341. Page 341 of 604 As a result, the EPA is proposing ASTM D6420-99 (2004)for use in 40 CFR part 63, subpart HHH. The EPA alsoproposes to allow Method 18 as an option in addition toASTM D6420-99 (2004). This would allow the continued use ofgas chromatography configurations other than gaschromatography/mass spectrometry. The EPA welcomes comments on this aspect of theproposed rulemaking and, specifically, invites the publicto identify potentially-applicable VCS and to explain whysuch standards should be used in this regulation.J. Executive Order 12898: Federal Actions to AddressEnvironmental Justice in Minority Populations and Low-Income Populations Executive Order 12898 (59 FR 7629, February 16, 1994)establishes Federal executive policy on EJ. Its mainprovision directs Federal agencies, to the greatest extentpracticable and permitted by law, to make EJ part of theirmission by identifying and addressing, as appropriate,disproportionately high and adverse human health orenvironmental effects of their programs, policies andactivities on minority populations and low-incomepopulations in the United States. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

342. Page 342 of 604 The EPA has determined that this proposed rule willnot have disproportionately high and adverse human healthor environmental effects on minority or low-incomepopulations because it increases the level of environmentalprotection for all affected populations without having anydisproportionately high and adverse human health orenvironmental effects on any population, including anyminority or low-income population. To examine the potential for any EJ issues that mightbe associated with each source category, we evaluated thedistributions of HAP-related cancer and noncancer risksacross different social, demographic and economic groupswithin the populations living near the facilities wherethese source categories are located. The methods used toconduct demographic analyses for this rule are described insection VII.C of the preamble for this rule. Thedevelopment of demographic analyses to inform theconsideration of EJ issues in EPA rulemakings is anevolving science. The EPA offers the demographic analysesin this proposed rulemaking as examples of how suchanalyses might be developed to inform such consideration,and invites public comment on the approaches used and theinterpretations made from the results, with the hope that This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

343. Page 343 of 604this will support the refinement and improve utility ofsuch analyses for future rulemakings. For the demographic analyses, we focused on thepopulations within 50 km of any facility estimated to haveexposures to HAP which result in cancer risks of 1-in-1million or greater, or noncancer HI of 1 or greater (basedon the emissions of the source category or the facility,respectively). We examined the distributions of those risksacross various demographic groups, comparing thepercentages of particular demographic groups to the totalnumber of people in those demographic groups nationwide.The results, including other risk metrics, such as averagerisks for the exposed populations, are documented in sourcecategory-specific technical reports in the docket for bothsource categories covered in this proposal. As described in the preamble, our risk assessmentsdemonstrate that the regulations for the oil and naturalgas production and natural gas transmission and storagesource categories, are associated with an acceptable levelof risk and that the proposed additional requirements willprovide an ample margin of safety to protect public health.Our analyses also show that, for these source categories,there is no potential for an adverse environmental effect This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

344. Page 344 of 604or human health multi-pathway effects, and that acute andchronic noncancer health impacts are unlikely. The EPA hasdetermined that, although there may be an existingdisparity in HAP risks from these sources between somedemographic groups, no demographic group is exposed to anunacceptable level of risk. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

345. Standards of Performance for New Stationary Sources: Oiland Natural Gas Production and Natural Gas Transmission andDistribution; National Emission Standards for Hazardous Air Pollutants From Oil and Natural Gas Production Facilities; and National Emission Standards for Hazardous Air Pollutants From Natural Gas Transmission and Storage Facilities--Page 345 of 604List of Subjects in 40 CFR Part 60Environmental protection, Air pollution control, Reportingand recordkeeping requirements, Volatile organic compounds.List of Subjects in 40 CFR Part 63Environmental protection, Air pollution control, Reportingand recordkeeping requirements, Volatile organic compounds.____________________________Dated:____________________________Lisa P. Jackson,Administrator. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

346. Page 346 of 604For the reasons set out in the preamble, title 40, chapterI of the Code of Federal Regulations is proposed to beamended as follows:PART 60--[AMENDED] 1. The authority citation for part 60 continues toread as follows:Authority: 42 U.S.C. 7401, et seq. 2. Section 60.17 is amended by: a. Revising paragraph (a)(7); and b. Revising paragraphs (a)(91) and (a)(92) to read asfollows:§60.17 Incorporations by reference.* * * * * (a) * * * (7) ASTM D86–78, 82, 90, 93, 95, 96, Distillation ofPetroleum Products, IBR approved for §§60.562–2(d),60.593(d), 60.593a(d), 60.633(h) and 60.5401(h).* * * * * (91) ASTM E169–63, 77, 93, General Techniques ofUltraviolet Quantitative Analysis, IBR approved for§§60.485a(d)(1), 60.593(b)(2), 60.593a(b)(2), 60.632(f) and60.5400(f). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

347. Page 347 of 604 (92) ASTM E260–73, 91, 96, General Gas ChromatographyProcedures, IBR approved for §§60.485a(d)(1), 60.593(b)(2),60.593a(b)(2), 60.632(f), 60.5400(f) and 60.5406(b).* * * * *Subpart KKK--Standards of Performance for Equipment Leaksof VOC From Onshore Natural Gas Processing Plants for whichConstruction, Reconstruction, or Modification CommencedAfter January 20, 1984, and on or Before [INSERT DATE OFPUBLICATION IN THE FEDERAL REGISTER]. 3. The heading for Subpart KKK is revised as set outabove. 4. Section 60.630 is amended by revising paragraph (b)to read as follows:§60.630 Applicability and designation of affected facility.* * * * * (b) Any affected facility under paragraph (a) of thissection that commences construction, reconstruction, ormodification after January 20, 1984, and on or before[INSERT DATE OF PUBLICATON IN THE FEDERAL REGISTER], issubject to the requirements of this subpart.* * * * *Subpart LLL--Standards of Performance for SO2 Emissionsfrom Onshore Natural Gas Processing for which Construction,Reconstruction, or Modification Commenced After January 20,1984, and on or Before [INSERT DATE OF PUBLICATION IN THEFEDERAL REGISTER]. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

348. Page 348 of 604 5. The heading for Subpart LLL is revised as set outabove. 6. Section 60.640 is amended by revising paragraph (d)to read as follows:§60.640 Applicability and designation of affectedfacilities.* * * * * (d) The provisions of this subpart apply to eachaffected facility identified in paragraph (a) of thissection which commences construction or modification afterJanuary 20, 1984, and on or before [INSERT DATE OFPUBLICATON IN THE FEDERAL REGISTER].* * * * * 7. Part 60 is amended by adding subpart OOOO to readas follows:Subpart OOOO—Standards of Performance for Crude Oil andNatural Gas Production, Transmission, and Distribution.Sec.60.5360 What is the purpose of this subpart?60.5365 Am I subject to this subpart?60.5370 When must I comply with this subpart?60.5375 What standards apply to gas wellhead affected facilities?60.5380 What standards apply to centrifugal compressor affected facilities?60.5385 What standards apply to reciprocating compressor affected facilities? This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

349. Page 349 of 60460.5390 What standards apply to pneumatic controller affected facilities?60.5395 What standards apply to storage vessel affected facilities?60.5400 What VOC standards apply to affected facilities at an onshore natural gas processing plant?60.5401 What are the exceptions to the VOC standards for affected facilities at onshore natural gas processing plants?60.5402 What are the alternative emission limitations for affected facilities at onshore natural gas processing plants?60.5405 What standards apply to sweetening units at onshore natural gas processing plants?60.5406 What test methods and procedures must I use for my sweetening units affected facilities at onshore natural gas processing plants?60.5407 What are the requirements for monitoring of emissions and operations from my sweetening unit affected facilities at onshore natural gas processing plants?60.5408 What is an optional procedure for measuring H2S in acid gas-Tutwiler Procedure.60.5410 How do I demonstrate initial compliance with the standards for my gas wellhead affected facility, my centrifugal compressor affected facility, my reciprocating compressor affected facility, my pneumatic controller affected facility, my storage vessel affected facility, and my affected facilities at onshore natural gas processing plants?60.5415 How do I demonstrate continuous compliance with the standards for my gas wellhead affected facility, my centrifugal compressor affected facility, my reciprocating compressor affected facility, my pneumatic controller affected facility, my storage vessel affected facility, and my affected facilities at onshore natural gas processing plants?60.5420 What are my notification, reporting, and recordkeeping requirements?60.5421 What are my additional recordkeeping requirements for my affected facility subject to VOC requirements for onshore natural gas processing This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

350. Page 350 of 604 plants?60.5422 What are my additional reporting requirements for my affected facility subject to VOC requirements for onshore natural gas processing plants?60.5423 What additional recordkeeping and reporting requirements apply to my sweetening unit affected facilities at onshore natural gas processing plants?60.5425 What part of the General Provisions apply to me?60.5430 What definitions apply to this subpart?Subpart OOOO—Standards of Performance for Crude Oil andNatural Gas Production, Transmission, and Distribution.§60.5360 What is the purpose of this subpart? This subpart establishes emission standards andcompliance schedules for the control of volatile organiccompounds (VOC) and sulfur dioxide (SO2) emissions fromaffected facilities that commenced construction,modification or reconstruction after [INSERT DATE OFPUBLICATION IN THE FEDERAL REGISTER].§60.5365 Am I subject to this subpart? If you are the owner or operator of one or more of theaffected facilities listed in paragraphs (a) through (g) ofthis section that commenced construction, modification, orreconstruction after [INSERT DATE OF PUBLICATION IN THEFEDERAL REGISTER] your affected facility is subject to theapplicable provisions of this subpart. For the purposes ofthis subpart, a well completion operation following This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

351. Page 351 of 604hydraulic fracturing or refracturing that occurs at a gaswellhead facility that commenced construction,modification, or reconstruction on or before [INSERT DATEOF PUBLICATION IN THE FEDERAL REGISTER] is considered amodification of the gas wellhead facility, but does notaffect other equipment, process units, storage vessels, orpneumatic devices located at the well site. (a) A gas wellhead affected facility, is a singlenatural gas well. (b) A centrifugal compressor affected facility, whichis defined as a single centrifugal compressor locatedbetween the wellhead and the city gate (as defined in§60.5430), except that a centrifugal compressor located ata well site (as defined in §60.5430) is not an affectedfacility under this subpart. For the purposes of thissubpart, your centrifugal compressor is considered to havecommenced construction on the date the compressor isinstalled at the facility. (c) A reciprocating compressor affected facility,which is defined as a single reciprocating compressorlocated between the wellhead and the city gate (as definedin §60.5430), except that a reciprocating compressor This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

352. Page 352 of 604located at a well site (as defined in §60.5430) is not anaffected facility under this subpart. For the purposes ofthis subpart, your reciprocating compressor is consideredto have commenced construction on the date the compressoris installed at the facility. (d) A pneumatic controller affected facility, which isdefined as a single pneumatic controller. (e) A storage vessel affected facility, which isdefined as a single storage vessel. (f) Compressors and equipment (as defined in §60.5430)located at onshore natural gas processing plants. (1) Each compressor in VOC service or in wet gasservice is an affected facility. (2) The group of all equipment, except compressors,within a process unit is an affected facility. (3) Addition or replacement of equipment, as definedin §60.5430, for the purpose of process improvement that isaccomplished without a capital expenditure shall not byitself be considered a modification under this subpart. (4) Equipment (as defined in §60.5430) associated witha compressor station, dehydration unit, sweetening unit,underground storage tank, field gas gathering system, or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

353. Page 353 of 604liquefied natural gas unit is covered by §§60.5400,60.5401, 60.5402, 60.5421 and 60.5422 of this subpart if itis located at an onshore natural gas processing plant.Equipment (as defined in §60.5430) not located at theonshore natural gas processing plant site is exempt fromthe provisions of §§60.5400, 60.5401, 60.5402, 60.5421 and60.5422 of this subpart. (5) Affected facilities located at onshore natural gasprocessing plants and described in (f)(1) and (f)(2) areexempt from this subpart if they are subject to andcontrolled according to 40 CFR subparts VVa, GGG or GGGa. (g) Sweetening units located onshore that processnatural gas produced from either onshore or offshore wells. (1) Each sweetening unit that processes natural gas isan affected facility; and (2) Each sweetening unit that processes natural gasfollowed by a sulfur recovery unit is an affected facility. (3) Facilities that have a design capacity less than 2long tons per day (LT/D) of hydrogen sulfide (H2S) in theacid gas (expressed as sulfur) are required to comply withrecordkeeping and reporting requirements specified in§60.5423(c) but are not required to comply with §§60.5405 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

354. Page 354 of 604through 60.5407 and paragraphs 60.5410(g) and 60.5415(g) ofthis subpart. (4) Sweetening facilities producing acid gas that iscompletely reinjected into oil-or-gas-bearing geologicstrata or that is otherwise not released to the atmosphereare not subject to §§60.5405 through 60.5407, andparagraphs 60.5410(g), 60.5415(g), and §60.5423 of thissubpart.§60.5370 When must I comply with this subpart? (a) You must be in compliance with the standards ofthis subpart no later than the date of publication of thefinal rule in the Federal Register or upon startup,whichever is later. (b) The provisions for exemption from complianceduring periods of startup, shutdown, and malfunctionsprovided for in 40 CFR 60.8(c) do not apply to thissubpart. (c) You are exempt from the obligation to obtain apermit under 40 CFR part 70 or 40 CFR part 71, provided youare not otherwise required by law to obtain a permit under40 CFR 70.3(a) or 40 CFR 71.3(a). Notwithstanding theprevious sentence, you must continue to comply with the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

355. Page 355 of 604provisions of this subpart.§60.5375 What standards apply to gas wellhead affectedfacilities? If you are the owner or operator of a gas wellheadaffected facility, you must comply with paragraphs (a)through (g) of this section. (a) Except as provided in paragraph (f) of thissection, for each well completion operation with hydraulicfracturing, as defined in §60.5430, you must controlemissions by the operational procedures found in paragraphs(a)(1) through (a)(3) of this section. (1) You must minimize the emissions associated withventing of hydrocarbon fluids and gas over the duration offlowback by routing the recovered liquids into storagevessels and routing the recovered gas into a gas gatheringline or collection system. (2) You must employ sand traps, surge vessels,separators, and tanks during flowback and cleanoutoperations to safely maximize resource recovery andminimize releases to the environment. All salable qualitygas must be routed to the gas gathering line as soon aspracticable. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

356. Page 356 of 604 (3) You must capture and direct flowback emissionsthat cannot be directed to the gathering line to acompletion combustion device, except in conditions that mayresult in a fire hazard or explosion. Completion combustiondevices must be equipped with a reliable continuousignition source over the duration of flowback. (b) You must maintain a log for each well completionoperation at each gas wellhead affected facility. The logmust be completed on a daily basis and must contain therecords specified in §60.5420(c)(1)(iii). (c) You must demonstrate initial compliance with thestandards that apply to gas wellhead affected facilities asrequired by §60.5410. (d) You must demonstrate continuous compliance withthe standards that apply to gas wellhead affectedfacilities as required by §60.5415. (e) You must perform the required notification,recordkeeping, and reporting as required by §60.5420. (f) For wells meeting the criteria for wildcat ordelineation wells, each well completion operation withhydraulic fracturing at a gas wellhead affected facilitymust reduce emissions by using a completion combustion This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

357. Page 357 of 604device meeting the requirements of paragraph (a)(3) of thissection. You must also maintain records specified in§60.5420(c)(1)(iii) for wildcat or delineation wells.§60.5380 What standards apply to centrifugal compressoraffected facilities? You must comply with the standards in paragraphs (a)through (d) of this section, as applicable for eachcentrifugal compressor affected facility. (a) You must equip each rotating compressor shaft witha dry seal system upon initial startup. (b) You must demonstrate initial compliance with thestandards that apply to centrifugal compressor affectedfacilities as required by §60.5410. (c) You must demonstrate continuous compliance withthe standards that apply to centrifugal compressor affectedfacilities as required by §60.5415. (d) You must perform the required notification,recordkeeping, and reporting as required by §60.5420.§60.5385 What standards apply to reciprocating compressoraffected facilities? You must comply with the standards in paragraphs (a)through (d) of this section for each reciprocating This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

358. Page 358 of 604compressor affected facility. (a) You must replace the reciprocating compressor rodpacking before the compressor has operated for 26,000hours. The number of hours of operation must becontinuously monitored beginning upon initial startup ofyour reciprocating compressor affected facility, or thedate of publication of the final rule in the FederalRegister, or the date of the previous reciprocatingcompressor rod packing replacement, whichever is later. (b) You must demonstrate initial compliance withstandards that apply to reciprocating compressor affectedfacilities as required by §60.5410. (c) You must demonstrate continuous compliance withstandards that apply to reciprocating compressor affectedfacilities as required by §60.5415. (d) You must perform the required notification,recordkeeping, and reporting as required by §60.5420.§60.5390 What standards apply to pneumatic controlleraffected facilities? For each pneumatic controller affected facility youmust comply with the VOC standards, based on natural gas asa surrogate for VOC, in either paragraph (b) or (c) of this This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

359. Page 359 of 604section, as applicable. Pneumatic controllers meeting theconditions in paragraph (a) are exempt from thisrequirement. (a) The requirements of paragraph (b) or (c) of thissection are not required if you demonstrate, to theAdministrator’s satisfaction, that the use of a high bleeddevice is predicated. The demonstration may include, but isnot limited to, response time, safety and actuation. (b) Each pneumatic controller affected facilitylocated at a natural gas processing plant (as defined in§60.5430) must have zero emissions of natural gas. (c) Each pneumatic controller affected facility notlocated at a natural gas processing plant (as defined in§60.5430) must have natural gas emissions no greater than 6standard cubic feet per hour. (d) You must demonstrate initial compliance withstandards that apply to pneumatic controller affectedfacilities as required by §60.5410. (e) You must demonstrate continuous compliance withstandards that apply to pneumatic controller affectedfacilities as required by §60.5415. (f) You must perform the required notification, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

360. Page 360 of 604recordkeeping, and reporting as required by §60.5420,except that you are not required to submit thenotifications specified in §60.5420(a).§60.5395 What standards apply to storage vessel affectedfacilities? You must comply with the standards in paragraphs (a)through (e) of this section for each storage vesselaffected facility. (a) You must comply with the standards for storagevessels specified in 40 CFR part 63, subpart HH, §63.766(b)and (c), except as specified in paragraph (b) of thissection. Storage vessels that meet either one or both ofthe throughput conditions specified in paragraphs (a)(1) or(a)(2) of this section are not subject to the standards ofthis section. (1) The annual average condensate throughput is lessthan 1 barrel per day per storage vessel. (2) The annual average crude oil throughput is lessthan 20 barrels per day per storage vessel. (b) This standard does not apply to storage vesselsalready subject to and controlled in accordance with therequirements for storage vessels in 40 CFR part 63, subpart This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

361. Page 361 of 604HH, §63.766(b)(1) or (2). (c) You must demonstrate initial compliance withstandards that apply to storage vessel affected facilitiesas required by §60.5410. (d) You must demonstrate continuous compliance withstandards that apply to storage vessel affected facilitiesas required by §60.5415. (e) You must perform the required notification,recordkeeping, and reporting as required by §60.5420.§60.5400 What VOC standards apply to affected facilities atan onshore natural gas processing plant? This section applies to each compressor in VOC serviceor in wet gas service and the group of all equipment (asdefined in §60.5430), except compressors, within a processunit. (a) You must comply with the requirements of §60.482-1a(a), (b), and (d),§60.482-2a, and §60.482-4a through60.482-11a, except as provided in §60.5401. (b) You may elect to comply with the requirements of§§60.483-1a and 60.483-2a, as an alternative. (c) You may apply to the Administrator for permissionto use an alternative means of emission limitation that This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

362. Page 362 of 604achieves a reduction in emissions of VOC at leastequivalent to that achieved by the controls required inthis subpart according to the requirements of §60.5402 ofthis subpart. (d) You must comply with the provisions of §60.485a ofthis part except as provided in paragraph (f) of thissection. (e) You must comply with the provisions of §§60.486aand 60.487a of this part except as provided in §§60.5401,60.5421, and 60.5422 of this part. (f) You must use the following provision instead of§60.485a(d)(1): Each piece of equipment is presumed to bein VOC service or in wet gas service unless an owner oroperator demonstrates that the piece of equipment is not inVOC service or in wet gas service. For a piece of equipmentto be considered not in VOC service, it must be determinedthat the VOC content can be reasonably expected never toexceed 10.0 percent by weight. For a piece of equipment tobe considered in wet gas service, it must be determinedthat it contains or contacts the field gas before theextraction step in the process. For purposes of determiningthe percent VOC content of the process fluid that is This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

363. Page 363 of 604contained in or contacts a piece of equipment, proceduresthat conform to the methods described in ASTM E169-63, 77,or 93, E168-67, 77, or 92, or E260-73, 91, or 96(incorporated by reference as specified in §60.17) must beused.§60.5401 What are the exceptions to the VOC standards foraffected facilities at onshore natural gas processingplants? (a) You may comply with the following exceptions to theprovisions of subpart VVa of this part. (b)(1) Each pressure relief device in gas/vapor servicemay be monitored quarterly and within 5 days after eachpressure release to detect leaks by the methods specifiedin §60.485a(b) except as provided in §60.5400(c) and inparagraph (b)(4) of this section, and §60.482-4a(a) through(c) of subpart VVa. (2) If an instrument reading of 5000 ppm or greater ismeasured, a leak is detected. (3)(i) When a leak is detected, it must be repaired assoon as practicable, but no later than 15 calendar daysafter it is detected, except as provided in §60.482-9a. (ii) A first attempt at repair must be made no later This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

364. Page 364 of 604than 5 calendar days after each leak is detected. (4)(i) Any pressure relief device that is located in anonfractionating plant that is monitored only by non-plantpersonnel may be monitored after a pressure release thenext time the monitoring personnel are on-site, instead ofwithin 5 days as specified in paragraph (b)(1) of thissection and §60.482-4a(b)(1) of subpart VVa. (ii) No pressure relief device described in paragraph(b)(4)(i) of this section must be allowed to operate formore than 30 days after a pressure release withoutmonitoring. (c) Sampling connection systems are exempt from therequirements of §60.482-5a. (d) Pumps in light liquid service, valves in gas/vaporand light liquid service, and pressure relief devices ingas/vapor service that are located at a nonfractionatingplant with a design capacity to process 283,200 standardcubic meters per day (scmd) (10 million standard cubic feetper day) or more of field gas are exempt from the routinemonitoring requirements of §§60.482-2a(a)(1) and 60.482-7a(a), and paragraph (b)(1) of this section. (e) Pumps in light liquid service, valves in gas/vapor This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

365. Page 365 of 604and light liquid service, and pressure relief devices ingas/vapor service within a process unit that is located inthe Alaskan North Slope are exempt from the routinemonitoring requirements of §§60.482-2a(a)(1), 60.482-7a(a),and paragraph (b)(1) of this section. (f) Flares used to comply with this subpart must complywith the requirements of §60.18. (g) An owner or operator may use the followingprovisions instead of §60.485a(e): (1) Equipment is in heavy liquid service if the weightpercent evaporated is 10 percent or less at 150oC (302oF) asdetermined by ASTM Method D86-78, 82, 90, 95, or 96(incorporated by reference as specified in §60.17). (2) Equipment is in light liquid service if the weightpercent evaporated is greater than 10 percent at 150oC(302oF) as determined by ASTM Method D86-78, 82, 90, 95, or96 (incorporated by reference as specified in §60.17).§60.5402 What are the alternative emission limitations forequipment leaks from onshore natural gas processing plants? (a) If, in the Administrators judgment, an alternativemeans of emission limitation will achieve a reduction inVOC emissions at least equivalent to the reduction in VOC This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

366. Page 366 of 604emissions achieved under any design, equipment, workpractice or operational standard, the Administrator willpublish, in the Federal Register, a notice permitting theuse of that alternative means for the purpose of compliancewith that standard. The notice may condition permission onrequirements related to the operation and maintenance ofthe alternative means. (b) Any notice under paragraph (a) of this section mustbe published only after notice and an opportunity for apublic hearing. (c) The Administrator will consider applications underthis section from either owners or operators of affectedfacilities, or manufacturers of control equipment. (d) The Administrator will treat applications underthis section according to the following criteria, except incases where the Administrator concludes that other criteriaare appropriate: (1) The applicant must collect, verify and submit testdata, covering a period of at least 12 months, necessary tosupport the finding in paragraph (a) of this section. (2) If the applicant is an owner or operator of anaffected facility, the applicant must commit in writing to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

367. Page 367 of 604operate and maintain the alternative means so as to achievea reduction in VOC emissions at least equivalent to thereduction in VOC emissions achieved under the design,equipment, work practice or operational standard.§60.5405 What standards apply to sweetening units atonshore natural gas processing plants? (a) During the initial performance test required by§60.8(b), you must achieve at a minimum, an SO2 emissionreduction efficiency (Zi) to be determined from Table 1 ofthis subpart based on the sulfur feed rate (X) and thesulfur content of the acid gas (Y) of the affectedfacility. (b) After demonstrating compliance with the provisionsof paragraph (a) of this section, you must achieve at aminimum, an SO2 emission reduction efficiency (Zc) to bedetermined from Table 2 of this subpart based on the sulfurfeed rate (X) and the sulfur content of the acid gas (Y) ofthe affected facility.60.5406 What test methods and procedures must I use for mysweetening units affected facilities at onshore natural gasprocessing plants? (a) In conducting the performance tests required in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

368. Page 368 of 604§60.8, you must use the test methods in appendix A of thispart or other methods and procedures as specified in thissection, except as provided in paragraph §60.8(b). (b) During a performance test required by §60.8, youmust determine the minimum required reduction efficiencies(Z) of SO2 emissions as required in §60.5405(a) and (b) asfollows: (1) The average sulfur feed rate (X) must be computedas follows:Where:X = average sulfur feed rate, Mg/D (LT/D).Qa = average volumetric flow rate of acid gas fromsweetening unit, dscm/day (dscf/day).Y = average H2S concentration in acid gas feed fromsweetening unit, percent by volume, expressed as a decimal.K = (32 kg S/kg-mole)/((24.04 dscm/kg-mole)(1000 kg S/ Mg)) = 1.331 × 10−3Mg/dscm, for metric units = (32 lb S/lb-mole)/((385.36 dscf/lb-mole)(2240 lb S/longton)) = 3.707 × 10−5long ton/dscf, for English units. (2) You must use the continuous readings from theprocess flowmeter to determine the average volumetric flowrate (Qa) in dscm/day (dscf/day) of the acid gas from thesweetening unit for each run. (3) You must use the Tutwiler procedure in §60.5408 ora chromatographic procedure following ASTM E–260 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

369. Page 369 of 604(incorporated by reference—see §60.17) to determine the H2Sconcentration in the acid gas feed from the sweetening unit(Y). At least one sample per hour (at equally spacedintervals) must be taken during each 4-hour run. Thearithmetic mean of all samples must be the average H2Sconcentration (Y) on a dry basis for the run. Bymultiplying the result from the Tutwiler procedure by 1.62× 10−3, the units gr/100 scf are converted to volumepercent. (4) Using the information from paragraphs (b)(1) and(b)(3) of this section, Tables 1 and 2 of this subpart mustbe used to determine the required initial (Zi) andcontinuous (Zc) reduction efficiencies of SO2 emissions. (c) You must determine compliance with the SO2standards in §60.5405(a) or (b) as follows: (1) You must compute the emission reduction efficiency(R) achieved by the sulfur recovery technology for each runusing the following equation: 100 / (2) You must use the level indicators or manualsoundings to measure the liquid sulfur accumulation rate inthe product storage tanks. You must use readings taken at This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

370. Page 370 of 604the beginning and end of each run, the tank geometry,sulfur density at the storage temperature, and sampleduration to determine the sulfur production rate (S) inkg/hr (lb/hr) for each run. (3) You must compute the emission rate of sulfur foreach run as follows: /Where:E = emission rate of sulfur per run, kg/hr.Ce = concentration of sulfur equivalent (SO2+ reducedsulfur), g/dscm (lb/dscf).Qsd = volumetric flow rate of effluent gas, dscm/hr(dscf/hr).K1 = conversion factor, 1000 g/kg (7000 gr/lb). (4) The concentration (Ce) of sulfur equivalent mustbe the sum of the SO2 and TRS concentrations, after beingconverted to sulfur equivalents. For each run and each ofthe test methods specified in this paragraph (c) of thissection, you must use a sampling time of at least 4 hours.You must use Method 1 of appendix A to part 60 of thischapter to select the sampling site. The sampling point inthe duct must be at the centroid of the cross-section ifthe area is less than 5 m2 (54 ft2 ) or at a point nocloser to the walls than 1 m (39 in) if the cross-sectional This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

371. Page 371 of 604area is 5 m2 or more, and the centroid is more than 1 m (39in.) from the wall. (i) You must use Method 6 of appendix A to part 60 ofthis chapter to determine the SO2 concentration. You musttake eight samples of 20 minutes each at 30-minuteintervals. The arithmetic average must be the concentrationfor the run. The concentration must be multiplied by 0.5 ×10−3 to convert the results to sulfur equivalent. (ii) You must use Method 15 of appendix A to part 60of this chapter to determine the TRS concentration fromreduction-type devices or where the oxygen content of theeffluent gas is less than 1.0 percent by volume. Thesampling rate must be at least 3 liters/min (0.1 ft3 /min)to insure minimum residence time in the sample line. Youmust take sixteen samples at 15-minute intervals. Thearithmetic average of all the samples must be theconcentration for the run. The concentration in ppm reducedsulfur as sulfur must be multiplied by 1.333 × 10−3 toconvert the results to sulfur equivalent. (iii) You must use Method 16A or Method 15 of appendixA to part 60 of this chapter to determine the reducedsulfur concentration from oxidation-type devices or where This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

372. Page 372 of 604the oxygen content of the effluent gas is greater than 1.0percent by volume. You must take eight samples of 20minutes each at 30-minute intervals. The arithmetic averagemust be the concentration for the run. The concentration inppm reduced sulfur as sulfur must be multiplied by 1.333 ×10−3 to convert the results to sulfur equivalent. (iv) You must use Method 2 of appendix A to part 60 ofthis chapter to determine the volumetric flow rate of theeffluent gas. A velocity traverse must be conducted at thebeginning and end of each run. The arithmetic average ofthe two measurements must be used to calculate thevolumetric flow rate (Qsd) for the run. For thedetermination of the effluent gas molecular weight, asingle integrated sample over the 4-hour period may betaken and analyzed or grab samples at 1-hour intervals maybe taken, analyzed, and averaged. For the moisture content,you must take two samples of at least 0.10 dscm (3.5 dscf)and 10 minutes at the beginning of the 4-hour run and nearthe end of the time period. The arithmetic average of thetwo runs must be the moisture content for the run.60.5407 What are the requirements for monitoring ofemissions and operations from my sweetening unit affected This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

373. Page 373 of 604facilities at onshore natural gas processing plants? (a) If your sweetening unit affected facility islocated at an onshore natural gas processing plant and issubject to the provisions of §60.5405(a) or (b) you mustinstall, calibrate, maintain, and operate monitoringdevices or perform measurements to determine the followingoperations information on a daily basis: (1) The accumulation of sulfur product over each 24-hour period. The monitoring method may incorporate the useof an instrument to measure and record the liquid sulfurproduction rate, or may be a procedure for measuring andrecording the sulfur liquid levels in the storage tankswith a level indicator or by manual soundings, withsubsequent calculation of the sulfur production rate basedon the tank geometry, stored sulfur density, and elapsedtime between readings. The method must be designed to beaccurate within ±2 percent of the 24-hour sulfuraccumulation. (2) The H2S concentration in the acid gas from thesweetening unit for each 24-hour period. At least onesample per 24-hour period must be collected and analyzedusing the equation specified in §60.5406(b)(1). The This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

374. Page 374 of 604Administrator may require you to demonstrate that the H2Sconcentration obtained from one or more samples over a 24-hour period is within ±20 percent of the average of 12samples collected at equally spaced intervals during the24-hour period. In instances where the H2S concentration ofa single sample is not within ±20 percent of the average ofthe 12 equally spaced samples, the Administrator mayrequire a more frequent sampling schedule. (3) The average acid gas flow rate from the sweeteningunit. You must install and operate a monitoring device tocontinuously measure the flow rate of acid gas. Themonitoring device reading must be recorded at least onceper hour during each 24-hour period. The average acid gasflow rate must be computed from the individual readings. (4) The sulfur feed rate (X). For each 24-hour period,you must compute X using the equation specified in§60.5406(b)(3). (5) The required sulfur dioxide emission reductionefficiency for the 24-hour period. You must use the sulfurfeed rate and the H2S concentration in the acid gas for the24-hour period, as applicable, to determine the requiredreduction efficiency in accordance with the provisions of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

375. Page 375 of 604§60.5405(b). (b) Where compliance is achieved through the use of anoxidation control system or a reduction control systemfollowed by a continually operated incineration device, youmust install, calibrate, maintain, and operate monitoringdevices and continuous emission monitors as follows: (1) A continuous monitoring system to measure thetotal sulfur emission rate (E) of SO2 in the gasesdischarged to the atmosphere. The SO2 emission rate must beexpressed in terms of equivalent sulfur mass flow rates(kg/hr (lb/hr)). The span of this monitoring system must beset so that the equivalent emission limit of §60.5405(b)will be between 30 percent and 70 percent of themeasurement range of the instrument system. (2) Except as provided in paragraph (b)(3) of thissection: A monitoring device to measure the temperature ofthe gas leaving the combustion zone of the incinerator, ifcompliance with §60.5405(a) is achieved through the use ofan oxidation control system or a reduction control systemfollowed by a continually operated incineration device. Themonitoring device must be certified by the manufacturer tobe accurate to within ±1 percent of the temperature being This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

376. Page 376 of 604measured. (3) When performance tests are conducted under theprovision of §60.8 to demonstrate compliance with thestandards under §60.5405, the temperature of the gasleaving the incinerator combustion zone must be determinedusing the monitoring device. If the volumetric ratio ofsulfur dioxide to sulfur dioxide plus total reduced sulfur(expressed as SO2) in the gas leaving the incinerator isequal to or less than 0.98, then temperature monitoring maybe used to demonstrate that sulfur dioxide emissionmonitoring is sufficient to determine total sulfuremissions. At all times during the operation of thefacility, you must maintain the average temperature of thegas leaving the combustion zone of the incinerator at orabove the appropriate level determined during the mostrecent performance test to ensure the sulfur compoundoxidation criteria are met. Operation at lower averagetemperatures may be considered by the Administrator to beunacceptable operation and maintenance of the affectedfacility. You may request that the minimum incineratortemperature be reestablished by conducting new performancetests under §60.8. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

377. Page 377 of 604 (3) Upon promulgation of a performance specificationof continuous monitoring systems for total reduced sulfurcompounds at sulfur recovery plants, you may, as analternative to paragraph (b)(2) of this section, install,calibrate, maintain, and operate a continuous emissionmonitoring system for total reduced sulfur compounds asrequired in paragraph (d) of this section in addition to asulfur dioxide emission monitoring system. The sum of theequivalent sulfur mass emission rates from the twomonitoring systems must be used to compute the total sulfuremission rate (E). (c) Where compliance is achieved through the use of areduction control system not followed by a continuallyoperated incineration device, you must install, calibrate,maintain, and operate a continuous monitoring system tomeasure the emission rate of reduced sulfur compounds asSO2 equivalent in the gases discharged to the atmosphere.The SO2 equivalent compound emission rate must be expressedin terms of equivalent sulfur mass flow rates (kg/hr(lb/hr)). The span of this monitoring system must be set sothat the equivalent emission limit of §60.5405(b) will bebetween 30 and 70 percent of the measurement range of the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

378. Page 378 of 604system. This requirement becomes effective uponpromulgation of a performance specification for continuousmonitoring systems for total reduced sulfur compounds atsulfur recovery plants. (d) For those sources required to comply withparagraph (b) or (c) of this section, you must calculatethe average sulfur emission reduction efficiency achieved(R) for each 24-hour clock internal. The 24-hour intervalmay begin and end at any selected clock time, but must beconsistent. You must compute the 24-hour average reductionefficiency (R) based on the 24-hour average sulfurproduction rate (S) and sulfur emission rate (E), using theequation in §60.5406(c)(1). (1) You must use data obtained from the sulfurproduction rate monitoring device specified in paragraph(a) of this section to determine S. (2) You must use data obtained from the sulfuremission rate monitoring systems specified in paragraphs(b) or (c) of this section to calculate a 24-hour averagefor the sulfur emission rate (E). The monitoring systemmust provide at least one data point in each successive 15-minute interval. You must use at least two data points to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

379. Page 379 of 604calculate each 1-hour average. You must use a minimum of 181-hour averages to compute each 24-hour average. (e) In lieu of complying with (b) or (c) of thissection, those sources with a design capacity of less than152 Mg/D (150 LT/D) of H2S expressed as sulfur may calculatethe sulfur emission reduction efficiency achieved for each24-hour period by:Where:R = The sulfur dioxide removal efficiency achieved duringthe 24-hour period, percent.K2 = Conversion factor, 0.02400 Mg/D per kg/hr (0.01071LT/D per lb/hr).S = The sulfur production rate during the 24-hour period,kg/hr (lb/hr).X = The sulfur feed rate in the acid gas, Mg/D (LT/D). (f) The monitoring devices required in paragraphs(b)(1), (b)(3) and (c) of this section must be calibratedat least annually according to the manufacturersspecifications, as required by §60.13(b). (g) The continuous emission monitoring systemsrequired in paragraphs (b)(1), (b)(3), and (c) of thissection must be subject to the emission monitoringrequirements of §60.13 of the General Provisions. Forconducting the continuous emission monitoring system This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

380. Page 380 of 604performance evaluation required by §60.13(c), PerformanceSpecification 2 of appendix B to part 60 of this chaptermust apply, and Method 6 must be used for systems requiredby paragraph (b) of this section.§60.5408 What is an optional procedure for measuringhydrogen sulfide in acid gas—Tutwiler Procedure.52 (a) When an instantaneous sample is desired and H2Sconcentration is ten grains per 1000 cubic foot or more, a100 ml Tutwiler burette is used. For concentrations lessthan ten grains, a 500 ml Tutwiler burette and more dilutesolutions are used. In principle, this method consists oftitrating hydrogen sulfide in a gas sample directly with astandard solution of iodine. (b) Apparatus. (See Figure 1 of this subpart) A 100or 500 ml capacity Tutwiler burette, with two-way glassstopcock at bottom and three-way stopcock at top whichconnect either with inlet tubulature or glass-stopperedcylinder, 10 ml capacity, graduated in 0.1 ml subdivision;rubber tubing connecting burette with leveling bottle. (c) Reagents. (1) Iodine stock solution, 0.1N. Weight55 Gas Engineers Handbook, Fuel Gas Engineering practices, TheIndustrial Press, 93 Worth Street, New York, NY, 1966, FirstEdition, Second Printing, page 6/25 (Docket A-80-20-A, Entry II-I-67). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

381. Page 381 of 60412.7 g iodine, and 20 to 25 g cp potassium iodide for eachliter of solution. Dissolve KI in as little water asnecessary; dissolve iodine in concentrated KI solution,make up to proper volume, and store in glass-stopperedbrown glass bottle. (2) Standard iodine solution, 1 ml=0.001771 g I.Transfer 33.7 ml of above 0.1N stock solution into a 250 mlvolumetric flask; add water to mark and mix well. Then, for100 ml sample of gas, 1 ml of standard iodine solution isequivalent to 100 grains H2S per cubic feet of gas. (3) Starch solution. Rub into a thin paste about oneteaspoonful of wheat starch with a little water; pour intoabout a pint of boiling water; stir; let cool and decantoff clear solution. Make fresh solution every few days. (d) Procedure. Fill leveling bulb with starchsolution. Raise (L), open cock (G), open (F) to (A), andclose (F) when solutions starts to run out of gas inlet.Close (G). Purge gas sampling line and connect with (A).Lower (L) and open (F) and (G). When liquid level isseveral ml past the 100 ml mark, close (G) and (F), anddisconnect sampling tube. Open (G) and bring starchsolution to 100 ml mark by raising (L); then close (G). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

382. Page 382 of 604Open (F) momentarily, to bring gas in burette toatmospheric pressure, and close (F). Open (G), bring liquidlevel down to 10 ml mark by lowering (L). Close (G), clamprubber tubing near (E) and disconnect it from burette.Rinse graduated cylinder with a standard iodine solution(0.00171 g I per ml); fill cylinder and record reading.Introduce successive small amounts of iodine thru (F);shake well after each addition; continue until a faintpermanent blue color is obtained. Record reading; subtractfrom previous reading, and call difference D. (e) With every fresh stock of starch solution performa blank test as follows: Introduce fresh starch solutioninto burette up to 100 ml mark. Close (F) and (G). Lower(L) and open (G). When liquid level reaches the 10 ml mark,close (G). With air in burette, titrate as during a testand up to same end point. Call ml of iodine used C. Then,Grains H2S per 100 cubic foot of gas=100 (D—C) (f) Greater sensitivity can be attained if a 500 mlcapacity Tutwiler burette is used with a more dilute(0.001N) iodine solution. Concentrations less than 1.0grains per 100 cubic foot can be determined in this way.Usually, the starch-iodine end point is much less distinct, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

383. Page 383 of 604and a blank determination of end point, with H2S-free gasor air, is required.Figure 1. Tutwiler burette (lettered items mentioned intext).§60.5410 How do I demonstrate initial compliance with the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

384. Page 384 of 604standards for my gas wellhead affected facility, mycentrifugal compressor affected facility, my reciprocatingcompressor affected facility, my pneumatic controlleraffected facility, my storage vessel affected facility, andmy affected facilities at onshore natural gas processingplants? You must determine initial compliance with thestandards for each affected facility using the requirementsin paragraphs (a) through (g) of this section. The initialcompliance period begins on the date of publication of thefinal rule in the Federal Register or upon initial startup,whichever is later, and ends on the date the first annualreport is due as specified in §60.5420(b). (a) You have achieved initial compliance withstandards for each well completion operation conducted atyour gas wellhead affected facility if you have compliedwith paragraphs (a)(1) and (a)(2) of this section. (1) You have notified the Administrator within 30 daysof the commencement of the well completion operation, thedate of the commencement of the well completion operation,the latitude and longitude coordinates of the well indecimal degrees to an accuracy and precision of five (5) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

385. Page 385 of 604decimals of a degree using the North American Datum (NAD)of 1983. (2) You have maintained a log of records as specifiedin §60.5375(b) or (f) for each well completion operationconducted during the initial compliance period. (3) You have submitted the initial annual report foryour wellhead affected facility as required in §60.5420(b). (b) You have achieved initial compliance withstandards for your centrifugal compressor affected facilityif the centrifugal compressor is fitted with a dry sealsystem upon initial startup as required by §60.5380. (c) You have achieved initial compliance withstandards for each reciprocating compressor affectedfacility if you have complied with paragraphs (c)(1) and(c)(2) of this section. (1) During the initial compliance period, you havecontinuously monitored the number of hours of operation. (2) You have included the cumulative number of hoursof operation for your reciprocating compressor affectedfacility during the initial compliance period in yourinitial annual report required in §60.5420(b). (d) You have achieved initial compliance with emission This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

386. Page 386 of 604standards for your pneumatic controller affected facilityif you comply with the requirements specified in paragraphs(d)(1) through (d)(4) of this section. (1) You have demonstrated, to the Administrator’ssatisfaction, the use of a high bleed device is predicatedas specified in §60.5490(a). (2) You own or operate a pneumatic controller affectedfacility located at a natural gas processing plant and yourpneumatic controller is driven other than by use of naturalgas and therefore emits zero natural gas. (3) You own or operate a pneumatic controller affectedfacility not located at a natural gas processing plant andthe manufacturer’s design specifications guarantee thecontroller emits less than or equal to 6.0 standard cubicfeet of gas per hour. (4) You have included the information in paragraphs(d)(1) through (d)(3) of this section in the initial annualreport submitted for your pneumatic controller affectedfacilities according to the requirements of §60.5420(b). (e) You have demonstrated initial compliance withemission standards for your storage vessel affectedfacility if you are complying with paragraphs (e)(1) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

387. Page 387 of 604through (e)(7) of this section. (1) You have equipped the storage vessel with a closedvent system that meets the requirements of §63.771(c)connected to a control device that meets the conditionsspecified in §63.771(d). (2) You have conducted an initial performance test asrequired in §63.772(e) within 180 days after initialstartup or the date of publication of the final rule in theFederal Register and have conducted the compliancedemonstration in §63.772(f). (3) You have conducted the initial inspectionsrequired in §63.773(c). (4) You have installed and operated continuousparameter monitoring systems in accordance with §63.773(d). (5) If you are exempt from the standards of §60.5395according to §60.5395(a)(1) or (a)(2), you have determinedthe condensate or crude oil throughput, as applicable,according to paragraphs (e)(5)(i) or (e)(5)(ii) of thissection and demonstrated to the Administrator’ssatisfaction that your annual average condensate throughputis less than 1 barrel per day per tank and your annualaverage crude oil throughput is less than 20 barrels per This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

388. Page 388 of 604day per tank. (i) You have installed and operated a flow meter tomeasure condensate or crude oil throughput in accordancewith the manufacturer’s procedures or specifications. (ii) You have used any other method approved by theAdministrator to determine annual average condensate orcrude oil throughput. (6) You have submitted the information in paragraphs(e)(1) through (e)(5) of this section in the initial annualreport for your storage vessel affected facility asrequired in §60.5420(b). (f) For affected facilities at onshore natural gasprocessing plants, initial compliance with the VOCrequirements is demonstrated if you are in compliance withthe requirements of §60.5400. (g) For sweetening unit affected facilities at onshorenatural gas processing plants, initial compliance isdemonstrated according to paragraphs (g)(1) through (g)(3)of this section. (1) To determine compliance with the standards for SO2specified in §60.5405(a), during the initial performancetest as required by §60.8, the minimum required sulfur This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

389. Page 389 of 604dioxide emission reduction efficiency (Zi) is compared tothe emission reduction efficiency (R) achieved by thesulfur recovery technology as specified in paragraphs(g)(1)(i) and (g)(1)(ii) of this section. (i) If R ≥ Zi, your affected facility is in compliance. (ii) If R < Zi, your affected facility is not incompliance. (2) The emission reduction efficiency (R) achieved bythe sulfur reduction technology must be determined usingthe procedures in §60.5406(c)(1). (3) You have submitted the results of paragraphs(g)(1) and (g)(2) in the initial annual report submittedfor your sweetening unit affected facilities at onshorenatural gas processing plants.§60.5415 How do I demonstrate continuous compliance withthe standards for my gas wellhead affected facility, mycentrifugal compressor affected facility, my stationaryreciprocating compressor affected facility, my pneumaticcontroller affected facility, my storage vessel affectedfacility, and my affected facilities at onshore natural gasprocessing plants? (a) For each gas wellhead affected facility, you must This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

390. Page 390 of 604demonstrate continuous compliance by maintaining therecords for each completion operation (as defined in§60.5430) specified in §60.5420. (b) For each centrifugal compressor affectedfacility, continuous compliance is demonstrated if therotating compressor shaft is equipped with a dry seal. (c) For each reciprocating compressor affectedfacility, you have demonstrated continuous complianceaccording to paragraphs (c)(1) and (2) of this section (1) You have continuously monitored the number ofhours of operation for each reciprocating compressoraffected facility since initial startup, or the date ofpublication of the final rule in the Federal Register, orthe date of the previous reciprocating compressor rodpacking replacement, whichever is later. The cumulativenumber of hours of operation must be included in the annualreport as required in §60.5420(b)(4). (2) You have replaced the reciprocating compressor rodpacking before the total number of hours of operationreaches 26,000 hours. (d) For each pneumatic controller affected facility,continuous compliance is demonstrated by maintaining the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

391. Page 391 of 604records demonstrating that you have installed and operatedthe pneumatic controllers as required in §60.5390(a), (b)or (c). (e) For each storage vessel affected facility,continuous compliance is demonstrated according 40 CFR part63, subpart HH, §63.772(f). (f) For affected facilities at onshore natural gasprocessing plants, continuous compliance with VOCrequirements is demonstrated if you are in compliance withthe requirements of §60.5400. (g) For each sweetening unit affected facility atonshore natural gas processing plants, you must demonstratecontinuous compliance with the standards for SO2 specifiedin §60.5405(b) according to paragraphs (g)(1) and (g)(2) ofthis section. (1) The minimum required SO2 emission reductionefficiency (Zc) is compared to the emission reductionefficiency (R) achieved by the sulfur recovery technology. (i) If R ≥ Zc, your affected facility is in compliance. (ii) If R < Zc, your affected facility is not incompliance. (2) The emission reduction efficiency (R) achieved by This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

392. Page 392 of 604the sulfur reduction technology must be determined usingthe procedures in §60.5406(c)(1). (h) Affirmative Defense for Exceedance of EmissionLimit During Malfunction. In response to an action toenforce the standards set forth in paragraph §§60.5375,60.5380, 60.5385, 60.5390, 60.5395, 60.5400, and 60.5405,you may assert an affirmative defense to a claim for civilpenalties for exceedances of such standards that are causedby malfunction, as defined at 40 CFR 60.2. Appropriatepenalties may be assessed, however, if you fail to meetyour burden of proving all of the requirements in theaffirmative defense. The affirmative defense shall not beavailable for claims for injunctive relief. (1) To establish the affirmative defense in any actionto enforce such a limit, you must timely meet thenotification requirements in §60.5420(a), and must prove bya preponderance of evidence that: (i) The excess emissions: (A) Were caused by a sudden, infrequent, andunavoidable failure of air pollution control and monitoringequipment, process equipment, or a process to operate in anormal or usual manner, and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

393. Page 393 of 604 (B) Could not have been prevented through carefulplanning, proper design or better operation and maintenancepractices; and (C) Did not stem from any activity or event that couldhave been foreseen and avoided, or planned for; and (D) Were not part of a recurring pattern indicative ofinadequate design, operation, or maintenance; and (ii) Repairs were made as expeditiously as possiblewhen the applicable emission limitations were beingexceeded. Off-shift and overtime labor were used, to theextent practicable to make these repairs; and (iii) The frequency, amount and duration of the excessemissions (including any bypass) were minimized to themaximum extent practicable during periods of suchemissions; and (iv) If the excess emissions resulted from a bypass ofcontrol equipment or a process, then the bypass wasunavoidable to prevent loss of life, personal injury, orsevere property damage; and (v) All possible steps were taken to minimize theimpact of the excess emissions on ambient air quality, theenvironment and human health; and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

394. Page 394 of 604 (vi) All emissions monitoring and control systems werekept in operation if at all possible, consistent withsafety and good air pollution control practices; and (vii) All of the actions in response to the excessemissions were documented by properly signed,contemporaneous operating logs; and (viii) At all times, the facility was operated in amanner consistent with good practices for minimizingemissions; and (ix) A written root cause analysis has been prepared,the purpose of which is to determine, correct, andeliminate the primary causes of the malfunction and theexcess emissions resulting from the malfunction event atissue. The analysis shall also specify, using bestmonitoring methods and engineering judgment, the amount ofexcess emissions that were the result of the malfunction. (2) The owner or operator of the facility experiencingan exceedance of its emission limit(s) during a malfunctionshall notify the Administrator by telephone or facsimile(FAX) transmission as soon as possible, but no later than 2business days after the initial occurrence of themalfunction, if it wishes to avail itself of an affirmative This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

395. Page 395 of 604defense to civil penalties for that malfunction. The owneror operator seeking to assert an affirmative defense shallalso submit a written report to the Administrator within 45days of the initial occurrence of the exceedance of thestandards in §§60.5375, 60.5380, 60.5385, 60.5390, 60.5395,and 60.5400 to demonstrate, with all necessary supportingdocumentation, that it has met the requirements set forthin paragraph (a) of this section. The owner or operator mayseek an extension of this deadline for up to 30 additionaldays by submitting a written request to the Administratorbefore the expiration of the 45-day period. Until a requestfor an extension has been approved by the Administrator,the owner or operator is subject to the requirement tosubmit such report within 45 days of the initial occurrenceof the exceedance.§60.5420 What are my notification, reporting, andrecordkeeping requirements? (a) You must submit the notifications required in§60.7(a)(1), (a)(3) and (a)(4), and according to paragraphs(a)(1) and (a)(2) of this section, if you own or operateone or more of the affected facilities specified in§60.5365. For the purposes of this subpart, a workover that This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

396. Page 396 of 604occurs after [INSERT DATE OF PUBLICATION IN THE FEDERALREGISTER] at each affected facility for which construction,reconstruction, or modification commenced on or before[INSERT DATE OF PUBLICATION IN THE FEDERAL REGISTER] isconsidered a modification for which a notification must besubmitted under §60.7(a)(4). (1) If you own or operate a pneumatic controlleraffected facility you are not required to submit thenotifications required in §60.7(a)(1), (a)(3) and (a)(4). (2) If you own or operate a gas wellhead affectedfacility, you must submit a notification to theAdministrator within 30 days of the commencement of thewell completion operation. The notification must includethe date of commencement of the well completion operation,the latitude and longitude coordinates of the well indecimal degrees to an accuracy and precision of five (5)decimals of a degree using the North American Datum of1983. (b) Reporting requirements. You must submit annualreports containing the information specified in paragraphs(b)(1) through (b)(6) of this section to the Administrator.The initial annual report is due 1 year after the initial This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

397. Page 397 of 604startup date for your affected facility or 1 year after thedate of publication of the final rule in the FederalRegister, whichever is later. Subsequent annual reports aredue on the same date each year as the initial annualreport. If you own or operate more than one affectedfacility, you may submit one report for multiple affectedfacilities provided the report contains all of theinformation required as specified in paragraphs (b)(1)through (b)(6) of this section. (1) The general information specified in paragraphs(b)(1)(i) through (b)(1)(iii) of this section. (i) The company name and address of the affectedfacility. (ii) An identification of each affected facility beingincluded in the annual report. (iii) Beginning and ending dates of the reportingperiod. (2) For each gas wellhead affected facility, theinformation in paragraphs (b)(2)(i) through (b)(2)(iii) ofthis section. (i) An identification of each well completionoperation, as defined in §60.5430, for each gas wellhead This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

398. Page 398 of 604affected facility conducted during the reporting period; (ii) A record of deviations in cases where wellcompletion operations with hydraulic fracturing were notperformed in compliance with the requirements specified in§60.5375 for each gas well affected facility. (iii) Records specified in §60.5375(b) for each wellcompletion operation that occurred during the reportingperiod. (3) For each centrifugal compressor affected facilityinstalled during the reporting period, documentation thatthe centrifugal compressor is equipped with dry seals. (4) For each reciprocating compressor affectedfacility, the information specified in paragraphs (b)(4)(i)and (b)(4)(ii) of this section. (i) The cumulative number of hours or operation sinceinitial startup, the date of publication of the final rulein the Federal Register, or since the previousreciprocating compressor rod packing replacement, whicheveris later. (ii) Documentation that the reciprocating compressorrod packing was replaced before the cumulative number ofhours of operation reached 24,000 hours. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

399. Page 399 of 604 (5) For each pneumatic controller affected facility,the information specified in paragraphs (b)(5)(i) through(b)(5)(iv) of this section. (i) The date, location and manufacturer specificationsfor each pneumatic controller installed. (ii) If applicable, documentation that the use of highbleed pneumatic devices is predicated and the reasons why. (iii) For pneumatic controllers not installed at anatural gas processing plant, the manufacturer’s guaranteethat the device is designed such that natural gas emissionsare less than 6 standard cubic feet per hour. (iv) For pneumatic controllers installed at a naturalgas processing plant, documentation that each controllershas zero natural gas emissions. (6) For each storage vessel affected facility, theinformation in paragraphs (b)(6)(i) and (b)(6)(ii) of thissection. (i) If required to reduce emissions by complying withparagraph §60.5395(a)(1), the records specified in 40 CFRpart 63, subpart §63.774(b)(2) through (b)(8). (ii) Documentation that the annual average condensatethroughput is less than 1 barrel per day per storage vessel This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

400. Page 400 of 604and crude oil throughput is less than 21 barrels per dayper storage for meeting the requirements in §60.5395(a)(1)or (a)(2). (c) Recordkeeping requirements. You must maintain therecords identified as specified in §60.7(f) and inparagraphs (c)(1) through (c)(5) of this section (1) The records for each gas wellhead affectedfacility as specified in paragraphs (c)(1)(i) through(c)(1)(iii). (i) Records identifying each well completion operationfor each gas wellhead affected facility conducted duringthe reporting period; (ii) Record of deviations in cases where wellcompletion operations with hydraulic fracturing were notperformed in compliance with the requirements specified in§60.5375. (iii) Records required in §60.5375(b) or (f) for eachwell completion operation conducted for each gas wellheadaffected facility that occurred during the reportingperiod. You must maintain the records specified inparagraphs (c)(1)(iii)(A) and (c)(1)(iii)(B) of thissection. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

401. Page 401 of 604 (A) For each gas wellheads affected facility requiredto comply with the requirements of §60.5375(a), you mustrecord: The location of the well; the duration of flowback;duration of recovery to the sales line; duration ofcombustion; duration of venting; and specific reasons forventing in lieu of capture or combustion. The duration mustbe specified in hours of time. (B) For each gas wellhead affected facility requiredto comply with the requirements of §60.5375(f), you mustmaintain the records specified in paragraph (c)(1)(iii)(A)of this section except that you do not have to record theduration of recovery to the sales line. In addition, youmust record the distance, in miles, of the nearestgathering line. (2) For each centrifugal compressor affected facility,you must maintain records on the type of seal systeminstalled. (3) For each reciprocating compressors affectedfacility, you must maintain the records in paragraphs(c)(3)(i) and (c)(3)(ii) of this section. (i) Records of the cumulative number of hours ofoperation since initial startup or the date of publication This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

402. Page 402 of 604of the final rule in the Federal Register, or the previousreplacement of the reciprocating compressor rod packing,whichever is later. (ii) Records of the date and time of eachreciprocating compressor rod packing replacement. (4) For each pneumatic controller affected facility,you must maintain the records identified in paragraphs(c)(4)(i) through (c)(4)(iv) of this section. (i) Records of the date, location and manufacturerspecifications for each pneumatic controller installed. (ii) Records of the determination that the use of highbleed pneumatic devices is predicated and the reasons why. (iii) If the pneumatic controller affected facility isnot located at a natural gas processing plant, records ofthe manufacturer’s guarantee that the device is designedsuch that natural gas emissions are less than 6 standardcubic feet per hour. (iv) If the pneumatic controller affected facility islocated at a natural gas processing plant, records of thedocumentation that only instrument air controllers areused. (5) For each storage vessel affected facility, you This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

403. Page 403 of 604must maintain the records identified in paragraphs(c)(5)(i) and (c)(5)(ii) of this section. (i) If required to reduce emissions by complying with§63.766, the records specified in 40 CFR part 63, subpartHH, §63.774(b)(2) through (8). (ii) Records of the determination that the annualaverage condensate throughput is less than 1 barrel per dayper storage vessel and crude oil throughput is less than 21barrels per day per storage vessel for the exemption under§60.5395(a)(1) and (a)(2).§60.5421 What are my additional recordkeeping requirementsfor my affected facility subject to VOC requirements foronshore natural gas processing plants? (a) You must comply with the requirements of paragraph(b)of this section in addition to the requirements of§60.486a. (b) The following recordkeeping requirements apply topressure relief devices subject to the requirements of§60.5401(b)(1) of this subpart. (1) When each leak is detected as specified in§60.5401(b)(2), a weatherproof and readily visibleidentification, marked with the equipment identification This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

404. Page 404 of 604number, must be attached to the leaking equipment. Theidentification on the pressure relief device may be removedafter it has been repaired. (2) When each leak is detected as specified in§60.5401(b)(2), the following information must be recordedin a log and shall be kept for 2 years in a readilyaccessible location: (i) The instrument and operator identification numbersand the equipment identification number. (ii) The date the leak was detected and the dates ofeach attempt to repair the leak. (iii) Repair methods applied in each attempt to repairthe leak. (iv) ”Above 500 ppm” if the maximum instrument readingmeasured by the methods specified in paragraph (a) of thissection after each repair attempt is 500 ppm or greater. (v) ”Repair delayed” and the reason for the delay if aleak is not repaired within 15 calendar days afterdiscovery of the leak. (vi) The signature of the owner or operator (ordesignate) whose decision it was that repair could not beeffected without a process shutdown. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

405. Page 405 of 604 (vii) The expected date of successful repair of theleak if a leak is not repaired within 15 days. (viii) Dates of process unit shutdowns that occur whilethe equipment is unrepaired. (ix) The date of successful repair of the leak. (x) A list of identification numbers for equipment thatare designated for no detectable emissions under theprovisions of §60.482-4a(a). The designation of equipmentsubject to the provisions of §60.482-4a(a) must be signedby the owner or operator.§60.5422 What are my additional reporting requirements formy affected facility subject to VOC requirements foronshore natural gas processing plants? (a) You must comply with the requirements ofparagraphs (b) and (c) of this section in addition to therequirements of §60.487a(a), (b), (c)(i) through (iv), and(c)(vii) through (viii). (b) An owner or operator must include the followinginformation in the initial semiannual report in addition tothe information required in §60.487a(b)(1) through (4):Number of pressure relief devices subject to therequirements of §60.5401(b) except for those pressure This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

406. Page 406 of 604relief devices designated for no detectable emissions underthe provisions of §60.482–4a(a) and those pressure reliefdevices complying with §60.482–4a(c). (c) An owner or operator must include the followinginformation in all semiannual reports in addition to theinformation required in §60.487a(c)(2)(i) through (vi): (1) Number of pressure relief devices for which leakswere detected as required in §60.5401(b)(2); and (2) Number of pressure relief devices for which leakswere not repaired as required in §60.5401(b)(3).§60.5423 What additional recordkeeping and reportingrequirements apply to my sweetening unit affectedfacilities at onshore natural gas processing plants? (a) You must retain records of the calculations andmeasurements required in §60.5405(a) and (b) and§60.5407(a) through (g) for at least 2 years following thedate of the measurements. This requirement is includedunder §60.7(d) of the General Provisions. (b) You must submit a written report of excessemissions to the Administrator semiannually. For thepurpose of these reports, excess emissions are defined as: (1) Any 24-hour period (at consistent intervals) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

407. Page 407 of 604during which the average sulfur emission reductionefficiency (R) is less than the minimum required efficiency(Z). (2) For any affected facility electing to comply withthe provisions of §60.5407(b)(2), any 24-hour period duringwhich the average temperature of the gases leaving thecombustion zone of an incinerator is less than theappropriate operating temperature as determined during themost recent performance test in accordance with theprovisions of §60.5407(b)(2). Each 24-hour period mustconsist of at least 96 temperature measurements equallyspaced over the 24 hours. (c) To certify that a facility is exempt from thecontrol requirements of these standards, for each facilitywith a design capacity less that 2 LT/D of H2S in the acidgas (expressed as sulfur) you must keep, for the life ofthe facility, an analysis demonstrating that the facilitysdesign capacity is less than 2 LT/D of H2S expressed assulfur. (d) If you elect to comply with §60.5407(e) you mustkeep, for the life of the facility, a record demonstratingthat the facilitys design capacity is less than 150 LT/D This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

408. Page 408 of 604of H2S expressed as sulfur. (e) The requirements of paragraph (b) of this sectionremain in force until and unless the EPA, in delegatingenforcement authority to a state under section 111(c) ofthe Act, approves reporting requirements or an alternativemeans of compliance surveillance adopted by such state. Inthat event, affected sources within the state will berelieved of obligation to comply with paragraph (b) of thissection, provided that they comply with the requirementsestablished by the state.§60.5425 What part of the General Provisions apply to me? Table 3 to this subpart shows which parts of theGeneral Provisions in §§60.1 through 60.19 apply to you.§60.5430 What definitions apply to this subpart? As used in this subpart, all terms not defined hereinshall have the meaning given them in the Act, in subpart Aor subpart VVa of part 60; and the following terms shallhave the specific meanings given them. Acid gas means a gas stream of hydrogen sulfide (H2S)and carbon dioxide (CO2) that has been separated from sournatural gas by a sweetening unit. Alaskan North Slope means the approximately 69,000 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

409. Page 409 of 604square-mile area extending from the Brooks Range to theArctic Ocean. API Gravity means the weight per unit volume ofhydrocarbon liquids as measured by a system recommended bythe American Petroleum Institute (API) and is expressed indegrees. Centrifugal Compressor means a piece of equipment thatcompresses a process gas by means of mechanical rotatingvanes or impellers. City gate means the delivery point at which naturalgas is transferred from a transmission pipeline to thelocal gas utility. Completion combustion device means any ignitiondevice, installed horizontally or vertically, used inexploration and production operations to combust otherwisevented emissions from completions or workovers. Compressor means a piece of equipment that compressesprocess gas and is usually a centrifugal compressor or areciprocating compressor. Compressor station means any permanent combination ofcompressors that move natural gas at increased pressurefrom fields, in transmission pipelines, or into storage. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

410. Page 410 of 604 Condensate means a hydrocarbon liquid separated fromnatural gas that condenses due to changes in thetemperature, pressure, or both, and remains liquid atstandard conditions, as specified in §60.2. For thepurposes of this subpart, a hydrocarbon liquid with an APIgravity equal to or greater than 40 degrees is consideredcondensate. Crude oil means crude petroleum oil any otherhydrocarbon liquid, which are produced at the well inliquid form by ordinary production methods, and which arenot the result of condensation of gas before or after itleaves the reservoir. For the purposes of this subpart, ahydrocarbon liquid with an API gravity less than 40 degreesis considered crude oil. Delineation well means a well drilled in order todetermine the boundary of a field or producing reservoir. Dehydrator means a device in which an absorbentdirectly contacts a natural gas stream and absorbs water ina contact tower or absorption column (absorber). Equipment means each pump, pressure relief device,open-ended valve or line, valve, compressor, and flange orother connector that is in VOC service or in wet gas This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

411. Page 411 of 604service, and any device or system required by this subpart. Field gas means feedstock gas entering the natural gasprocessing plant. Field Gas Gathering means the system used transportfield gas from a field to the main pipeline in the area. Flare means a thermal oxidation system using an open(without enclosure) flame. Flowback means the process of allowing fluids to flowfrom the well following a treatment, either in preparationfor a subsequent phase of treatment or in preparation forcleanup and returning the well to production. Flow line means surface pipe through which oil and/ornatural gas travels from the well. Gas-driven pneumatic controller means a pneumaticcontroller powered by pressurized natural gas. Gas processing plant process unit means equipmentassembled for the extraction of natural gas liquids fromfield gas, the fractionation of the liquids into naturalgas products, or other operations associated with theprocessing of natural gas products. A process unit canoperate independently if supplied with sufficient feed orraw materials and sufficient storage facilities for the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

412. Page 412 of 604products. Gas well means a well, the principal production ofwhich at the mouth of the well is gas. High-Bleed Pneumatic Devices means automated,continuous bleed flow control devices powered bypressurized natural gas and used for maintaining a processcondition such as liquid level, pressure, delta-pressureand temperature. Part of the gas power stream which isregulated by the process condition flows to a valveactuator controller where it vents continuously (bleeds) tothe atmosphere at a rate in excess of six standard cubicfeet per hour. Hydraulic fracturing means process of directingpressurized liquids, containing water, proppant, and anyadded chemicals, to penetrate tight sand, shale, or coalformations that involve high rate, extended back flow toexpel fracture fluids and sand during completions and wellworkovers. In light liquid service means that the piece ofequipment contains a liquid that meets the conditionsspecified in §60.485a(e) or §60.5401(h)(2) of this part. In wet gas service means that a compressor or piece of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

413. Page 413 of 604equipment contains or contacts the field gas before theextraction step at a gas processing plant process unit. Liquefied natural gas unit means a unit used to coolnatural gas to the point at which it is condensed into aliquid which is colorless, odorless, non-corrosive and non-toxic. Low-bleed pneumatic controller means automated flowcontrol devices powered by pressurized natural gas and usedfor maintaining a process condition such as liquid level,pressure, delta-pressure and temperature. Part of the gaspower stream which is regulated by the process conditionflows to a valve actuator controller where it ventscontinuously (bleeds) to the atmosphere at a rate equal toor less than six standard cubic feet per hour. Modification means any physical change in, or changein the method of operation of, an affected facility whichincreases the amount of VOC or natural gas emitted into theatmosphere by that facility or which results in theemission of VOC or natural gas into the atmosphere notpreviously emitted. For the purposes of this subpart, eachrecompletion of a fractured or refractured existing gaswell is considered to be a modification. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

414. Page 414 of 604 Natural gas liquids means the hydrocarbons, such asethane, propane, butane, and pentane that are extractedfrom field gas. Natural gas processing plant (gas plant) means anyprocessing site engaged in the extraction of natural gasliquids from field gas, fractionation of mixed natural gasliquids to natural gas products, or both. Nonfractionating plant means any gas plant that doesnot fractionate mixed natural gas liquids into natural gasproducts. Non gas-driven pneumatic device means an instrumentthat is actuated using other sources of power thanpressurized natural gas; examples include solar, electric,and instrument air. Onshore means all facilities except those that arelocated in the territorial seas or on the outer continentalshelf. Plunger lift system means an intermittent gas liftthat uses gas pressure buildup in the casing-tubing annulusto push a steel plunger, and the column of fluid ahead ofit, up the well tubing to the surface. Pneumatic controller means an automated instrument This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

415. Page 415 of 604used for maintaining a process condition such as liquidlevel, pressure, delta-pressure and temperature. Pneumatic pump means a pump that uses pressurizednatural gas to move a piston or diaphragm, which pumpsliquids on the opposite side of the piston or diaphragm. Process unit means components assembled for theextraction of natural gas liquids from field gas, thefractionation of the liquids into natural gas products, orother operations associated with the processing of naturalgas products. A process unit can operate independently ifsupplied with sufficient feed or raw materials andsufficient storage facilities for the products. Reciprocating compressor means a piece of equipmentthat increases the pressure of a process gas by positivedisplacement, employing linear movement of the driveshaft. Reciprocating compressor rod packing means a series offlexible rings in machined metal cups that fit around thereciprocating compressor piston rod to create a seallimiting the amount of compressed natural gas that escapesto the atmosphere. Reduced emissions completion means a well completionwhere gas flowback that is otherwise vented is captured, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

416. Page 416 of 604cleaned, and routed to the sales line. Reduced emissions recompletion means a well completionfollowing refracturing of a gas well where gas flowbackthat is otherwise vented is captured, cleaned, and routedto the sales line. Reduced sulfur compounds means H2S, carbonyl sulfide(COS), and carbon disulfide (CS2). Routed to a process or route to a process means theemissions are conveyed to any enclosed portion of a processunit where the emissions are predominantly recycled and/orconsumed in the same manner as a material that fulfills thesame function in the process and/or transformed by chemicalreaction into materials that are not regulated materialsand/or incorporated into a product; and/or recovered. Sales line means pipeline, generally small indiameter, used to transport oil or gas from the well to aprocessing facility or a mainline pipeline. Salable quality gas means natural gas that meets thecomposition, moisture, or other limits set by the purchaserof the natural gas. Storage vessel means a stationary vessel or series ofstationary vessels that are either manifolded together or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

417. Page 417 of 604are located at a single well site and that have potentialfor VOC emissions equal to or greater than 10 tpy. Sulfur production rate means the rate of liquid sulfuraccumulation from the sulfur recovery unit. Sulfur recovery unit means a process device thatrecovers element sulfur from acid gas. Sweetening unit means a process device that removeshydrogen sulfide and/or carbon dioxide from the natural gasstream. Surface site means any combination of one or moregraded pad sites, gravel pad sites, foundations, platforms,or the immediate physical location upon which equipment isphysically affixed. Total SO2 equivalents means the sum of volumetric ormass concentrations of the sulfur compounds obtained byadding the quantity existing as SO2 to the quantity of SO2that would be obtained if all reduced sulfur compounds wereconverted to SO2 (ppmv or kg/dscm (lb/dscf)). Total Reduced Sulfur (TRS) means the sum of the sulfurcompounds hydrogen sulfide, methyl mercaptan, dimethylsulfide, and dimethyl disulfide as measured by Method 16 ofappendix A to part 60 of this chapter. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

418. Page 418 of 604 Underground storage tank means a storage tank storedbelow ground. Well means an oil or gas well, a hole drilled for thepurpose of producing oil or gas, or a well into whichfluids are injected. Well completion means the process that allows for theflow of petroleum or natural gas from newly drilled wellsto expel drilling and reservoir fluids and tests thereservoir flow characteristics, steps which may ventproduced gas to the atmosphere via an open pit or tank.Well completion also involves connecting the well bore tothe reservoir, which may include treating the formation orinstalling tubing, packer(s), or lifting equipment. Well completion operation means any well completion orwell workover occurring at a gas wellhead affectedfacility. Well site means the areas that are directly disturbedduring the drilling and subsequent operation of, oraffected by, production facilities directly associated withany oil well, gas well, or injection well and itsassociated well pad. Wellhead means the piping, casing, tubing and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

419. Page 419 of 604connected valves protruding above the earth’s surface foran oil and/or natural gas well. The wellhead ends where theflow line connects to a wellhead valve. The wellhead doesnot include other equipment at the well site except for anyconveyance through which gas is vented to the atmosphere. Wildcat well means a well outside known fields or thefirst well drilled in an oil or gas field where no otheroil and gas production exists. Table 1. Required Minimum Initial SO2 Emission Reduction Efficiency (Zi) H2Scontent of Sulfur feed rate (X), LT/D acid gas (Y), % 2.0<X<5.0 5.0<X<15.0 15.0<X<300.0 X>300.0 Y>50 79.0 ............ 88.51X0.0101Y0.0125 ............. or 99.9, whichever is smaller 20<Y<50 79.0 .......... 88.5X0.0101Y0.0125 .......... 97.9 or 97.9, whichever is smaller 10<Y<20 79.0 88.5X0.0101Y0.0125 93.5 93.5 or 97.9, whichever is smaller Y<10 79.0 79.0 79.0 79.0Table 2. Required Minimum SO2 Emission Reduction Efficiency (Zc) H2Scontent of Sulfur feed rate (X), LT/D acid gas (Y), % 2.0<X<5.0 5.0<X<15.0 15.0<X<300.0 X>300.0 Y>50 74.0 ............ 85.35X0.0144Y0.0128 ............. or 99.9, whichever is smaller This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

420. Page 420 of 604 20<Y<50 74.0 .......... 85.35X0.0144Y0.0128 ......... 97.5 or 97.9, whichever is smaller 10<Y<20 74.0 85.35X0.0144Y0.0128 90.8 90.8 or 90.8, whichever is smaller Y<10 74.0 74.0 74.0 74.0E = The sulfur emission rate expressed as elemental sulfur,kilograms per hour (kg/hr) [pounds per hour (lb/hr)],rounded to one decimal place.R = The sulfur emission reduction efficiency achieved inpercent, carried to one decimal place.S = The sulfur production rate, kilograms per hour (kg/hr)[pounds per hour (lb/hr)], rounded to one decimal place.X = The sulfur feed rate from the sweetening unit (i.e.,the H2S in the acid gas), expressed as sulfur, Mg/D(LT/D),rounded to one decimal place.Y = The sulfur content of the acid gas from the sweeteningunit, expressed as mole percent H2S (dry basis) rounded toone decimal place.Z = The minimum required sulfur dioxide (SO2) emissionreduction efficiency, expressed as percent carried to onedecimal place. Zi refers to the reduction efficiencyrequired at the initial performance test. Zc refers to thereduction efficiency required on a continuous basis aftercompliance with Zi has been demonstrated.Table 3 to Subpart OOOO of part 60 – Applicability ofGeneral Provisions to Subpart OOOOAs stated in §60.5425, you must comply with the followingapplicable General Provisions: General AppliesProvisions Subject of to Citation Citation subpart? Explanation General applicability of§60.1 Yes the General Provisions§60.2 Definitions Yes Additional terms defined in §60.5430 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

421. Page 421 of 604§60.3 Units and Yes abbreviations§60.4 Address Yes§60.5 Determination of Yes construction or modification§60.6 Review of plans Yes§60.7 Notification and Yes Except that §60.7 record keeping only applies as specified in §60.5420(a).§60.8 Performance tests No Performance testing is required for storage vessels as specified in 40 CFR part 63, subpart HH.§60.9 Availability of Yes information§60.10 State authority Yes§60.11 Compliance with No Requirements are standards and specified in maintenance subpart OOOO. requirements§60.12 Circumvention Yes§60.13 Monitoring Yes Continuous requirements monitors are required for storage vessels§60.14 Modification Yes§60.15 Reconstruction Yes§60.16 Priority list Yes§60.17 Incorporations by Yes reference§60.18 General control Yes device requirements§60.19 General Yes notification and reporting requirement This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

422. Page 422 of 604PART 63--[AMENDED] 8. The authority citation for part 63 continues toread as follows:Authority: 42 U.S.C. 7401, et seq. 9. Section 63.14 is amended by: a. Adding paragraphs (b)(67), (b)(68), (b)(69) and(b)(70); and b. Revising paragraph (i)(1) to read as follows:§63.14 Incorporations by reference.* * * * * (b) * * ** * * * * (67) ASTM D1945 – 03(2010) Standard Test Method forAnalysis of Natural Gas by Gas Chromatography, IBR approvedfor §§63.772 and 63.1282. (68) ASTM D5504 - 08 Standard Test Method forDetermination of Sulfur Compounds in Natural Gas andGaseous Fuels by Gas Chromatography and Chemiluminescence ,IBR approved for §§63.772 and 63.1282. (69) ASTM D3588 - 98(2003) Standard Practice forCalculating Heat Value, Compressibility Factor, andRelative Density of Gaseous Fuels, IBR approved for This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

423. Page 423 of 604§§63.772 and 63.1282. (70) ASTM D4891 - 89(2006) Standard Test Method forHeating Value of Gases in Natural Gas Range byStoichiometric Combustion, IBR approved for §§63.772 and63.1282.* * * * * (i) * * * (1) ANSI/ASME PTC 19.10–1981, Flue and Exhaust GasAnalyses [Part 10, Instruments and Apparatus], issuedAugust 31, 1981 IBR approved for §§63.309(k)(1)(iii),63.771(e), 63.865(b), 63.1281(d), 63.3166(a)(3),63.3360(e)(1)(iii), 63.3545(a)(3), 63.3555(a)(3),63.4166(a)(3), 63.4362(a)(3), 63.4766(a)(3), 63.4965(a)(3),63.5160(d)(1)(iii), 63.9307(c)(2), 63.9323(a)(3),63.11148(e)(3)(iii), 63.11155(e)(3), 63.11162(f)(3)(iii)and (f)(4), 63.11163(g)(1)(iii) and (g)(2),63.11410(j)(1)(iii), 63.11551(a)(2)(i)(C),63.11646(a)(1)(iii), table 5 to subpart DDDDD of this part,and table 1 to subpart ZZZZZ of this part.* * * * *Subpart HH-–[AMENDED] 10. Section 63.760 is amended by: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

424. Page 424 of 604 a. Revising paragraph (a)(1) introductory text; b. Revising paragraph (a)(1)(iii); c. Revising paragraph (a)(2); d. Revising paragraph (b)(1)(ii); e. Revising paragraph (f) introductory text; f. Revising paragraph (f)(1); g. Revising paragraph (f)(2); and h. Adding paragraphs (f)(7), (f)(8), (f)(9) and(f)(10) to read as follows:§63.760 Applicability and designation of affected source. (a) * * * (1) Facilities that are major or area sources ofhazardous air pollutants (HAP) as defined in §63.761.Emissions for major source determination purposes can beestimated using the maximum natural gas or hydrocarbonliquid throughput, as appropriate, calculated in paragraphs(a)(1)(i) through (iii) of this section. As an alternativeto calculating the maximum natural gas or hydrocarbonliquid throughput, the owner or operator of a new orexisting source may use the facilitys design maximumnatural gas or hydrocarbon liquid throughput to estimatethe maximum potential emissions. Other means to determine This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

425. Page 425 of 604the facilitys major source status are allowed, providedthe information is documented and recorded to theAdministrators satisfaction in accordance with§63.10(b)(3). A facility that is determined to be an areasource, but subsequently increases its emissions or itspotential to emit above the major source levels, andbecomes a major source, must comply thereafter with allprovisions of this subpart applicable to a major sourcestarting on the applicable compliance date specified inparagraph (f) of this section. Nothing in this paragraph isintended to preclude a source from limiting its potentialto emit through other appropriate mechanisms that may beavailable through the permitting authority.* * * * * (iii) The owner or operator shall determine themaximum values for other parameters used to calculateemissions as the maximum for the period over which themaximum natural gas or hydrocarbon liquid throughput isdetermined in accordance with paragraph (a)(1)(i)(A) or (B)of this section. Parameters, other than glycol circulationrate, shall be based on either highest measured values orannual average. For estimating maximum potential emissions This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

426. Page 426 of 604from glycol dehydration units, the glycol circulation rateused in the calculation shall be the unit’s maximum rateunder its physical and operational design consistent withthe definition of potential to emit in §63.2. (2) Facilities that process, upgrade, or storehydrocarbon liquids prior to the point where hydrocarbonliquids enter either the Organic Liquids Distribution (Non-gasoline) or Petroleum Refineries source categories.* * * * * (b) * * * (1) * * ** * * * * (ii) Each storage vessel;* * * * * (f) The owner or operator of an affected major sourceshall achieve compliance with the provisions of thissubpart by the dates specified in paragraphs (f)(1),(f)(2), and (f)(7) through (f)(10) of this section. Theowner or operator of an affected area source shall achievecompliance with the provisions of this subpart by the datesspecified in paragraphs (f)(3) through (f)(6) of thissection. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

427. Page 427 of 604 (1) Except as specified in paragraphs (f)(7) through(10) of this section, the owner or operator of an affectedmajor source, the construction or reconstruction of whichcommenced before February 6, 1998, shall achieve compliancewith the applicable provisions of this subpart no laterthan June 17, 2002, except as provided for in §63.6(i). Theowner or operator of an area source, the construction orreconstruction of which commenced before February 6, 1998,that increases its emissions of (or its potential to emit)HAP such that the source becomes a major source that issubject to this subpart shall comply with this subpart 3years after becoming a major source. (2) Except as specified in paragraphs (f)(7) through(10) of this section, the owner or operator of an affectedmajor source, the construction or reconstruction of whichcommences on or after February 6, 1998, shall achievecompliance with the applicable provisions of this subpartimmediately upon initial startup or June 17, 1999,whichever date is later. Area sources, other thanproduction field facilities identified in (f)(9) of thissection, the construction or reconstruction of whichcommences on or after February 6, 1998, that become major This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

428. Page 428 of 604sources shall comply with the provisions of this standardimmediately upon becoming a major source.* * * * * (7) Each affected small glycol dehydration unit andeach storage vessel that is not a storage vessel with thepotential for flash emissions located at a major source,that commenced construction before [INSERT DATE OFPUBLICATION IN THE FEDERAL REGISTER] must achievecompliance no later than 3 years after the date ofpublication of the final rule in the Federal Register,except as provided in §63.6(i). (8) Each affected small glycol dehydration unit andeach storage vessel that is not a storage vessel with thepotential for flash emissions, both as defined in §63.761,located at a major source, that commenced construction onor after [INSERT DATE OF PUBLICATION IN THE FEDERALREGISTER] must achieve compliance immediately upon initialstartup or the date of publication of the final rule in theFederal Register, whichever is later. (9) A production field facility, as defined in§63.761, constructed before [INSERT DATE OF PUBLICATION INTHE FEDERAL REGISTER] that was previously determined to be This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

429. Page 429 of 604an area source but becomes a major source (as defined inparagraph 3 of the major source definition in §63.761) onthe date of publication of the final rule in the FederalRegister must achieve compliance no later than 3 yearsafter the date of publication of the final rule in theFederal Register, except as provided in §63.6(i).(10) Each large glycol dehydration unit, as defined in§63.761, that has complied with the provisions of thissubpart prior to [INSERT DATE OF PUBLICATION IN THE FEDERALREGISTER] by reducing its benzene emissions to less than0.9 megagrams per year must achieve compliance no laterthan 90 days after the date of publication of the finalrule in the Federal Register, except as provided in§63.6(i).* * * * * 11. Section 63.761 is amended by: a. Adding, in alphabetical order, new definitions forthe terms “affirmative defense,” “BTEX,” “flare,” “largeglycol dehydration units” and “small glycol dehydrationunits”; b. Revising the definition for “associated equipment,” This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

430. Page 430 of 604“facility,” “glycol dehydration unit baseline operations,”and “temperature monitoring device”; and c. Revising paragraph (3) of the definition for “majorsource” to read as follows:§63.761 Definitions.* * * * * Affirmative defense means, in the context of anenforcement proceeding, a response or defense put forwardby a defendant, regarding which the defendant has theburden of proof, and the merits of which are independentlyand objectively evaluated in a judicial or administrativeproceeding.* * * * * Associated equipment, as used in this subpart and asreferred to in section 112(n)(4) of the Act, meansequipment associated with an oil or natural gas explorationor production well, and includes all equipment from thewellbore to the point of custody transfer, except glycoldehydration units and storage vessels.* * * * * BTEX means benzene, toluene, ethyl benzene and xylene.* * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

431. Page 431 of 604 Facility means any grouping of equipment wherehydrocarbon liquids are processed, upgraded (i.e., removeimpurities or other constituents to meet contractspecifications), or stored; or where natural gas isprocessed, upgraded, or stored. For the purpose of a majorsource determination, facility (including a building,structure, or installation) means oil and natural gasproduction and processing equipment that is located withinthe boundaries of an individual surface site as defined inthis section. Equipment that is part of a facility willtypically be located within close proximity to otherequipment located at the same facility. Pieces ofproduction equipment or groupings of equipment located ondifferent oil and gas leases, mineral fee tracts, leasetracts, subsurface or surface unit areas, surface feetracts, surface lease tracts, or separate surface sites,whether or not connected by a road, waterway, power line orpipeline, shall not be considered part of the samefacility. Examples of facilities in the oil and natural gasproduction source category include, but are not limited to,well sites, satellite tank batteries, central tankbatteries, a compressor station that transports natural gas This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

432. Page 432 of 604to a natural gas processing plant, and natural gasprocessing plants.* * * * * Flare means a thermal oxidation system using an openflame (i.e., without enclosure).* * * * * Glycol dehydration unit baseline operations meansoperations representative of the large glycol dehydrationunit operations as of June 17, 1999 and the small glycoldehydrator unit operations as of [INSERT DATE OFPUBLICATION IN THE FEDERAL REGISTER]. For the purposes ofthis subpart, for determining the percentage of overall HAPemission reduction attributable to process modifications,baseline operations shall be parameter values (including,but not limited to, glycol circulation rate or glycol-HAPabsorbency) that represent actual long-term conditions(i.e., at least 1 year). Glycol dehydration units inoperation for less than 1 year shall document that theparameter values represent expected long-term operatingconditions had process modifications not been made.* * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

433. Page 433 of 604 Large glycol dehydration unit means a glycoldehydration unit with an actual annual average natural gasflowrate equal to or greater than 85 thousand standardcubic meters per day and actual annual average benzeneemissions equal to or greater than 0.90 Mg/yr, determinedaccording to §63.772(b).* * * * * Major source * * * (3) For facilities that are production fieldfacilities, only HAP emissions from glycol dehydrationunits and storage vessels shall be aggregated for a majorsource determination. For facilities that are notproduction field facilities, HAP emissions from all HAPemission units shall be aggregated for a major sourcedetermination.* * * * * Small glycol dehydration unit means a glycoldehydration unit, located at a major source, with an actualannual average natural gas flowrate less than 85 thousandstandard cubic meters per day or actual annual averagebenzene emissions less than 0.90 Mg/yr, determinedaccording to §63.772(b). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

434. Page 434 of 604* * * * * Temperature monitoring device means an instrument usedto monitor temperature and having a minimum accuracy of ±1percent of the temperature being monitored expressed in °C,or ±2.5 °C, whichever is greater. The temperaturemonitoring device may measure temperature in degreesFahrenheit or degrees Celsius, or both. 12. Section 63.762 is amended by: a. Revising paragraphs (a) through (d); and b. Removing paragraph (e) to read as follows:§63.762 Startups and shutdowns. (a) The provisions set forth in this subpart shallapply at all times. (b) The owner or operator shall not shut down items ofequipment that are required or utilized for compliance withthe provisions of this subpart during times when emissionsare being routed to such items of equipment, if theshutdown would contravene requirements of this subpartapplicable to such items of equipment. This paragraph doesnot apply if the owner or operator must shut down theequipment to avoid damage due to a contemporaneous startupor shutdown, of the affected source or a portion thereof. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

435. Page 435 of 604 (c) During startups and shutdowns, the owner oroperator shall implement measures to prevent or minimizeexcess emissions to the maximum extent practical. (d) In response to an action to enforce the standardsset forth in this subpart, you may assert an affirmativedefense to a claim for civil penalties for exceedances ofsuch standards that are caused by malfunction, as definedin 40 CFR 63.2. Appropriate penalties may be assessed,however, if you fail to meet your burden of proving all therequirements in the affirmative defense. The affirmativedefense shall not be available for claims for injunctiverelief. (1) To establish the affirmative defense in any actionto enforce such a limit, you must timely meet thenotification requirements in paragraph (d)(2) of thissection, and must prove by a preponderance of evidencethat: (i) The excess emissions: (A) Were caused by a sudden, infrequent, andunavoidable failure of air pollution control and monitoringequipment, process equipment, or a process to operate in anormal or usual manner; and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

436. Page 436 of 604 (B) Could not have been prevented through carefulplanning, proper design or better operation and maintenancepractices; and (C) Did not stem from any activity or event that couldhave been foreseen and avoided, or planned for; and (D) Were not part of a recurring pattern indicative ofinadequate design, operation, or maintenance; and (ii) Repairs were made as expeditiously as possiblewhen the applicable emission limitations were beingexceeded. Off-shift and overtime labor were used, to theextent practicable to make these repairs; and (iii) The frequency, amount and duration of the excessemissions (including any bypass) were minimized to themaximum extent practicable during periods of suchemissions; and (iv) If the excess emissions resulted from a bypass ofcontrol equipment or a process, then the bypass wasunavoidable to prevent loss of life, personal injury, orsevere property damage; and (v) All possible steps were taken to minimize theimpact of the excess emissions on ambient air quality, theenvironment, and human health; and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

437. Page 437 of 604 (vi) All emissions monitoring and control systems werekept in operation if at all possible, consistent withsafety and good air pollution control practices; and (vii) All of the actions in response to the excessemissions were documented by properly signed,contemporaneous operating logs; and (viii) At all times, the affected source was operatedin a manner consistent with good practices for minimizingemissions; and (ix) A written root cause analysis has been preparedto determine, correct, and eliminate the primary causes ofthe malfunction and the excess emissions resulting from themalfunction event at issue. The analysis shall alsospecify, using best monitoring methods and engineeringjudgment, the amount of excess emissions that were theresult of the malfunction. (2) Notification. The owner or operator of theaffected source experiencing exceedance of its emissionlimit(s) during a malfunction shall notify theAdministrator by telephone or facsimile transmission assoon as possible, but no later than two business days afterthe initial occurrence of the malfunction, if it wishes to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

438. Page 438 of 604avail itself of an affirmative defense to civil penaltiesfor that malfunction. The owner or operator seeking toassert an affirmative defense shall also submit a writtenreport to the Administrator within 45 days of the initialoccurrence of the exceedance of the standard in thissubpart to demonstrate, with all necessary supportingdocumentation, that it has met the requirements set forthin paragraph (d)(1) of this section. The owner or operatormay seek an extension of this deadline for up to 30additional days by submitting a written request to theAdministrator before the expiration of the 45 day period.Until a request for an extension has been approved by theAdministrator, the owner or operator is subject to therequirement to submit such report within 45 days of theinitial occurrence of the exceedance.* * * * * 13. Section 63.764 is amended by: a. Revising paragraph (c)(2) introductory text; b. Revising paragraph (e) introductory text; and c. Revising paragraph (i) introductory text; and d. Adding paragraph (j) to read as follows:§63.764 General standards. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

439. Page 439 of 604* * * * * (c) * * ** * * * * (2) For each storage vessel subject to this subpart,the owner or operator shall comply with the requirementsspecified in paragraphs (c)(2)(i) through (iii) of thissection.* * * * * (e) Exemptions. (1) The owner or operator of an areasource is exempt from the requirements of paragraph (d) ofthis section if the criteria listed in paragraph (e)(1)(i)or (ii) of this section are met, except that the records ofthe determination of these criteria must be maintained asrequired in §63.774(d)(1).* * * * * (i) In all cases where the provisions of this subpartrequire an owner or operator to repair leaks by a specifiedtime after the leak is detected, it is a violation of thisstandard to fail to take action to repair the leak(s)within the specified time. If action is taken to repair theleak(s) within the specified time, failure of that actionto successfully repair the leak(s) is not a violation of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

440. Page 440 of 604this standard. However, if the repairs are unsuccessful,and a leak is detected, the owner or operator shall takefurther action as required by the applicable provisions ofthis subpart. (j) At all times the owner or operator must operateand maintain any affected source, including associated airpollution control equipment and monitoring equipment, in amanner consistent with safety and good air pollutioncontrol practices for minimizing emissions. Determinationof whether such operation and maintenance procedures arebeing used will be based on information available to theAdministrator which may include, but is not limited to,monitoring results, review of operation and maintenanceprocedures, review of operation and maintenance records,and inspection of the source.* * * * * 14. Section 63.765 is amended by: a. Revising paragraph (a); b. Revising paragraph (b)(1) introductory text; c. Revising paragraphs (b)(1)(i) and (ii); d. Adding paragraph (b)(1)(iii); e. Revising paragraph (c)(2); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

441. Page 441 of 604 f. Revising paragraph (c)(3) introductory text; g. Revising paragraphs (c)(3)(i) and (ii); and h. Adding paragraph (c)(3)(iii) to read as follows:§63.765 Glycol dehydration unit process vent standards. (a) This section applies to each glycol dehydrationunit subject to this subpart that must be controlled forair emissions as specified in either paragraph (c)(1)(i) orparagraph (d)(1)(i) of §63.764. (b) * * * (1) For each glycol dehydration unit process vent, theowner or operator shall control air emissions by eitherparagraph (b)(1)(i), (ii), or (iii) of this section. (i) The owner or operator of a large glycoldehydration unit, as defined in §63.761, shall connect theprocess vent to a control device or a combination ofcontrol devices through a closed-vent system. The closed-vent system shall be designed and operated in accordancewith the requirements of §63.771(c). The control device(s)shall be designed and operated in accordance with therequirements of §63.771(d). (ii) The owner or operator of a glycol dehydrationunit located at an area source, that must be controlled as This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

442. Page 442 of 60 04spec cified in §63.764 n 4(d)(1)(i i), shall connect the pro l t ocessvent to a co t ontrol de evice or combinat tion of c control d devicesthro ough a cl losed-ven system and the outlet benzene nt m eemis ssions fr rom the c control d device(s) shall b reduce to a ) be edleve less than 0.90 megagra el t 0 ams per y year. The closed- e -ventsyst tem shall be desi l igned and operate in acc d ed cordance withthe requirem (c). The control device(s shall ments of §63.771( s)be d designed and oper rated in accordan nce with therequ uirements of §63. s .771(d), except t that the performa anceleve ified in §63.771( els speci (d)(1)(i) and (ii do not apply. ) i) t (iii) You must limit BT Y TEX emiss sions fro each s om smallglyc col dehyd dration u unit proc cess vent as def t, fined in§63. .761, to the limi determ it mined in Equation 1 of th n hissect tion. The limit m e must be m met in ac ccordance with on of e nethe alternat tives spe ecified i paragr in raphs (b) )(1)(iii) )(A) ough (D) of this section.thro .Equa ation 1Wher re:ELBTE = Unit EX t-specifi BTEX e ic emission limit, m megagrams per syearr;1.100x10-4 = BTEX emission limit, gra ams BTEX/ /standard cubic dmete -ppmv; er ; Th his document is a prepuublication n version n, signed d by EP PA Ad dministrat tor, Lisa P. Jackso on 07/2 on 28/2011. We have t taken step ps to ensure the accura o t acy of thi version but it is not th officia is n, he al ve ersion.

443. Page 443 of 604Throughput = Annual average daily natural gas throughput,standard cubic meters per day;Ci,BTEX = BTEX concentration of the natural gas at the inletto the glycol dehydration unit, ppmv. (A) Connect the process vent to a control device orcombination of control devices through a closed-ventsystem. The closed vent system shall be designed andoperated in accordance with the requirements of §63.771(c).The control device(s) shall be designed and operated inaccordance with the requirements of §63.771(f). (B) Meet the emissions limit through processmodifications in accordance with the requirements specifiedin §63.771(e). (C) Meet the emissions limit for each small glycoldehydration unit using a combination of processmodifications and one or more control devices through therequirements specified in paragraphs (b)(1)(iii)(A) and (B)of this section. (D) Demonstrate that the emissions limit is metthrough actual uncontrolled operation of the small glycoldehydration unit. Document operational parameters inaccordance with the requirements specified in §63.771(e)and emissions in accordance with the requirements specifiedin §63.772(b)(2). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

444. Page 444 of 604* * * * * (c) * * ** * * * * (2) The owner or operator shall demonstrate, to theAdministrators satisfaction, that the total HAP emissionsto the atmosphere from the large glycol dehydration unitprocess vent are reduced by 95.0 percent through processmodifications, or a combination of process modificationsand one or more control devices, in accordance with therequirements specified in §63.771(e). (3) Control of HAP emissions from a GCG separator(flash tank) vent is not required if the owner or operatordemonstrates, to the Administrators satisfaction, thattotal emissions to the atmosphere from the glycoldehydration unit process vent are reduced by one of thelevels specified in paragraph (c)(3)(i), (ii), or (iii) ofthis section, through the installation and operation ofcontrols as specified in paragraph (b)(1) of this section. (i) For any large glycol dehydration unit, HAPemissions are reduced by 95.0 percent or more. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

445. Page 445 of 604 (ii) For area source dehydration units, benzeneemissions are reduced to a level less than 0.90 megagramsper year. (iii) For each small glycol dehydration unit, BTEXemissions are reduced to a level less than the limitcalculated by paragraph (b)(1)(iii) of this section. 15. Section 63.766 is amended by: a. Revising paragraph (a); b. Revising paragraph (b) introductory text; c. Revising paragraph (b)(1); and d. Revising paragraph (d) to read as follows:§63.766 Storage vessel standards. (a) This section applies to each storage vessel (asdefined in §63.761) subject to this subpart. (b) The owner or operator of a storage vessel (asdefined in §63.761) shall comply with one of the controlrequirements specified in paragraphs (b)(1) and (2) of thissection. (1) The owner or operator shall equip the affectedstorage vessel with a cover that is connected, through aclosed-vent system that meets the conditions specified in§63.771(c), to a control device or a combination of control This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

446. Page 446 of 604devices that meets any of the conditions specified in§63.771(d). The cover shall be designed and operated inaccordance with the requirements of §63.771(b).* * * * * (d) This section does not apply to storage vessels forwhich the owner or operator is subject to and controlledunder the requirements specified in 40 CFR part 60, subpartKb; or the requirements specified under 40 CFR part 63subparts G or CC.* * * * * 16. Section 63.769 is amended by: a. Revising paragraph (b); b. Revising paragraph (c); and b. Revising paragraph (c)(8) to read as follows:§63.769 Equipment leak standards.* * * * * (b) This section does not apply to ancillary equipmentand compressors for which the owner or operator is subjectto and controlled under the requirements specified insubpart H of this part; or the requirements specified in 40CFR part 60, subpart KKK. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

447. Page 447 of 604 (c) For each piece of ancillary equipment and eachcompressor subject to this section located at an existingor new source, the owner or operator shall meet therequirements specified in 40 CFR part 61, subpart V,§§61.241 through 61.247, except as specified in paragraphs(c)(1) through (8) of this section, except for valvessubject to §61.247-2(b) a leak is detected if an instrumentreading of 500 ppm or greater is measured.* * * * * (8) Flares, as defined in §63.761, used to comply withthis subpart shall comply with the requirements of§63.11(b).* * * * * 17. Section 63.771 is amended by: a. Revising paragraph (c)(1) introductory text; b. Revising paragraph (d) introductory text; c. Revising paragraph (d)(1)(i); d. Revising paragraph (d)(1)(i)(C); e. Revising paragraph (d)(1)(ii); f. Revising paragraph (d)(1)(iii); g. Revising paragraph (d)(4)(i); h. Revising paragraph (d)(5)(i); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

448. Page 448 of 604 i. Revising paragraph (e)(2); j. Revising paragraph (e)(3) introductory text; k. Revising paragraph (e)(3)(ii); and l. Adding paragraph (f) to read as follows:§63.771 Control equipment requirements.* * * * * (c) Closed-vent system requirements. (1) The closed-vent system shall route all gases, vapors, and fumesemitted from the material in an emissions unit to a controldevice that meets the requirements specified in paragraph(d) of this section.* * * * * (d) Control device requirements for sources exceptsmall glycol dehydration units. Owners and operators ofsmall glycol dehydration units, shall comply with thecontrol device requirements in paragraph (f) of thissection. (1) * * * (i) An enclosed combustion device (e.g., thermal vaporincinerator, catalytic vapor incinerator, boiler, orprocess heater) that is designed and operated in accordancewith one of the following performance requirements: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

449. Page 449 of 604* * * * * (C) For a control device that can demonstrate auniform combustion zone temperature during the performancetest conducted under §63.772(e), operates at a minimumtemperature of 760 degrees C.* * * * * (ii) A vapor recovery device (e.g., carbon adsorptionsystem or condenser) or other non-destructive control devicethat is designed and operated to reduce the mass content ofeither TOC or total HAP in the gases vented to the deviceby 95.0 percent by weight or greater as determined inaccordance with the requirements of §63.772(e). (iii) A flare, as defined in §63.761, that is designedand operated in accordance with the requirements of§63.11(b).* * * * * (4) * * * (i) Each control device used to comply with thissubpart shall be operating at all times when gases, vapors,and fumes are vented from the HAP emissions unit or unitsthrough the closed-vent system to the control device, asrequired under §63.765, §63.766, and §63.769. An owner or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

450. Page 450 of 604operator may vent more than one unit to a control deviceused to comply with this subpart.* * * * * (5) * * * (i) Following the initial startup of the controldevice, all carbon in the control device shall be replacedwith fresh carbon on a regular, predetermined time intervalthat is no longer than the carbon service life establishedfor the carbon adsorption system. Records identifying theschedule for replacement and records of each carbonreplacement shall be maintained as required in§63.774(b)(7)(ix). The schedule for replacement shall besubmitted with the Notification of Compliance Status Reportas specified in §63.775(d)(5)(iv). Each carbon replacementmust be reported in the Periodic Reports as specified in§63.772(e)(2)(xii).* * * * * (e) * * ** * * * * (2) The owner or operator shall document, to theAdministrators satisfaction, the conditions for whichglycol dehydration unit baseline operations shall be This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

451. Page 451 of 604modified to achieve the 95.0 percent overall HAP emissionreduction, or BTEX limit determined in §63.765(b)(1)(iii),as applicable, either through process modifications orthrough a combination of process modifications and one ormore control devices. If a combination of processmodifications and one or more control devices are used, theowner or operator shall also establish the emissionreduction to be achieved by the control device to achievean overall HAP emission reduction of 95.0 percent for theglycol dehydration unit process vent or, if applicable, theBTEX limit determined in §63.765(b)(1)(iii) for the smallglycol dehydration unit process vent. Only modifications inglycol dehydration unit operations directly related toprocess changes, including but not limited to changes inglycol circulation rate or glycol-HAP absorbency, shall beallowed. Changes in the inlet gas characteristics ornatural gas throughput rate shall not be considered indetermining the overall emission reduction due to processmodifications. (3) The owner or operator that achieves a 95.0 percentHAP emission reduction or meets the BTEX limit determinedin §63.765(b)(1)(iii), as applicable, using process This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

452. Page 452 of 604modifications alone shall comply with paragraph (e)(3)(i)of this section. The owner or operator that achieves a 95.0percent HAP emission reduction or meets the BTEX limitdetermined in §63.765(b)(1)(iii), as applicable, using acombination of process modifications and one or morecontrol devices shall comply with paragraphs (e)(3)(i) and(e)(3)(ii) of this section.* * * * * (ii) The owner or operator shall comply with thecontrol device requirements specified in paragraph (d) or(f) of this section, as applicable, except that theemission reduction or limit achieved shall be the emissionreduction or limit specified for the control device(s) inparagraph (e)(2) of this section. (f) Control device requirements for small glycoldehydration units. (1) The control device used to meet BTEX the emissionlimit calculated in §63.765(b)(1)(iii) shall be one of thecontrol devices specified in paragraphs (f)(1)(i) through(iii) of this section. (i) An enclosed combustion device (e.g., thermal vaporincinerator, catalytic vapor incinerator, boiler, or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

453. Page 453 of 604process heater) that is designed and operated to reduce themass content of BTEX in the gases vented to the device asdetermined in accordance with the requirements of§63.772(e). If a boiler or process heater is used as thecontrol device, then the vent stream shall be introducedinto the flame zone of the boiler or process heater; or (ii) A vapor recovery device (e.g., carbon adsorptionsystem or condenser) or other non-destructive control devicethat is designed and operated to reduce the mass content ofBTEX in the gases vented to the device as determined inaccordance with the requirements of §63.772(e); or (iii) A flare, as defined in §63.761, that is designedand operated in accordance with the requirements of§63.11(b). (2) The owner or operator shall operate each controldevice in accordance with the requirements specified inparagraphs (f)(2)(i) and (ii) of this section. (i) Each control device used to comply with thissubpart shall be operating at all times. An owner oroperator may vent more than one unit to a control deviceused to comply with this subpart. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

454. Page 454 of 604 (ii) For each control device monitored in accordancewith the requirements of §63.773(d), the owner or operatorshall demonstrate compliance according to the requirementsof either §63.772(f) or (h). (3) For each carbon adsorption system used as acontrol device to meet the requirements of paragraph(f)(1)(ii) of this section, the owner or operator shallmanage the carbon as required under (d)(5)(i) and (ii) ofthis section. 18. Section 63.772 is amended by: a. Revising paragraph (b) introductory text; b. Revising paragraph (b)(1)(ii); c. Revising paragraph (b)(2); d. Revising paragraphs (b)(2)(i) through (ii); e. Adding paragraph (d); f. Revising paragraph (e) introductory text; g. Revising paragraphs (e)(1)(i) through (v); h. Revising paragraph (e)(2); i. Revising paragraph (e)(3) introductory text; j. Revising paragraph (e)(3)(i)(B); k. Revising paragraph (e)(3)(iv)(C)(1); l. Adding paragraphs (e)(3)(v); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

455. Page 455 of 604 m. Revising paragraph (e)(4) introductory text; n. Revising paragraph (e)(4)(i); o. Revising paragraph (e)(5); p. Revising paragraph (f) introductory text; q. Adding paragraphs (f)(2) through (f)(6); r. Revising paragraph (g) introductory text; s. Revising paragraph (g)(1) and paragraph (g)(2)introductory text; t. Revising paragraph (g)(2)(iii); u. Revising paragraph (g)(3); v. Adding paragraph (h); and w. Adding paragraph (i) to read as follows:§63.772 Test methods, compliance procedures, and compliancedemonstrations.* * * * * (b) Determination of glycol dehydration unit flowrate,benzene emissions, or BTEX emissions. The procedures ofthis paragraph shall be used by an owner or operator todetermine glycol dehydration unit natural gas flowrate,benzene emissions, or BTEX emissions. (1) * * ** * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

456. Page 456 of 604 (ii) The owner or operator shall document, to theAdministrators satisfaction, the actual annual averagenatural gas flowrate to the glycol dehydration unit.* * * * * (2) The determination of actual average benzene orBTEX emissions from a glycol dehydration unit shall be madeusing the procedures of either paragraph (b)(2)(i) or(b)(2)(ii) of this section. Emissions shall be determinedeither uncontrolled, or with federally enforceable controlsin place. (i) The owner or operator shall determine actualaverage benzene or BTEX emissions using the model GRI-GLYCalcTM , Version 3.0 or higher, and the procedurespresented in the associated GRI-GLYCalcTM TechnicalReference Manual. Inputs to the model shall berepresentative of actual operating conditions of the glycoldehydration unit and may be determined using the proceduresdocumented in the Gas Research Institute (GRI) reportentitled “Atmospheric Rich/Lean Method for DeterminingGlycol Dehydrator Emissions” (GRI–95/0368.1); or (ii) The owner or operator shall determine an averagemass rate of benzene or BTEX emissions in kilograms per This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

457. Page 457 of 604hour through direct measurement using the methods in§63.772(a)(1)(i) or (ii), or an alternative methodaccording to §63.7(f). Annual emissions in kilograms peryear shall be determined by multiplying the mass rate bythe number of hours the unit is operated per year. Thisresult shall be converted to megagrams per year.* * * * * (d) Test procedures and compliance demonstrations forsmall glycol dehydration units. This paragraph applies tothe test procedures for small dehydration units. (1) If the owner or operator is using a controldevice to comply with the emission limit in§63.765(b)(1)(iii), the requirements of paragraph (e) ofthis section apply. Compliance is demonstrated using themethods specified in paragraph (f) of this section. (2) If no control device is used to comply with theemission limit in §63.765(b)(1)(iii), the owner or operatormust determine the glycol dehydration unit BTEX emissionsas specified in paragraphs (d)(2)(i) through (iii) of thissection. Compliance is demonstrated if the BTEX emissionsdetermined as specified in paragraphs (d)(2)(i) through This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

458. Page 458 of 604(iii) are less than the emission limit calculated using theequation in §63.765(b)(1)(iii). (i) Method 1 or 1A, 40 CFR part 60, appendix A, asappropriate, shall be used for selection of the samplingsites at the outlet of the glycol dehydration unit processvent. Any references to particulate mentioned in Methods 1and 1A do not apply to this section. (ii) The gas volumetric flowrate shall be determinedusing Method 2, 2A, 2C, or 2D, 40 CFR part 60, appendix A,as appropriate. (iii) The BTEX emissions from the outlet of theglycol dehydration unit process vent shall be determinedusing the procedures specified in paragraph (e)(3)(v) ofthis section. As an alternative, the mass rate of BTEX atthe outlet of the glycol dehydration unit process vent maybe calculated using the model GRI-GLYCalcTM , Version 3.0 orhigher, and the procedures presented in the associated GRI-GLYCalcTM Technical Reference Manual. Inputs to the modelshall be representative of actual operating conditions ofthe glycol dehydration unit and shall be determined usingthe procedures documented in the Gas Research Institute(GRI) report entitled “Atmospheric Rich/Lean Method for This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

459. Page 459 of 604Determining Glycol Dehydrator Emissions” (GRI–95/0368.1).When the BTEX mass rate is calculated for glycoldehydration units using the model GRI-GLYCalcTM , all BTEXmeasured by Method 18, 40 CFR part 60, appendix A, shall besummed. (e) Control device performance test procedures. Thisparagraph applies to the performance testing of controldevices. The owners or operators shall demonstrate that acontrol device achieves the performance requirements of§63.771(d)(1), (e)(3)(ii) or (f)(1) using a performancetest as specified in paragraph (e)(3) of this section.Owners or operators using a condenser have the option touse a design analysis as specified in paragraph (e)(4) ofthis section. The owner or operator may elect to use thealternative procedures in paragraph (e)(5) of this sectionfor performance testing of a condenser used to controlemissions from a glycol dehydration unit process vent. Asan alternative to conducting a performance test under thissection for combustion control devices, a control devicethat can be demonstrated to meet the performancerequirements of §63.771(d)(1), (e)(3)(ii) or (f)(1) through This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

460. Page 460 of 604a performance test conducted by the manufacturer, asspecified in paragraph (h) of this section can be used. (1) * * * (i) Except as specified in paragraph (e)(2) of thissection, a flare, as defined in §63.761, that is designedand operated in accordance with §63.11(b); (ii) Except for control devices used for small glycoldehydration units, a boiler or process heater with a designheat input capacity of 44 megawatts or greater; (iii) Except for control devices used for small glycoldehydration units, a boiler or process heater into whichthe vent stream is introduced with the primary fuel or isused as the primary fuel; (iv) Except for control devices used for small glycoldehydration units, a boiler or process heater burninghazardous waste for which the owner or operator has eitherbeen issued a final permit under 40 CFR part 270 andcomplies with the requirements of 40 CFR part 266, subpartH; or has certified compliance with the interim statusrequirements of 40 CFR part 266, subpart H; (v) Except for control devices used for small glycoldehydration units, a hazardous waste incinerator for which This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

461. Page 461 of 604the owner or operator has been issued a final permit under40 CFR part 270 and complies with the requirements of 40CFR part 264, subpart O; or has certified compliance withthe interim status requirements of 40 CFR part 265, subpartO.* * * * * (2) An owner or operator shall design and operate eachflare, as defined in §63.761, in accordance with therequirements specified in §63.11(b) and the compliancedetermination shall be conducted using Method 22 of 40 CFRpart 60, appendix A, to determine visible emissions. (3) For a performance test conducted to demonstratethat a control device meets the requirements of§63.771(d)(1), (e)(3)(ii) or (f)(1), the owner or operatorshall use the test methods and procedures specified inparagraphs (e)(3)(i) through (v) of this section. Theinitial and periodic performance tests shall be conductedaccording to the schedule specified in paragraph (e)(3)(vi)of this section. (i) * * ** * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

462. Page 462 of 604 (B) To determine compliance with the enclosedcombustion device total HAP concentration limit specifiedin §63.765(b)(1)(iii), or the BTEX emission limit specifiedin §63.771(f)(1) the sampling site shall be located at theoutlet of the combustion device.* * * * * (iv) * * ** * * * * (C) * * * (1) The emission rate correction factor for excessair, integrated sampling and analysis procedures of Method3A or 3B, 40 CFR part 60, appendix A, shall be used todetermine the oxygen concentration. The samples shall betaken during the same time that the samples are taken fordetermining TOC concentration or total HAP concentration.* * * * * (v) To determine compliance with the BTEX emissionlimit specified in §63.771(f)(1) the owner or operatorshall use one of the following methods: Method 18, 40 CFRpart 60, appendix A; ASTM D6420–99 (2004), as specified in§63.772(a)(1)(ii); or any other method or data that havebeen validated according to the applicable procedures in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

463. Page 463 of 60 04Meth hod 301, 40 CFR p part 63, appendix A. The followin x ng cedures shall be used to calculat BTEX eproc s te emissions s: (A) The minimum samplin time f e m ng for each run shal be 1 llhour in whic either an inte r ch r egrated s sample or a minim r mum offour grab sa r amples sh hall be t taken. If grab sa f ampling i used, isthen the sam n mples sha all be ta aken at a approxima ately equ ualinte ervals in time, s n such as 1 15-minute interva e als durin the ngrun. . (B) The mass ra e ate of BT TEX (Eo) shall be compute using e edthe equation and pr ns rocedures specifi s ied in pa aragraphs s(e)( (3)(v)(B) )(1) and (2) of t this sect tion. (1) The followi e ing equat tion shal be use ll ed:Wher re:Eo= Mass rat of BTE at the outlet of the c te EX e control d device,dry basis, kilogram per hour k r.Coj= Concentration of sample compone ent j of the gas streamat tthe outle of the control device, dry bas et e l , sis, part per tsmilllion by volume. vMoj= Molecular weight of sam mple comp ponent j of the ggasstreeam at th outlet of the control device, gram/gra he t am-mole.Qo= Flowrate of gas stream a the ou e at utlet of the cont troldeviice, dry standard cubic m d meter per minute. r . −6K2= Constant 2.494× t, ×10 (parrts per m million) (gram-moole perstanndard cub bic meter (kilog r) gram/gram (minut m) te/hour), where ,stanndard tem mperature (gram-m e mole per standard cubic m d meter)is 2 degree C. 20 es Th his document is a prepuublication n version n, signed d by EP PA Ad dministrat tor, Lisa P. Jackso on 07/2 on 28/2011. We have t taken step ps to ensure the accura o t acy of thi version but it is not th officia is n, he al ve ersion.

464. Page 464 of 604n = Number of components in sample. (2) When the BTEX mass rate is calculated, only BTEXcompounds measured by Method 18, 40 CFR part 60, appendixA, or ASTM D6420–99 (2004) as specified in§63.772(a)(1)(ii), shall be summed using the equations inparagraph (e)(3)(v)(B)(1) of this section. (vi) The owner or operator shall conduct performancetests according to the schedule specified in paragraphs(e)(3)(vi)(A) and (B) of this section. (A) An initial performance test shall be conductedwithin 180 days after the compliance date that is specifiedfor each affected source in §63.760(f)(7) through (8),except that the initial performance test for existingcombustion control devices at existing major sources shallbe conducted no later than 3 years after the date ofpublication of the final rule in the Federal Register. Ifthe owner or operator of an existing combustion controldevice at an existing major source chooses to replace suchdevice with a control device whose model is tested under§63.772(h), then the newly installed device shall complywith all provisions of this subpart no later than 3 yearsafter the date of publication of the final rule in the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

465. Page 465 of 604Federal Register. The performance test results shall besubmitted in the Notification of Compliance Status Reportas required in §63.775(d)(1)(ii). (B) Periodic performance tests shall be conducted forall control devices required to conduct initial performancetests except as specified in paragraphs (e)(3)(vi)(B)(1)and (2) of this section. The first periodic performancetest shall be conducted no later than 60 months after theinitial performance test required in paragraph(e)(3)(vi)(A) of this section. Subsequent periodicperformance tests shall be conducted at intervals no longerthan 60 months following the previous periodic performancetest or whenever a source desires to establish a newoperating limit. The periodic performance test results mustbe submitted in the next Periodic Report as specified in§63.775(e)(2)(xi). Combustion control devices meeting thecriteria in either paragraph (e)(3)(vi)(B)(1) or (2) ofthis section are not required to conduct periodicperformance tests. (1) A control device whose model is tested under, andmeets the criteria of, §63.772(h), or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

466. Page 466 of 604 (2) A combustion control device tested under§63.772(e) that meets the outlet TOC or HAP performancelevel specified in §63.771(d)(1)(i)(B) and that establishesa correlation between firebox or combustion chambertemperature and the TOC or HAP performance level. (4) For a condenser design analysis conducted to meetthe requirements of §63.771(d)(1), (e)(3)(ii), or (f)(1),the owner or operator shall meet the requirements specifiedin paragraphs (e)(4)(i) and (e)(4)(ii) of this section.Documentation of the design analysis shall be submitted asa part of the Notification of Compliance Status Report asrequired in §63.775(d)(1)(i). (i) The condenser design analysis shall include ananalysis of the vent stream composition, constituentconcentrations, flowrate, relative humidity, andtemperature, and shall establish the design outlet organiccompound concentration level, design average temperature ofthe condenser exhaust vent stream, and the design averagetemperatures of the coolant fluid at the condenser inletand outlet. As an alternative to the condenser designanalysis, an owner or operator may elect to use theprocedures specified in paragraph (e)(5) of this section. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

467. Page 467 of 604* * * * * (5) As an alternative to the procedures in paragraph(e)(4)(i) of this section, an owner or operator may electto use the procedures documented in the GRI reportentitled, “Atmospheric Rich/Lean Method for DeterminingGlycol Dehydrator Emissions” (GRI–95/0368.1) as inputs forthe model GRI-GLYCalcTM , Version 3.0 or higher, to generatea condenser performance curve. (f) Compliance demonstration for control deviceperformance requirements. This paragraph applies to thedemonstration of compliance with the control deviceperformance requirements specified in §63.771(d)(1)(i),(e)(3) and (f)(1). Compliance shall be demonstrated usingthe requirements in paragraphs (f)(1) through (3) of thissection. As an alternative, an owner or operator thatinstalls a condenser as the control device to achieve therequirements specified in §63.771(d)(1)(ii), (e)(3) or(f)(1) may demonstrate compliance according to paragraph(g) of this section. An owner or operator may switchbetween compliance with paragraph (f) of this section andcompliance with paragraph (g) of this section only after atleast 1 year of operation in compliance with the selected This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

468. Page 468 of 604approach. Notification of such a change in the compliancemethod shall be reported in the next Periodic Report, asrequired in §63.775(e), following the change.* * * * * (2) The owner or operator shall calculate the dailyaverage of the applicable monitored parameter in accordancewith §63.773(d)(4) except that the inlet gas flow rate tothe control device shall not be averaged. (3) Compliance with the operating parameter limit isachieved when the daily average of the monitoring parametervalue calculated under paragraph (f)(2) of this section iseither equal to or greater than the minimum or equal to orless than the maximum monitoring value established underparagraph (f)(1) of this section. For inlet gas flow rate,compliance with the operating parameter limit is achievedwhen the value is equal to or less than the valueestablished under §63.772(h). (4) Except for periods of monitoring systemmalfunctions, repairs associated with monitoring systemmalfunctions, and required monitoring system qualityassurance or quality control activities (including, asapplicable, system accuracy audits and required zero and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

469. Page 469 of 604span adjustments), the CMS required in §63.773(d) must beoperated at all times the affected source is operating. Amonitoring system malfunction is any sudden, infrequent,not reasonably preventable failure of the monitoring systemto provide valid data. Monitoring system failures that arecaused in part by poor maintenance or careless operationare not malfunctions. Monitoring system repairs arerequired to be completed in response to monitoring systemmalfunctions and to return the monitoring system tooperation as expeditiously as practicable. (5) Data recorded during monitoring systemmalfunctions, repairs associated with monitoring systemmalfunctions, or required monitoring system qualityassurance or control activities may not be used incalculations used to report emissions or operating levels.All the data collected during all other required datacollection periods must be used in assessing the operationof the control device and associated control system. (6) Except for periods of monitoring systemmalfunctions, repairs associated with monitoring systemmalfunctions, and required quality monitoring systemquality assurance or quality control activities (including, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

470. Page 470 of 604as applicable, system accuracy audits and required zero andspan adjustments), failure to collect required data is adeviation of the monitoring requirements. (g) Compliance demonstration with percent reduction oremission limit performance requirements—condensers. Thisparagraph applies to the demonstration of compliance withthe performance requirements specified in§63.771(d)(1)(ii),(e)(3) or (f)(1) for condensers.Compliance shall be demonstrated using the procedures inparagraphs (g)(1) through (3) of this section. (1) The owner or operator shall establish a site-specific condenser performance curve according to§63.773(d)(5)(ii). For sources required to meet the BTEXlimit in accordance with §63.771(e) or (f)(1) the owner oroperator shall identify the minimum percent reductionnecessary to meet the BTEX limit. (2) Compliance with the requirements in§63.771(d)(1)(ii),(e)(3) or (f)(1) shall be demonstrated bythe procedures in paragraphs (g)(2)(i) through (iii) ofthis section.* * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

471. Page 471 of 604 (iii) Except as provided in paragraphs (g)(2)(iii)(A)and (B) of this section, at the end of each operating day,the owner or operator shall calculate the 365-day averageHAP, or BTEX, emission reduction, as appropriate, from thecondenser efficiencies as determined in paragraph(g)(2)(ii) of this section for the preceding 365 operatingdays. If the owner or operator uses a combination ofprocess modifications and a condenser in accordance withthe requirements of §63.771(e), the 365-day average HAP, orBTEX, emission reduction shall be calculated using theemission reduction achieved through process modificationsand the condenser efficiency as determined in paragraph(g)(2)(ii) of this section, both for the previous 365operating days. (A) After the compliance dates specified in§63.760(f), an owner or operator with less than 120 days ofdata for determining average HAP, or BTEX, emissionreduction, as appropriate, shall calculate the average HAP,or BTEX emission reduction, as appropriate, for the first120 days of operation after the compliance dates. Forsources required to meet the overall 95.0 percent reductionrequirement, compliance is achieved if the 120-day average This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

472. Page 472 of 604HAP emission reduction is equal to or greater than 90.0percent. For sources required to meet the BTEX limit under§63.765(b)(1)(iii), compliance is achieved if the averageBTEX emission reduction is at least 95.0 percent of therequired 365-day value identified under paragraph (g)(1) ofthis section (i.e., at least 76.0 percent if the 365-daydesign value is 80.0 percent). (B) After 120 days and no more than 364 days ofoperation after the compliance dates specified in§63.760(f), the owner or operator shall calculate theaverage HAP emission reduction as the HAP emissionreduction averaged over the number of days between thecurrent day and the applicable compliance date. For sourcesrequired to meet the overall 95.0-percent reductionrequirement, compliance with the performance requirementsis achieved if the average HAP emission reduction is equalto or greater than 90.0 percent. For sources required tomeet the BTEX limit under §63.765(b)(1)(iii), compliance isachieved if the average BTEX emission reduction is at least95.0 percent of the required 365-day value identified underparagraph (g)(1) of this section (i.e., at least 76.0percent if the 365-day design value is 80.0 percent). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

473. Page 473 of 604 (3) If the owner or operator has data for 365 days ormore of operation, compliance is achieved based on theapplicable criteria in paragraphs (g)(3)(i) or (ii) of thissection. (i) For sources meeting the HAP emission reductionspecified in §63.771(d)(1)(ii) or (e)(3) the average HAPemission reduction calculated in paragraph (g)(2)(iii) ofthis section is equal to or greater than 95.0 percent. (ii) For sources required to meet the BTEX limit under§63.771(e)(3) or (f)(1), compliance is achieved if theaverage BTEX emission reduction calculated in paragraph(g)(2)(iii) of this section is equal to or greater than theminimum percent reduction identified in paragraph (g)(1) ofthis section.* * * * *(h) Performance testing for combustion control devices -manufacturers’ performance test. (1) This paragraph applies to the performance testingof a combustion control device conducted by the devicemanufacturer. The manufacturer shall demonstrate that aspecific model of control device achieves the performancerequirements in (h)(7) of this section by conducting a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

474. Page 474 of 604performance test as specified in paragraphs (h)(2) through(6) of this section. (2) Performance testing shall consist of three one-hour (or longer) test runs for each of the four followingfiring rate settings making a total of 12 test runs pertest. Propene (propylene) gas shall be used for the testingfuel. All fuel analyses shall be performed by anindependent third-party laboratory (not affiliated with thecontrol device manufacturer or fuel supplier). (i) 90 - 100 percent of maximum design rate (fixedrate). (ii) 70 - 100 - 70 percent (ramp up, ramp down). Beginthe test at 70 percent of the maximum design rate. Withinthe first 5 minutes, ramp the firing rate to 100 percent ofthe maximum design rate. Hold at 100 percent for 5 minutes.In the 10-15 minute time range, ramp back down to 70percent of the maximum design rate. Repeat three more timesfor a total of 60 minutes of sampling. (iii) 30 - 70 - 30 percent (ramp up, ramp down). Beginthe test at 30 percent of the maximum design rate. Withinthe first 5 minutes, ramp the firing rate to 70 percent ofthe maximum design rate. Hold at 70 percent for 5 minutes. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

475. Page 475 of 604In the 10-15 minute time range, ramp back down to 30percent of the maximum design rate. Repeat three more timesfor a total of 60 minutes of sampling. (iv) 0 - 30 - 0 percent (ramp up, ramp down). Beginthe test at 0 percent of the maximum design rate. Withinthe first 5 minutes, ramp the firing rate to 100 percent ofthe maximum design rate. Hold at 30 percent for 5 minutes.In the 10-15 minute time range, ramp back down to 0 percentof the maximum design rate. Repeat three more times for atotal of 60 minutes of sampling. (3) All models employing multiple enclosures shall betested simultaneously and with all burners operational.Results shall be reported for the each enclosureindividually and for the average of the emissions from allinterconnected combustion enclosures/chambers. Controldevice operating data shall be collected continuouslythroughout the performance test using an electronic DataAcquisition System and strip chart. Data shall be submittedwith the test report in accordance with paragraph (8)(iii)of this section. (4) Inlet testing shall be conducted as specified inparagraphs (4)(i) through (iii) of this section. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

476. Page 476 of 604 (i) The fuel flow metering system shall be located inaccordance with Method 2A, 40 CFR part 60, appendix A-1,(or other approved procedure) to measure fuel flow rate atthe control device inlet location. The fitting for fillingfuel sample containers shall be located a minimum of 8 pipediameters upstream of any inlet fuel flow monitoring meter. (ii) Inlet flow rate shall be determined using Method2A, 40 CFR part 60, appendix A-1. Record the start and stopreading for each 60-minute THC test. Record the gaspressure and temperature at 5-minute intervals throughouteach 60-minute THC test. (iii) Inlet fuel sampling shall be conducted inaccordance with the criteria in paragraphs (h)(4)(iii)(A)and (B) of this section. (A) At the inlet fuel sampling location, securelyconnect a Silonite-coated stainless steel evacuatedcanister fitted with a flow controller sufficient to fillthe canister over a 1 hour period. Filling shall beconducted as specified in the following: (1) Open the canister sampling valve at the beginningof the total hydrocarbon (THC) test, and close the canisterat the end of the THC test. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

477. Page 477 of 604 (2) Fill one canister for each THC test run. (3) Label the canisters individually and record on achain of custody form. (B) Each fuel sample shall be analyzed using thefollowing methods. The results shall be included in thetest report. (1) Hydrocarbon compounds containing between one andfive atoms of carbon plus benzene using ASTM D1945-03. (2) Hydrogen (H2), carbon monoxide (CO), carbon dioxide(CO2), nitrogen (N2), oxygen (O2) using ASTM D1945-03. (3) Carbonyl sulfide, carbon disulfide plus mercaptansusing ASTM D5504. (4) Higher heating value using ASTM D3588-98 or ASTMD4891-89. (5) Outlet testing shall be conducted in accordancewith the criteria in paragraphs (h)(5)(i) through (v) ofthis section. (i) Sampling and flowrate measured in accordance withthe following: (A) The outlet sampling location shall be a minimum of4 equivalent stack diameters downstream from the highestpeak flame or any other flow disturbance, and a minimum of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

478. Page 478 of 604one equivalent stack diameter upstream of the exit or anyother flow disturbance. A minimum of two sample ports shallbe used. (B) Flow rate shall be measured using Method 1, 40 CFRpart 60, Appendix 1, for determining flow measurementtraverse point location; and Method 2, 40 CFR part 60,Appendix 1, shall be used to measure duct velocity. If lowflow conditions are encountered (i.e., velocity pressuredifferentials less than 0.05 inches of water) during theperformance test, a more sensitive manometer shall be usedto obtain an accurate flow profile. (ii) Molecular weight shall be determined as specifiedin paragraphs (h)(4)(iii)(B), (5)(ii)(A) and (B) of thissection. (A) An integrated bag sample shall be collected duringthe Method 4, 40 CFR part 60, Appendix A, moisture test.Analyze the bag sample using a gas chromatograph-thermalconductivity detector (GC-TCD) analysis meeting thefollowing criteria: (1) Collect the integrated sample throughout theentire test, and collect representative volumes from eachtraverse location. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

479. Page 479 of 604 (2) The sampling line shall be purged with stack gasbefore opening the valve and beginning to fill the bag. (3) The bag contents shall be kneaded orotherwise vigorously mixed prior to the GC analysis. (4) The GC-TCD calibration procedure in Method 3C, 40CFR part 60, Appendix A, shall be modified by using EPAAlt-045 as follows: For the initial calibration, triplicateinjections of any single concentration must agree within 5percent of their mean to be valid. The calibration responsefactor for a single concentration re-check must be within10 percent of the original calibration response factor forthat concentration. If this criterion is not met, theinitial calibration using at least three concentrationlevels shall be repeated. (B) Report the molecular weight of: O2, CO2, methane(CH4), and N2 and include in the test report submitted under§63.775(d)(iii). Moisture shall be determined using Method4, 40 CFR part 60, Appendix A. Traverse both ports with theMethod 4, 40 CFR part 60, Appendix A, sampling train duringeach test run. Ambient air shall not be introduced into theMethod 3C, 40 CFR part 60, Appendix A, integrated bagsample during the port change. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

480. Page 480 of 604 (iv) Carbon monoxide shall be determined using Method10, 40 CFR part 60, Appendix A. The test shall be run atthe same time and with the sample points used for the EPAMethod 25A, 40 CFR part 60, Appendix A, testing. Aninstrument range of 0-10 per million by volume-dry (ppmvd)shall be used. (v) Visible emissions shall be determined using Method22, 40 CFR part 60, Appendix A. The test shall be performedcontinuously during each test run. A digital colorphotograph of the exhaust point, taken from the position ofthe observer and annotated with date and time, will betaken once per test run and the four photos included in thetest report. (6) Total hydrocarbons (THC) shall be determined asspecified by the following criteria: (i) Conduct THC sampling using Method 25A, 40 CFR part60, Appendix A, except the option for locating the probe inthe center 10 percent of the stack shall not be allowed.The THC probe must be traversed to 16.7 percent, 50percent, and 83.3 percent of the stack diameter during thetesting. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

481. Page 481 of 604 (ii) A valid test shall consist of three Method 25A,40 CFR part 60, Appendix A, tests, each no less than 60minutes in duration. (iii) A 0-10 parts per million by volume-wet (ppmvw)(as propane) measurement range is preferred; as analternative a 0-30 ppmvw (as carbon) measurement range maybe used. (iv) Calibration gases will be propane in air and becertified through EPA Protocol 1 – “EPA TraceabilityProtocol for Assay and Certification of Gaseous CalibrationStandards,” September 1997, as amended August 25, 1999,EPA–600/R–97/121 (or more recent if updated since 1999). (v) THC measurements shall be reported in terms ofppmvw as propane. (vi) THC results shall be corrected to 3 percent CO2,as measured by Method 3C, 40 CFR part 60, Appendix A. (vii) Subtraction of methane/ethane from the THC datais not allowed in determining results. (7) Performance test criteria: (i) The control device model tested must meet thecriteria in paragraphs (h)(7)(i)(A) through (C) of thissection: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

482. Page 482 of 604 (A) Method 22, 40 CFR part 60, Appendix A, resultsunder paragraph (h)(5)(v) of this section with noindication of visible emissions, and (B) Average Method 25A, 40 CFR part 60, Appendix A,results under paragraph (h)(6) of this section equal to orless than 10.0 ppmvw THC as propane corrected to 3.0percent CO2, and (C) Average CO emissions determined under paragraph(h)(5)(iv) of this section equal to or less than 10 partsppmvd, corrected to 3.0 percent CO2. (ii) The manufacturer shall determine a maximum inletgas flow rate which shall not be exceeded for each controldevice model to achieve the criteria in paragraph (h)(7)(i)of this section. (iii) A control device meeting the criteria inparagraph (h)(7)(i)(A) through (C) of this section willhave demonstrated a destruction efficiency of 98.0 percentfor HAP regulated under this subpart. (8) The owner or operator of a combustion controldevice model tested under this section shall submit theinformation listed in paragraphs (h)(8)(i) through (iii) inthe test report required under §63.775(d)(1)(iii). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

483. Page 483 of 604 (i) Full schematic of the control device anddimensions of the device components. (ii) Design net heating value (minimum and maximum) ofthe device. (iii) Test fuel gas flow range (in both mass andvolume). Include the minimum and maximum allowable inletgas flow rate. (iv) Air/stream injection/assist ranges, if used. (v) The test parameter ranges listed in paragraphs(h)(8)(vi)(A) through (O) of this section, as applicablefor the tested model. (A) Fuel gas delivery pressure and temperature. (B) Fuel gas moisture range. (C) Purge gas usage range. (D) Condensate (liquid fuel) separation range. (E) Combustion zone temperature range. This isrequired for all devices that measure this parameter. (F) Excess combustion air range. (G) Flame arrestor(s). (H) Burner manifold pressure. (I) Pilot flame sensor. (J) Pilot flame design fuel and fuel usage. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

484. Page 484 of 604 (K) Tip velocity range. (L) Momentum flux ratio. (M) Exit temperature range. (N) Exit flow rate. (O) Wind velocity and direction. (vi) The test report shall include all calibrationquality assurance/quality control data, calibration gasvalues, gas cylinder certification, and strip chartsannotated with test times and calibration values. (i) Compliance demonstration for combustion controldevices - manufacturers’ performance test. This paragraphapplies to the demonstration of compliance for a combustioncontrol device tested under the provisions in paragraph (h)of this section. Owners or operators shall demonstrate thata control device achieves the performance requirements of§63.771(d)(1), (e)(3)(ii) or (f)(1), by installing a devicetested under paragraph (h) of this section and complyingwith the following criteria: (1) The inlet gas flow rate shall meet the rangespecified by the manufacturer. Flow rate shall be measuredas specified in §63.773(d)(3)(i)(H)(1). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

485. Page 485 of 604 (2) A pilot flame shall be present at all times ofoperation. The pilot flame shall be monitored in accordancewith §63.773(d)(3)(i)(H)(2). (3) Devices shall be operated with no visibleemissions, except for periods not to exceed a total of 5minutes during any 2 consecutive hours. A visible emissionstest using Method 22, 40 CFR part 60, Appendix A, shall beperformed monthly. The observation period shall be 2 hoursand shall be used according to Method 22. (4) Compliance with the operating parameter limit isachieved when the following criteria are met: (i) The inlet gas flow rate monitored under paragraph(i)(1) of this section is equal to or below the maximumestablished by the manufacturer; and (ii) The pilot flame is present at all times; and (iii) During the visible emissions test performedunder paragraph (i)(3) of this section the duration ofvisible emissions does not exceed a total of 5 minutesduring the observation period. Devices failing the visibleemissions test shall follow the requirements in paragraphs(i)(4)(iii)(A) and (B) of this section. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

486. Page 486 of 604 (A) Following the first failure, the fuel nozzle(s)and burner tubes shall be replaced. (B) If, following replacement of the fuel nozzle(s)and burner tubes as specified in paragraph (i)(4)(iii)(A),the visible emissions test is not passed in the nextscheduled test, either a performance test shall beperformed under paragraph (e) of this section, or thedevice shall be replaced with another control device whosemodel was tested, and meets, the requirements in paragraph(h) of this section. 19. Section 63.773 is amended by: a. Revising paragraph (b); b. Revising paragraph (d)(1) introductory text; c. Revising paragraph (d)(1)(ii) and adding paragraphs(d)(1)(iii) and (iv); d. Revising paragraphs (d)(2)(i) and (d)(2)(ii); e. Revising paragraphs (d)(3)(i)(A) and (B); f. Revising paragraphs (d)(3)(i)(D) and (E); g. Revising paragraphs (d)(3)(i)(F)(1) and (2); h. Revising paragraph (d)(3)(i)(G); i. Adding paragraph (d)(3)(i)(H); j. Revising paragraph (d)(4); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

487. Page 487 of 604 k. Revising paragraph (d)(5)(i) introductory text; l. Revising paragraphs (d)(5)(i)(A) and (B); m. Adding paragraph (d)(5)(i)(C); n. Revising paragraphs (d)(5)(ii)(A) through (C); o. Revising paragraphs (d)(6)(ii) and (iii); p. Adding paragraph (d)(6)(vi); q. Revising paragraph (d)(8)(i)(A); and r. Revising paragraph (d)(8)(ii) to read as follows:§63.773 Inspection and monitoring requirements.* * * * * (b) The owner or operator of a control device whosemodel was tested under 63.772(h) shall develop aninspection and maintenance plan for each control device.At a minimum, the plan shall contain the control devicemanufacturer’s recommendations for ensuring properoperation of the device. Semi-annual inspections shall beconducted for each control device with maintenance andreplacement of control device components made in accordancewith the plan. (d) Control device monitoring requirements. (1) Foreach control device, except as provided for in paragraph(d)(2) of this section, the owner or operator shall install This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

488. Page 488 of 604and operate a continuous parameter monitoring system inaccordance with the requirements of paragraphs (d)(3)through (9) of this section. Owners or operators thatinstall and operate a flare in accordance with§63.771(d)(1)(iii) or (f)(1)(iii) are exempt from therequirements of paragraphs (d)(4) and (5) of this section.The continuous monitoring system shall be designed andoperated so that a determination can be made on whether thecontrol device is achieving the applicable performancerequirements of §63.771(d), (e)(3) or (f)(1). Eachcontinuous parameter monitoring system shall meet thefollowing specifications and requirements:* * * * * (ii) A site-specific monitoring plan must be preparedthat addresses the monitoring system design, datacollection, and the quality assurance and quality controlelements outlined in paragraph (d) of this section and in§63.8(d). Each CPMS must be installed, calibrated,operated, and maintained in accordance with the proceduresin your approved site-specific monitoring plan. Using theprocess described in §63.8(f)(4), you may request approvalof monitoring system quality assurance and quality control This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

489. Page 489 of 604procedures alternative to those specified in paragraphs(d)(1)(ii)(A) through (E) of this section in your site-specific monitoring plan. (A) The performance criteria and design specificationsfor the monitoring system equipment, including the sampleinterface, detector signal analyzer, and data acquisitionand calculations; (B) Sampling interface (e.g., thermocouple) locationsuch that the monitoring system will provide representativemeasurements; (C) Equipment performance checks, system accuracyaudits, or other audit procedures; (D) Ongoing operation and maintenance procedures inaccordance with provisions in §63.8(c)(1) and (c)(3); and (E) Ongoing reporting and recordkeeping procedures inaccordance with provisions in §63.10(c), (e)(1), and(e)(2)(i). (iii) The owner or operator must conduct the CPMSequipment performance checks, system accuracy audits, orother audit procedures specified in the site-specificmonitoring plan at least once every 12 months. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

490. Page 490 of 604 (iv) The owner or operator must conduct a performanceevaluation of each CPMS in accordance with the site-specific monitoring plan. (2) * * * (i) Except for control devices for small glycoldehydration units, a boiler or process heater in which allvent streams are introduced with the primary fuel or isused as the primary fuel; or (ii) Except for control devices for small glycoldehydration units, a boiler or process heater with a designheat input capacity equal to or greater than 44 megawatts. (3) * * * (i) * * * (A) For a thermal vapor incinerator that demonstratesduring the performance test conducted under §63.772(e) thatthe combustion zone temperature is an accurate indicator ofperformance, a temperature monitoring device equipped witha continuous recorder. The monitoring device shall have aminimum accuracy of ±1 percent of the temperature beingmonitored in degrees C, or ±2.5 degrees C, whichever valueis greater. The temperature sensor shall be installed at alocation representative of the combustion zone temperature. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

491. Page 491 of 604 (B) For a catalytic vapor incinerator, a temperaturemonitoring device equipped with a continuous recorder. Thedevice shall be capable of monitoring temperature at twolocations and have a minimum accuracy of ±1 percent of thetemperature being monitored in degrees C, or ±2.5 degreesC, whichever value is greater. One temperature sensor shallbe installed in the vent stream at the nearest feasiblepoint to the catalyst bed inlet and a second temperaturesensor shall be installed in the vent stream at the nearestfeasible point to the catalyst bed outlet.* * * * * (D) For a boiler or process heater a temperaturemonitoring device equipped with a continuous recorder. Thetemperature monitoring device shall have a minimum accuracyof ±1 percent of the temperature being monitored in degreesC, or ±2.5 degrees C, whichever value is greater. Thetemperature sensor shall be installed at a locationrepresentative of the combustion zone temperature. (E) For a condenser, a temperature monitoring deviceequipped with a continuous recorder. The temperaturemonitoring device shall have a minimum accuracy of ±1percent of the temperature being monitored in degrees C, or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

492. Page 492 of 604±2.8 degrees C, whichever value is greater. The temperaturesensor shall be installed at a location in the exhaust ventstream from the condenser. (F) * * * (1) A continuous parameter monitoring system tomeasure and record the average total regeneration streammass flow or volumetric flow during each carbon bedregeneration cycle. The flow sensor must have a measurementsensitivity of 5 percent of the flow rate or 10 cubic feetper minute, whichever is greater. The mechanicalconnections for leakage must be checked at least everymonth, and a visual inspection must be performed at leastevery 3 months of all components of the flow CPMS forphysical and operational integrity and all electricalconnections for oxidation and galvanic corrosion if yourflow CPMS is not equipped with a redundant flow sensor; and (2) A continuous parameter monitoring system tomeasure and record the average carbon bed temperature forthe duration of the carbon bed steaming cycle and tomeasure the actual carbon bed temperature afterregeneration and within 15 minutes of completing thecooling cycle. The temperature monitoring device shall have This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

493. Page 493 of 604a minimum accuracy of ±1 percent of the temperature beingmonitored in degrees C, or ±2.5 degrees C, whichever valueis greater. (G) For a nonregenerative-type carbon adsorptionsystem, the owner or operator shall monitor the designcarbon replacement interval established using a performancetest performed in accordance with §63.772(e)(3) shall bebased on the total carbon working capacity of the controldevice and source operating schedule. (H) For a control device model whose model is testedunder §63.772(h): (1) A continuous monitoring system that measures gasflow rate at the inlet to the control device. Themonitoring instrument shall have an accuracy of plus orminus 2 percent or better. (2) A heat sensing monitoring device equipped with acontinuous recorder that indicates the continuous ignitionof the pilot flame.* * * * * (4) Using the data recorded by the monitoring system,except for inlet gas flow rate, the owner or operator mustcalculate the daily average value for each monitored This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

494. Page 494 of 604operating parameter for each operating day. If theemissions unit operation is continuous, the operating dayis a 24-hour period. If the emissions unit operation is notcontinuous, the operating day is the total number of hoursof control device operation per 24-hour period. Valid datapoints must be available for 75 percent of the operatinghours in an operating day to compute the daily average. (5) For each operating parameter monitor installed inaccordance with the requirements of paragraph (d)(3) ofthis section, the owner or operator shall comply withparagraph (d)(5)(i) of this section for all controldevices, and when condensers are installed, the owner oroperator shall also comply with paragraph (d)(5)(ii) ofthis section. (i) The owner or operator shall establish a minimumoperating parameter value or a maximum operating parametervalue, as appropriate for the control device, to define theconditions at which the control device must be operated tocontinuously achieve the applicable performancerequirements of §63.771(d)(1), (e)(3)(ii) or (f)(1). Eachminimum or maximum operating parameter value shall beestablished as follows: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

495. Page 495 of 604 (A) If the owner or operator conducts performancetests in accordance with the requirements of §63.772(e)(3)to demonstrate that the control device achieves theapplicable performance requirements specified in§63.771(d)(1), (e)(3)(ii) or (f)(1), then the minimumoperating parameter value or the maximum operatingparameter value shall be established based on valuesmeasured during the performance test and supplemented, asnecessary, by a condenser design analysis or control devicemanufacturer recommendations or a combination of both. (B) If the owner or operator uses a condenser designanalysis in accordance with the requirements of§63.772(e)(4) to demonstrate that the control deviceachieves the applicable performance requirements specifiedin §63.771(d)(1), (e)(3)(ii) or (f)(1), then the minimumoperating parameter value or the maximum operatingparameter value shall be established based on the condenserdesign analysis and may be supplemented by the condensermanufacturers recommendations. (C) If the owner or operator operates a control devicewhere the performance test requirement was met under§63.772(h) to demonstrate that the control device achieves This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

496. Page 496 of 604the applicable performance requirements specified in§63.771(d)(1), (e)(3)(ii) or (f)(1), then the maximum inletgas flow rate shall be established based on the performancetest and supplemented, as necessary, by the manufacturerrecommendations. (ii) * * * (A) If the owner or operator conducts a performancetest in accordance with the requirements of §63.772(e)(3)to demonstrate that the condenser achieves the applicableperformance requirements in §63.771(d)(1), (e)(3)(ii) or(f)(1), then the condenser performance curve shall be basedon values measured during the performance test andsupplemented as necessary by control device designanalysis, or control device manufacturers recommendations,or a combination or both. (B) If the owner or operator uses a control devicedesign analysis in accordance with the requirements of§63.772(e)(4)(i) to demonstrate that the condenser achievesthe applicable performance requirements specified in§63.771(d)(1), (e)(3)(ii) or (f)(1), then the condenserperformance curve shall be based on the condenser design This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

497. Page 497 of 604analysis and may be supplemented by the control devicemanufacturers recommendations. (C) As an alternative to paragraph (d)(5)(ii)(B) ofthis section, the owner or operator may elect to use theprocedures documented in the GRI report entitled,“Atmospheric Rich/Lean Method for Determining GlycolDehydrator Emissions” (GRI–95/0368.1) as inputs for themodel GRI-GLYCalcTM , Version 3.0 or higher, to generate acondenser performance curve.* * * * * (6) * * ** * * * * (ii) For sources meeting §63.771(d)(1)(ii), anexcursion occurs when the 365-day average condenserefficiency calculated according to the requirementsspecified in §63.772(g)(2)(iii) is less than 95.0 percent.For sources meeting §63.771(f)(1), an excursion occurs whenthe 365-day average condenser efficiency calculatedaccording to the requirements specified in§63.772(g)(2)(iii) is less than 95.0 percent of theidentified 365-day required percent reduction. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

498. Page 498 of 604 (iii) For sources meeting §63.771(d)(1)(ii), if anowner or operator has less than 365 days of data, anexcursion occurs when the average condenser efficiencycalculated according to the procedures specified in§63.772(g)(2)(iii)(A) or (B) is less than 90.0 percent. Forsources meeting §63.771(d)(1)(ii), an excursion occurs whenthe 365-day average condenser efficiency calculatedaccording to the requirements specified in§63.772(g)(2)(iii) is less than the identified 365-dayrequired percent reduction.* * * * * (vi) For control device whose model is tested under§63.772(h) an excursion occurs when: (A) The inlet gas flow rate exceeds the maximumestablished during the test conducted under §63.772(h). (B) Failure of the monthly visible emissions testconducted under §63.772(i)(3) occurs.* * * * * (8) * * * (i) * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

499. Page 499 of 604 (A) During a malfunction when the affected facility isoperated during such period in accordance with §63.6(e)(1);or* * * * * (ii) For each control device, or combinations ofcontrol devices installed on the same emissions unit, oneexcused excursion is allowed per semiannual period for anyreason. The initial semiannual period is the 6-monthreporting period addressed by the first Periodic Reportsubmitted by the owner or operator in accordance with§63.775(e) of this subpart.* * * * * 20. Section 63.774 is amended by: a. Revising paragraph (b)(3) introductory text; b. Deleting and reserving paragraph (b)(3)(ii); c. Revising paragraph (b)(4)(ii) introductory text; d. Adding paragraph (b)(4)(ii)(C); e. Adding paragraph (b)(7)(ix); and f. Adding paragraphs (g) and (h) to read as follows:§63.774 Recordkeeping requirements.* * * * * (b) * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

500. Page 500 of 604* * * * * (3) Records specified in §63.10(c) for each monitoringsystem operated by the owner or operator in accordance withthe requirements of §63.773(d). Notwithstanding therequirements of §63.10(c), monitoring data recorded duringperiods identified in paragraphs (b)(3)(i) through(b)(3)(iv) of this section shall not be included in anyaverage or percent leak rate computed under this subpart.Records shall be kept of the times and durations of allsuch periods and any other periods during process orcontrol device operation when monitors are not operating orfailed to collect required data.* * * * * (ii) [Reserved]* * * * * (4) * * ** * * * * (ii) Records of the daily average value of eachcontinuously monitored parameter for each operating daydetermined according to the procedures specified in§63.773(d)(4) of this subpart, except as specified inparagraphs (b)(4)(ii)(A) through (C) of this section. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

501. Page 501 of 604* * * * * (C) For control device whose model is tested under§63.772(h), the records required in paragraph (h) of thissection.* * * * * (7) * * * (ix) Records identifying the carbon replacementschedule under §63.771(d)(5) and records of each carbonreplacement.* * * * * (g) The owner or operator of an affected sourcesubject to this subpart shall maintain records of theoccurrence and duration of each malfunction of operation(i.e., process equipment) or the air pollution controlequipment and monitoring equipment. The owner or operatorshall maintain records of actions taken during periods ofmalfunction to minimize emissions in accordance with§63.764(a), including corrective actions to restoremalfunctioning process and air pollution control andmonitoring equipment to its normal or usual manner ofoperation. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

502. Page 502 of 604 (h) Record the following when using a control devicewhose model is tested under §63.772(h) to comply with§63.771(d), (e)(3)(ii) and (f)(1): (1) All visible emission readings and flowratemeasurements made during the compliance determinationrequired by §63.772(i); and (2) All hourly records and other recorded periods whenthe pilot flame is absent. (i) The date the semi-annual maintenance inspectionrequired under §63.773(b) is performed. Include a list ofany modifications or repairs made to the control deviceduring the inspection and other maintenance performed suchas cleaning of the fuel nozzles. 21. Section 63.775 is amended by: a. Revising paragraph (b)(1); b. Revising paragraph (b)(6); c. Deleting and reserving paragraph (b)(7); d. Revising paragraph (c)(1); e. Revising paragraph (c)(6); f. Revising paragraph (c)(7)(i); g. Revising paragraph (d)(1) introductory text; h. Revising paragraph (d)(1)(i); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

503. Page 503 of 604 i. Revising paragraph (d)(1)(ii) introductory text; j. Revising paragraph (d)(2)(ii); k. Adding paragraph (d)(5)(iv); l. Revising paragraph (d)(11); m. Adding paragraphs (d)(13) and (d)(14); n. Revising paragraphs (e)(2)(ii)(B) and (C); o. Adding paragraphs (e)(2)(ii)(E) and (F); p. Adding paragraphs (e)(2)(xi) and (xii); and q. Adding paragraph (g) to read as follows:§63.775 Reporting requirements.* * * * * (b) * * * (1) The initial notifications required for existingaffected sources under §63.9(b)(2) shall be submitted asprovided in paragraphs (b)(1)(i) and (ii) of this section. (i) Except as otherwise provided in paragraph (ii),the initial notifications shall be submitted by 1 yearafter an affected source becomes subject to the provisionsof this subpart or by June 17, 2000, whichever is later.Affected sources that are major sources on or before June17, 2000 and plan to be area sources by June 17, 2002 shallinclude in this notification a brief, nonbinding This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

504. Page 504 of 604description of a schedule for the action(s) that areplanned to achieve area source status. (ii) An affected source identified under §63.760(f)(7)or (9) shall submit an initial notification required forexisting affected sources under §63.9(b)(2) within 1 yearafter the affected source becomes subject to the provisionsof this subpart or by one year after publication of thefinal rule in the Federal Register, whichever is later. Anaffected source identified under §63.760(f)(7) or (9) thatplans to be an area source by three years after publicationof the final rule in the Federal Register, shall include inthis notification a brief, nonbinding description of aschedule for the action(s) that are planned to achieve areasource status.* * * * * (6) If there was a malfunction during the reportingperiod, the Periodic Report specified in paragraph (e) ofthis section shall include the number, duration, and abrief description for each type of malfunction whichoccurred during the reporting period and which caused ormay have caused any applicable emission limitation to beexceeded. The report must also include a description of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

505. Page 505 of 604actions taken by an owner or operator during a malfunctionof an affected source to minimize emissions in accordancewith §63.764(j), including actions taken to correct amalfunction. (7) [Reserved]* * * * * (c) * * * (1) The initial notifications required under§63.9(b)(2) not later than January 3, 2008. In addition tosubmitting your initial notification to the addresseesspecified under §63.9(a), you must also submit a copy ofthe initial notification to the EPAs Office of Air QualityPlanning and Standards. Send your notification via e-mailto Oil and Gas Sector@epa.gov or via U.S. mail or othermail delivery service to U.S. EPA, Sector Policies andPrograms Division/Fuels and Incineration Group (E143–01),Attn: Oil and Gas Project Leader, Research Triangle Park,NC 27711.* * * * * (6) If there was a malfunction during the reportingperiod, the Periodic Report specified in paragraph (e) ofthis section shall include the number, duration, and a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

506. Page 506 of 604brief description for each type of malfunction whichoccurred during the reporting period and which caused ormay have caused any applicable emission limitation to beexceeded. The report must also include a description ofactions taken by an owner or operator during a malfunctionof an affected source to minimize emissions in accordancewith §63.764(j), including actions taken to correct amalfunction. (7) * * * (i) Documentation of the sources location relative tothe nearest UA plus offset and UC boundaries. Thisinformation shall include the latitude and longitude of theaffected source; whether the source is located in an urbancluster with 10,000 people or more; the distance in milesto the nearest urbanized area boundary if the source is notlocated in an urban cluster with 10,000 people or more; andthe name of the nearest urban cluster with 10,000 people ormore and nearest urbanized area.* * * * * (d) * * * (1) If a closed-vent system and a control device otherthan a flare are used to comply with §63.764, the owner or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

507. Page 507 of 604operator shall submit the information in paragraph(d)(1)(iii) of this section and the information in eitherparagraph (d)(1)(i) or (ii) of this section. (i) The condenser design analysis documentationspecified in §63.772(e)(4) of this subpart, if the owner oroperator elects to prepare a design analysis. (ii) If the owner or operator is required to conduct aperformance test, the performance test results includingthe information specified in paragraphs (d)(1)(ii)(A) and(B) of this section. Results of a performance testconducted prior to the compliance date of this subpart canbe used provided that the test was conducted using themethods specified in §63.772(e)(3) and that the testconditions are representative of current operatingconditions. If the owner or operator operates a combustioncontrol device model tested under §63.772(h), an electroniccopy of the performance test results shall be submitted viaemail to Oil and Gas PT@EPA.GOV.* * * * * (iii) (2) * * ** * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

508. Page 508 of 604 (ii) A statement of whether a flame, was present atthe pilot light over the full period of the compliancedetermination. (5) * * ** * * * * (ii) An explanation of the rationale for why the owneror operator selected each of the operating parameter valuesestablished in §63.773(d)(5). This explanation shallinclude any data and calculations used to develop the valueand a description of why the chosen value indicates thatthe control device is operating in accordance with theapplicable requirements of §63.771(d)(1), (e)(3)(ii) or(f)(1).* * * * * (iv) For each carbon adsorber, the predeterminedcarbon replacement schedule as required in§63.771(d)(5)(i).* * * * * (11) The owner or operator shall submit the analysisprepared under §63.771(e)(2) to demonstrate the conditionsby which the facility will be operated to achieve the HAPemission reduction of 95.0 percent, or the BTEX limit in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

509. Page 509 of 604§63.765(b)(1)(iii), through process modifications or acombination of process modifications and one or morecontrol devices.* * * * * (13) If the owner or operator installs a combustioncontrol device model tested under the procedures in§63.772(h), the data listed under §63.772(h)(8). (14) For each combustion control device model testedunder §63.772(h), the information listed in paragraphs(d)(14)(i) through (vi) of this section. (i) Name, address and telephone number of the controldevice manufacturer. (ii) Control device model number. (iii) Control device serial number. (iv) Date of control device certification test. (v) Manufacturer’s HAP destruction efficiency rating. (vi) Control device operating parameters, maximumallowable inlet gas flowrate. (e) * * ** * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

510. Page 510 of 604 (2) The owner or operator shall include theinformation specified in paragraphs (e)(2)(i) through(xiii) of this section, as applicable.* * * * * (ii) * * ** * * * * (B) For each excursion caused when the 365-day averagecondenser control efficiency is less than the valuespecified in §63.773(d)(6)(ii), the report must include the365-day average values of the condenser control efficiency,and the date and duration of the period that the excursionoccurred. (C) For each excursion caused when condenser controlefficiency is less than the value specified in§63.773(d)(6)(iii), the report must include the averagevalues of the condenser control efficiency, and the dateand duration of the period that the excursion occurred.* * * * * (E) For each excursion caused when the maximum inletgas flow rate identified under §63.772(h) is exceeded, thereport must include the values of the inlet gas identified This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

511. Page 511 of 604and the date and duration of the period that the excursionoccurred. (F) For each excursion caused when visible emissionsdetermined under §63.772(i) exceed the maximum allowableduration, the report must include the date and duration ofthe period that the excursion occurred.* * * * * (xi) The results of any periodic test as required in§63.772(e)(3) conducted during the reporting period. (xii) For each carbon adsorber used to meet thecontrol device requirements of §63.771(d)(1), records ofeach carbon replacement that occurred during the reportingperiod. (xiii) For combustion control device inspectionsconducted in accordance with §63.773(b) the recordsspecified in §63.774(i).* * * * * (g) Electronic Reporting. (1) As of January 1, 2012and within 60 days after the date of completing eachperformance test, as defined in §63.2 and as required inthis subpart, you must submit performance test data, exceptopacity data, electronically to the EPA’s Central Data This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

512. Page 512 of 604Exchange (CDX) by using the Electronic Reporting Tool (ERT)(see http://www.epa.gov/ttn/chief/ert/ert tool.html/). Onlydata collected using test methods compatible with ERT aresubject to this requirement to be submitted electronicallyinto the EPA’s WebFIRE database. (2) All reports required by this subpart not subjectto the requirements in paragraphs (g)(1) of this sectionmust be sent to the Administrator at the appropriateaddress listed in §63.13. If acceptable to both theAdministrator and the owner or operator of a source, thesereports may be submitted on electronic media. TheAdministrator retains the right to require submittal ofreports subject to paragraph (g)(1) of this section inpaper format.* * * * * 22. Appendix to subpart HH of part 63 – Tables isamended by revising Table 2 to read as follows:Appendix to Subpart HH of Part 63—Tables* * * * * Table 2 to Subpart HH of Part 63—Applicability of 40 CFR Part 63 General Provisions to Subpart HH Applicabl General provisions e to reference subpart H Explanation This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

513. Page 513 of 604 H§63.1(a)(1) ..........Yes.§63.1(a)(2) ..........Yes.§63.1(a)(3) ..........Yes.§63.1(a)(4) ..........Yes.§63.1(a)(5) ..........No ........Section reserved.§63.1(a)(6) ..........Yes.§63.1(a)(7) through ..No ........Section reserved.(a)(9)§63.1(a)(10) .........Yes.§63.1(a)(11) .........Yes.§63.1(a)(12) .........Yes.§63.1(b)(1) ..........No ........Subpart HH specifies applicability.§63.1(b)(2) ..........No ........Section reserved.§63.1(b)(3) ..........Yes.§63.1(c)(1) ..........No ........Subpart HH specifies applicability.§63.1(c)(2) ..........Yes. ......Subpart HH exempts area sources from the requirement to obtain a Title V permit unless otherwise required by law as specified in §63.760(h).§63.1(c)(3) and No ........Section reserved.(c)(4) ...............§63.1(c)(5) ..........Yes.§63.1(d) .............No ........Section reserved.§63.1(e) .............Yes.§63.2 ................Yes .......Except definition of major source is unique for this source category and there are additional definitions in subpart HH.§63.3(a) through (c) .Yes.§63.4(a)(1) through ..Yes.(a)(2)§63.4(a)(3) through ..No ........Section reserved.(a)(5)§63.4(b) .............Yes.§63.4(c) .............Yes.§63.5(a)(1) ..........Yes.§63.5(a)(2) ..........Yes.§63.5(b)(1) ..........Yes. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

514. Page 514 of 604§63.5(b)(2) ..........No ........Section reserved.§63.5(b)(3) ..........Yes.§63.5(b)(4) ..........Yes.§63.5(b)(5) ..........No ........Section Reserved.§63.5(b)(6) ..........Yes.§63.5(c) .............No ........Section reserved.§63.5(d)(1) ..........Yes.§63.5(d)(2) ..........Yes.§63.5(d)(3) ..........Yes.§63.5(d)(4) ..........Yes.§63.5(e) .............Yes.§63.5(f)(1) ..........Yes.§63.5(f)(2) ..........Yes.§63.6(a) .............Yes.§63.6(b)(1) ..........Yes.§63.6(b)(2) ..........Yes.§63.6(b)(3) ..........Yes.§63.6(b)(4) ..........Yes.§63.6(b)(5) ..........Yes.§63.6(b)(6) ..........No ........Section reserved.§63.6(b)(7) ..........Yes.§63.6(c)(1) ..........Yes.§63.6(c)(2) ..........Yes.§63.6(c)(3) through ..No ........Section reserved.(c)(4)§63.6(c)(5) ..........Yes.§63.6(d) .............No ........Section reserved.§63.6(e) .............Yes.§63.6(e)(1)(i) .......No ........See §63.764(j) for general duty requirement.§63.6(e)(1)(ii) ......No.§63.6(e)(1)(iii) .....Yes.§63.6(e)(2) ..........No ........Section reserved.§63.6(e)(3) ..........No.§63.6(f)(1) ..........No§63.6(f)(2) ..........Yes.§63.6(f)(3) ..........Yes.§63.6(g) .............Yes.§63.6(h) .............No ........Subpart HH does not contain opacity or visible emission standards.§63.6(i)(1) through ..Yes.(i)(14)§63.6(i)(15) .........No ........Section reserved. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

515. Page 515 of 604§63.6(i)(16) .........Yes.§63.6(j) .............Yes.§63.7(a)(1) ..........Yes.§63.7(a)(2) ..........Yes .......But the performance test results must be submitted within 180 days after the compliance date.§63.7(a)(3) ..........Yes.§63.7(b) .............Yes.§63.7(c) .............Yes.§63.7(d) .............Yes.§63.7(e)(1) ..........No.§63.7(e)(2) ..........Yes.§63.7(e)(3) ..........Yes.§63.7(e)(4) ..........Yes.§63.7(f) .............Yes.§63.7(g) .............Yes.§63.7(h) .............Yes.§63.8(a)(1) ..........Yes.§63.8(a)(2) ..........Yes.§63.8(a)(3) ..........No ........Section reserved.§63.8(a)(4) ..........Yes.§63.8(b)(1) ..........Yes.§63.8(b)(2) ..........Yes.§63.8(b)(3) ..........Yes.§63.8(c)(1) ..........No.§63.8(c)(1)(i) No.§63.8(c)(1)(ii) Yes.§63.8(c)(1)(iii) .....Pending§63.8(c)(2) ..........Yes.§63.8(c)(3) ..........Yes.§63.8(c)(4) ..........Yes.§63.8(c)(4)(i) .......No ........Subpart HH does not require continuous opacity monitors.§63.8(c)(4)(ii) ......Yes.§63.8(c)(5) through ..Yes.(c)(8)§63.8(d) .............Yes.§63.8(d)(3) ..........Yes. Except for last sentence, which refers to an SSM plan. SSM plans are not required.§63.8(e) .............Yes Subpart HH does not specifically require continuous emissions monitor This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

516. Page 516 of 604 performance evaluation, however, the Administrator can request that one be conducted.§63.8(f)(1) through ..Yes.(f)(5)§63.8(f)(6) ..........Yes.§63.8(g) .............No ........Subpart HH specifies continuous monitoring system data reduction requirements.§63.9(a) .............Yes.§63.9(b)(1) ..........Yes.§63.9(b)(2) ..........Yes .......Existing sources are given 1 year (rather than 120 days) to submit this notification. Major and area sources that meet §63.764(e) do not have to submit initial notifications.§63.9(b)(3) ..........No ........Section reserved.§63.9(b)(4) ..........Yes.§63.9(b)(5) ..........Yes.§63.9(c) .............Yes.§63.9(d) .............Yes.§63.9(e) .............Yes.§63.9(f) .............No ........Subpart HH does not have opacity or visible emission standards.§63.9(g)(1) ..........Yes.§63.9(g)(2) ..........No ........Subpart HH does not have opacity or visible emission standards.§63.9(g)(3) ..........Yes.§63.9(h)(1) through ..Yes .......Area sources located outside(h)(3) UA plus offset and UC boundaries are not required to submit notifications of compliance status.§63.9(h)(4) ..........No ........Section reserved.§63.9(h)(5) through ..Yes.(h)(6)§63.9(i) .............Yes.§63.9(j) .............Yes.§63.10(a) ............Yes. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

517. Page 517 of 604§63.10(b)(1) .........Yes .......§63.774(b)(1) requires sources to maintain the most recent 12 months of data on- site and allows offsite storage for the remaining 4 years of data.§63.10(b)(2) .........Yes.§63.10(b)(2)(i) ......No.§63.10(b)(2)(ii) .....No ........See §63.774(g) for recordkeeping of occurrence, duration, and actions taken during malfunctions.§63.10(b)(2)(iii) ....Yes.§63.10(b)(2)(iv) .....No.through (b)(2)(v)§63.10(b)(2)(vi) .....Yes.through (b)(2)(xiv)§63.10(b)(3) .........Yes .......§63.774(b)(1) requires sources to maintain the most recent 12 months of data on- site and allows offsite storage for the remaining 4 years of data.§63.10(c)(1) .........Yes.§63.10(c)(2) through No ........Sections reserved.(c)(4)§63.10(c)(5) through Yes.(8) (c)(8)§63.10(c)(9) .........No ........Section reserved.§63.10(c)(10) through No ........See §63.774(g) for(11) recordkeeping of malfunctions.§63.10(c)(12) through Yes.(14)§63.10(c)(15) ........No.§63.10(d)(1) .........Yes.§63.10(d)(2) .........Yes .......Area sources located outside UA plus offset and UC boundaries do not have to submit performance test reports.§63.10(d)(3) .........Yes.§63.10(d)(4) .........Yes.§63.10(d)(5) .........No ........See §63.775(b)(6) or (c)(6) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

518. Page 518 of 604 for reporting of malfunctions.§63.10(e)(1) .........Yes .......Area sources located outside UA plus offset and UC boundaries are not required to submit reports.§63.10(e)(2) .........Yes .......Area sources located outside UA plus offset and UC boundaries are not required to submit reports.§63.10(e)(3)(i) ......Yes .......Subpart HH requires major sources to submit Periodic Reports semi-annually. Area sources are required to submit Periodic Reports annually. Area sources located outside UA plus offset and UC boundaries are not required to submit reports.§63.10(e)(3)(i)(A) ...Yes.§63.10(e)(3)(i)(B) ...Yes.§63.10(e)(3)(i)(C) ...No ........Section reserved.§63.10(e)(3)(ii) .....Yes.through (viii)§63.10(f) ............Yes.§63.11(a) and (b) ....Yes.§63.11(c), (d), and Yes.(e)§63.12(a) through (c) Yes.§63.13(a) through (c) Yes.§63.14(a) and (b) ....Yes.§63.15(a) and (b) ....Yes.§63.16 ...............Yes. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

519. Page 519 of 604Subpart HHH–-[AMENDED] 23. Section 63.1270 is amended by: a. Revising paragraph (a) introductory text; b. Revising paragraph (a)(4); c. Revising paragraphs (d)(1) and (d)(2); and d. Adding paragraphs (d)(3),(4) and (5) to read asfollows:§63.1270 Applicability and designation of affected source. (a) This subpart applies to owners and operators ofnatural gas transmission and storage facilities thattransport or store natural gas prior to entering thepipeline to a local distribution company or to a final enduser (if there is no local distribution company), and thatare major sources of hazardous air pollutants (HAP)emissions as defined in §63.1271. Emissions for majorsource determination purposes can be estimated using themaximum natural gas throughput calculated in eitherparagraph (a)(1) or (2) of this section and paragraphs(a)(3) and (4) of this section. As an alternative tocalculating the maximum natural gas throughput, the owneror operator of a new or existing source may use thefacility design maximum natural gas throughput to estimate This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

520. Page 520 of 604the maximum potential emissions. Other means to determinethe facilitys major source status are allowed, providedthe information is documented and recorded to theAdministrators satisfaction in accordance with§63.10(b)(3). A compressor station that transports naturalgas prior to the point of custody transfer or to a naturalgas processing plant (if present) is not considered a partof the natural gas transmission and storage sourcecategory. A facility that is determined to be an areasource, but subsequently increases its emissions or itspotential to emit above the major source levels (withoutobtaining and complying with other limitations that keepits potential to emit HAP below major source levels), andbecomes a major source, must comply thereafter with allapplicable provisions of this subpart starting on theapplicable compliance date specified in paragraph (d) ofthis section. Nothing in this paragraph is intended topreclude a source from limiting its potential to emitthrough other appropriate mechanisms that may be availablethrough the permitting authority.* * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

521. Page 521 of 604 (4) The owner or operator shall determine the maximumvalues for other parameters used to calculate potentialemissions as the maximum over the same period for whichmaximum throughput is determined as specified in paragraph(a)(1) or (a)(2) of this section. These parameters shall bebased on an annual average or the highest single measuredvalue. For estimating maximum potential emissions fromglycol dehydration units, the glycol circulation rate usedin the calculation shall be the unit’s maximum rate underits physical and operational design consistent with thedefinition of potential to emit in §63.2.* * * * * (d) * * * (1) Except as specified in paragraphs (d)(3) through(5) of this section, the owner or operator of an affectedsource, the construction or reconstruction of whichcommenced before February 6, 1998, shall achieve compliancewith this provisions of the subpart no later than June 17,2002 except as provided for in §63.6(i). The owner oroperator of an area source, the construction orreconstruction of which commenced before February 6, 1998,that increases its emissions of (or its potential to emit) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

522. Page 522 of 604HAP such that the source becomes a major source that issubject to this subpart shall comply with this subpart 3years after becoming a major source. (2) Except as specified in paragraphs (d)(3) through(5) of this section, the owner or operator of an affectedsource, the construction or reconstruction of whichcommences on or after February 6, 1998, shall achievecompliance with the provisions of this subpart immediatelyupon initial startup or June 17, 1999, whichever date islater. Area sources, the construction or reconstruction ofwhich commences on or after February 6, 1998, that becomemajor sources shall comply with the provisions of thisstandard immediately upon becoming a major source. (3) Each affected small glycol dehydration unit, asdefined in §63.1271, located at a major source, thatcommenced construction before [INSERT DATE OF PUBLICATIONIN THE FEDERAL REGISTER] must achieve compliance no laterthan 3 years after the date of publication of the finalrule in the Federal Register, except as provided in§63.6(i). (4) Each affected small glycol dehydration unit, asdefined in §63.1271, located at a major source, that This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

523. Page 523 of 604commenced construction on or after [INSERT DATE OFPUBLICATION IN THE FEDERAL REGISTER] must achievecompliance immediately upon initial startup or the date ofpublication of the final rule in the Federal Register,whichever is later. (5) Each large glycol dehydration unit, as defined in§63.1271, that has complied with the provisions of thissubpart prior to [INSERT DATE OF PUBLICATION IN THE FEDERALREGISTER] by reducing its benzene emissions to less than0.9 megagrams per year must achieve compliance no laterthan 90 days after the date of publication of the finalrule in the Federal Register, except as provided in§63.6(i).* * * * * 24. Section 63.1271 is amended by: a. Adding, in alphabetical order, new definitions forthe terms “affirmative defense,” “BTEX,” “flare,” “largeglycol dehydration units,” “small glycol dehydrationunits”; and b. Revising, in alphabetical order, the definition for“glycol dehydration unit baseline operations” and“temperature monitoring device” to read as follows: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

524. Page 524 of 604§63.1271 Definitions.* * * * * Affirmative defense means, in the context of anenforcement proceeding, a response or defense put forwardby a defendant, regarding which the defendant has theburden of proof, and the merits of which are independentlyand objectively evaluated in a judicial or administrativeproceeding.* * * * * BTEX means benzene, toluene, ethyl benzene and xylene.* * * * * Flare means a thermal oxidation system using an openflame (i.e., without enclosure).* * * * * Glycol dehydration unit baseline operations meansoperations representative of the large glycol dehydrationunit operations as of June 17, 1999 and the small glycoldehydration unit operations as of [INSERT DATE OFPUBLICATION IN THE FEDERAL REGISTER]. For the purposes ofthis subpart, for determining the percentage of overall HAPemission reduction attributable to process modifications,glycol dehydration unit baseline operations shall be This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

525. Page 525 of 604parameter values (including, but not limited to, glycolcirculation rate or glycol-HAP absorbency) that representactual long-term conditions (i.e., at least 1 year). Glycoldehydration units in operation for less than 1 year shalldocument that the parameter values represent expected long-term operating conditions had process modifications notbeen made.* * * * * Large glycol dehydration unit means a glycoldehydration unit with an actual annual average natural gasflowrate equal to or greater than 283.0 thousand standardcubic meters per day and actual annual average benzeneemissions equal to or greater than 0.90 Mg/yr, determinedaccording to §63.1282(a).* * * * * Small glycol dehydration unit means a glycoldehydration unit, located at a major source, with an actualannual average natural gas flowrate less than 283.0thousand standard cubic meters per day or actual annualaverage benzene emissions less than 0.90 Mg/yr, determinedaccording to §63.1282(a). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

526. Page 526 of 604 Temperature monitoring device means an instrument usedto monitor temperature and having a minimum accuracy of ±1percent of the temperature being monitored expressed in °C,or ±2.5 °C, whichever is greater. The temperaturemonitoring device may measure temperature in degreesFahrenheit or degrees Celsius, or both. 25. Section 63.1272 is amended by: a. Revising paragraph (a) through (d); and b. Deleting paragraph (e) to read as follows:§63.1272 Startups and shutdowns. (a) The provisions set forth in this subpart shallapply at all times. (b) The owner or operator shall not shut down items ofequipment that are required or utilized for compliance withthe provisions of this subpart during times when emissionsare being routed to such items of equipment, if theshutdown would contravene requirements of this subpartapplicable to such items of equipment. This paragraph doesnot apply if the owner or operator must shut down theequipment to avoid damage due to a contemporaneous startupor shutdown of the affected source or a portion thereof. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

527. Page 527 of 604 (c) During startups and shutdowns, the owner oroperator shall implement measures to prevent or minimizeexcess emissions to the maximum extent practical. (d) In response to an action to enforce the standardsset forth in this subpart, you may assert an affirmativedefense to a claim for civil penalties for exceedances ofsuch standards that are caused by malfunction, as definedin §63.2. Appropriate penalties may be assessed, however,if you fail to meet your burden of proving all therequirements in the affirmative defense. The affirmativedefense shall not be available for claims for injunctiverelief. (1) To establish the affirmative defense in any actionto enforce such a limit, the owner or operator must timelymeet the notification requirements in paragraph (d)(2) ofthis section, and must prove by a preponderance of evidencethat: (i) The excess emissions: (A) Were caused by a sudden, infrequent, andunavoidable failure of air pollution control and monitoringequipment, process equipment, or a process to operate in anormal or usual manner; and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

528. Page 528 of 604 (B) Could not have been prevented through carefulplanning, proper design or better operation and maintenancepractices; and (C) Did not stem from any activity or event that couldhave been foreseen and avoided, or planned for; and (D) Were not part of a recurring pattern indicative ofinadequate design, operation, or maintenance; and (ii) Repairs were made as expeditiously as possiblewhen the applicable emission limitations were beingexceeded. Off-shift and overtime labor were used, to theextent practicable to make these repairs; and (iii) The frequency, amount and duration of the excessemissions (including any bypass) were minimized to themaximum extent practicable during periods of suchemissions; and (iv) If the excess emissions resulted from a bypass ofcontrol equipment or a process, then the bypass wasunavoidable to prevent loss of life, personal injury, orsevere property damage; and (v) All possible steps were taken to minimize theimpact of the excess emissions on ambient air quality, theenvironment, and human health; and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

529. Page 529 of 604 (vi) All emissions monitoring and control systems werekept in operation if at all possible, consistent withsafety and good air pollution control practices; and (vii) All of the actions in response to the excessemissions were documented by properly signed,contemporaneous operating logs; and (viii) At all times, the affected source was operatedin a manner consistent with good practices for minimizingemissions; and (ix) A written root cause analysis has been preparedto determine, correct, and eliminate the primary causes ofthe malfunction and the excess emissions resulting from themalfunction event at issue. The analysis shall alsospecify, using best monitoring methods and engineeringjudgment, the amount of excess emissions that were theresult of the malfunction. (2) Notification. The owner or operator of theaffected source experiencing an exceedance of its emissionlimit(s) during a malfunction shall notify theAdministrator by telephone or facsimile transmission assoon as possible, but no later than two business days afterthe initial occurrence of the malfunction, if it wishes to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

530. Page 530 of 604avail itself of an affirmative defense to civil penaltiesfor that malfunction. The owner or operator seeking toassert an affirmative defense shall also submit a writtenreport to the Administrator within 45 days of the initialoccurrence of the exceedance of the standard in thissubpart to demonstrate, with all necessary supportingdocumentation, that it has met the requirements set forthin paragraph (d)(1) of this section. The owner or operatormay seek an extension of this deadline for up to 30additional days by submitting a written request to theAdministrator before the expiration of the 45 day period.Until a request for an extension has been approved by theAdministrator, the owner or operator is subject to therequirement to submit such report within 45 days of theinitial occurrence of the exceedance.* * * * * 26. Section 63.1274 is amended by: a. Revising paragraph (c) introductory text; b. Deleting and reserving paragraph (d); c. Revising paragraph (g); and d. Adding paragraph (h) to read as follows: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

531. Page 531 of 604§63.1274 General standards.* * * * * (c) The owner or operator of an affected source (i.e.,glycol dehydration unit) located at an existing or newmajor source of HAP emissions shall comply with therequirements in this subpart as follows:* * * * * (d) [Reserved]* * * * * (g) In all cases where the provisions of this subpartrequire an owner or operator to repair leaks by a specifiedtime after the leak is detected, it is a violation of thisstandard to fail to take action to repair the leak(s)within the specified time. If action is taken to repair theleak(s) within the specified time, failure of that actionto successfully repair the leak(s) is not a violation ofthis standard. However, if the repairs are unsuccessful,and a leak is detected, the owner or operator shall takefurther action as required by the applicable provisions ofthis subpart. (h) At all times the owner or operator must operateand maintain any affected source, including associated air This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

532. Page 532 of 604pollution control equipment and monitoring equipment, in amanner consistent with safety and good air pollutioncontrol practices for minimizing emissions. Determinationof whether such operation and maintenance procedures arebeing used will be based on information available to theAdministrator which may include, but is not limited to,monitoring results, review of operation and maintenanceprocedures, review of operation and maintenance records,and inspection of the source. 27. Section 63.1275 is amended by: a. Revising paragraph (a); b. Revising paragraph (b)(1) introductory text; c. Revising paragraph (b)(1)(i); d. Deleting and reserving paragraph (b)(1)(ii); e. Adding paragraph (b)(1)(iii); f. Revising paragraph (c)(2); g. Revising paragraph (c)(3) introductory text; h. Revising paragraph (c)(1)(i); i. Deleting and reserving paragraph (c)(3)(ii); and j. Adding paragraph (c)(3)(iii) to read as follows: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

533. Page 533 of 604§63.1275 Glycol dehydration unit process vent standards. (a) This section applies to each glycol dehydrationunit subject to this subpart that must be controlled forair emissions as specified in paragraph (c)(1) of §63.1274. (b) * * * (1) For each glycol dehydration unit process vent, theowner or operator shall control air emissions by eitherparagraph (b)(1)(i) or (b)(1)(iii) of this section. (i) The owner or operator of a large glycoldehydration unit, as defined in §63.1271, shall connect theprocess vent to a control device or a combination ofcontrol devices through a closed-vent system. The closed-vent system shall be designed and operated in accordancewith the requirements of §63.1281(c). The control device(s)shall be designed and operated in accordance with therequirements of §63.1281(d). (ii) [Reserved] (iii) You must limit BTEX emissions from each smallglycol dehydration unit, as defined in §63.1271, to thelimit determined in Equation 1 of this section. The limitmust be met in accordance with one of the alternatives This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

534. Page 534 of 60 04spec cified in paragra n aphs (b)( (i)(iii)( (A) throu ugh (D) o this ofsect tion.Equa ation 1Wherre:ELBTE = Unit EX t-specifi BTEX e ic emission limit, m megagrams per syearr;6.422x10-5 = BTEX emis B ssion lim mit, gram BTEX/s ms standard cubicmete -ppmv; er ;Throoughput = Annual average daily naatural ga throug as ghput,stanndard cubbic meter per da rs ayCi,BT = BTEX concent TEX X tration o the na of atural ga at the inlet as eto tthe glyco dehydr ol ration un nit, ppmv v. (A) Con nnect the process vent to a contr e s o rol devic or cecomb bination of contr rol devic ces throu ugh a clo osed-vent tsyst tem. The closed v vent syst tem shall be desi l igned and doper rated in accordan nce with the requ uirements of s§63. .1281(c). The con . ntrol dev vice(s) s shall be designed and doper rated in accordan nce with the requ uirements of s§63. .1281(f). . (B) Mee the em et missions limit th hrough pr rocess ification in accmodi ns cordance with the require e ements sp pecifiedin § §63.1281( (e). (C) Mee the em et mission l limit for each sm r mall glyc coldehy ydration unit usi ing a com mbination of proc n cessmodi ification and on or mor contro device throug the ns ne re ol es gh Th his document is a prepuublication n version n, signed d by EP PA Ad dministrat tor, Lisa P. Jackso on 07/2 on 28/2011. We have t taken step ps to ensure the accura o t acy of thi version but it is not th officia is n, he al ve ersion.

535. Page 535 of 604requirements specified in paragraphs (b)(1)(iii)(A) and (B)of this section. (D) Demonstrate that the emissions limit is metthrough actual uncontrolled operation of the small glycoldehydration unit. Document operational parameters inaccordance with the requirements specified in §63.1281(e)and emissions in accordance with the requirements specifiedin §63.1282(a)(3).* * * * * (c) * * * (2) The owner or operator shall demonstrate, to theAdministrators satisfaction, that the total HAP emissionsto the atmosphere from the large glycol dehydration unitprocess vent are reduced by 95.0 percent through processmodifications or a combination of process modifications andone or more control devices, in accordance with therequirements specified in §63.1281(e). (3) Control of HAP emissions from a GCG separator(flash tank) vent is not required if the owner or operatordemonstrates, to the Administrators satisfaction, thattotal emissions to the atmosphere from the glycoldehydration unit process vent are reduced by one of the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

536. Page 536 of 604levels specified in paragraph (c)(3)(i) or (iii) throughthe installation and operation of controls as specified inparagraph (b)(1) of this section. (i) For any large glycol dehydration unit, HAPemissions are reduced by 95.0 percent or more. (ii) [Reserved] (iii) For each small glycol dehydration unit, BTEXemissions are reduced to a level less than the limitcalculated in paragraph (b)(1)(iii) of this section. 28. Section 63.1281 is amended by: a. Revising paragraph (c)(1); b. Revising paragraph (d)(1) introductory text; c. Revising paragraph (d)(1)(i); d. Revising paragraph (d)(1)(i)(C); e. Revising paragraphs (d)(1)(ii) and (iii); f. Revising paragraph (d)(4)(i); g. Revising paragraph (d)(5)(i); h. Revising paragraph (e)(2); i. Revising paragraph (e)(3) introductory text; j. Revising paragraph (e)(3)(ii); and k. Adding paragraph (f) to read as follows: This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

537. Page 537 of 604§63.1281 Control equipment requirements.* * * * * (c) * * * (1) The closed-vent system shall route all gases,vapors, and fumes emitted from the material in an emissionsunit to a control device that meets the requirementsspecified in paragraph (d) of this section.* * * * * (d) Control device requirements for sources exceptsmall glycol dehydration units. Owners and operators ofsmall glycol dehydration units shall comply with thecontrol requirements in paragraph (f) of this section. (1) * * * (i) An enclosed combustion device (e.g., thermal vaporincinerator, catalytic vapor incinerator, boiler, orprocess heater) that is designed and operated in accordancewith one of the following performance requirements:* * * * * (C) For a control device that can demonstrate auniform combustion zone temperature during the performancetest conducted under §63.1282(d), operates at a minimumtemperature of 760°C. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

538. Page 538 of 604* * * * * (ii) A vapor recovery device (e.g., carbon adsorptionsystem or condenser) or other non-destructive controldevice that is designed and operated to reduce the masscontent of either TOC or total HAP in the gases vented tothe device by 95.0 percent by weight or greater asdetermined in accordance with the requirements of§63.1282(d). (iii) A flare, as defined in §63.1271, that isdesigned and operated in accordance with the requirementsof §63.11(b).* * * * * (4) * * * (i) Each control device used to comply with thissubpart shall be operating at all times when gases, vapors,and fumes are vented from the emissions unit or unitsthrough the closed vent system to the control device asrequired under §63.1275. An owner or operator may vent morethan one unit to a control device used to comply with thissubpart. (5) * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

539. Page 539 of 604 (i) Following the initial startup of the controldevice, all carbon in the control device shall be replacedwith fresh carbon on a regular, predetermined time intervalthat is no longer than the carbon service life establishedfor the carbon adsorption system. Records identifying theschedule for replacement and records of each carbonreplacement shall be maintained as required in§63.1284(b)(7)(ix). The schedule for replacement shall besubmitted with the Notification of Compliance Status Reportas specified in §63.1285(d)(4)(iv). Each carbon replacementmust be reported in the Periodic Reports as specified in§63.1285(e)(2)(xi).* * * * * (e) * * ** * * * * (2) The owner or operator shall document, to theAdministrators satisfaction, the conditions for whichglycol dehydration unit baseline operations shall bemodified to achieve the 95.0 percent overall HAP emissionreduction, or BTEX limit determined in §63.1275(b)(1)(iii),as applicable, either through process modifications orthrough a combination of process modifications and one or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

540. Page 540 of 604more control devices. If a combination of processmodifications and one or more control devices are used, theowner or operator shall also establish the emissionreduction to be achieved by the control device to achievean overall HAP emission reduction of 95.0 percent for theglycol dehydration unit process vent or, if applicable, theBTEX limit determined in §63.1275(b)(1)(iii) for the smallglycol dehydration unit process vent. Only modifications inglycol dehydration unit operations directly related toprocess changes, including but not limited to changes inglycol circulation rate or glycol-HAP absorbency, shall beallowed. Changes in the inlet gas characteristics ornatural gas throughput rate shall not be considered indetermining the overall emission reduction due to processmodifications. (3) The owner or operator that achieves a 95.0 percentHAP emission reduction or meets the BTEX limit determinedin §63.1275(b)(1)(iii), as applicable, using processmodifications alone shall comply with paragraph (e)(3)(i)of this section. The owner or operator that achieves a 95.0percent HAP emission reduction or meets the BTEX limitdetermined in §63.1275(b)(1)(iii), as applicable, using a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

541. Page 541 of 604combination of process modifications and one or morecontrol devices shall comply with paragraphs (e)(3)(i) and(e)(3)(ii) of this section.* * * * * (ii) The owner or operator shall comply with thecontrol device requirements specified in paragraph (d) or(f) of this section, as applicable, except that theemission reduction or limit achieved shall be the emissionreduction or limit specified for the control device(s) inparagraph (e)(2) of this section. (f) Control device requirements for small glycoldehydration units. (1) The control device used to meet BTEX the emissionlimit calculated in §63.1275(b)(1)(iii) shall be one of thecontrol devices specified in paragraphs (f)(1)(i) through(iii) of this section. (i) An enclosed combustion device (e.g., thermal vaporincinerator, catalytic vapor incinerator, boiler, orprocess heater) that is designed and operated to reduce themass content of BTEX in the gases vented to the device asdetermined in accordance with the requirements of§63.1282(d). If a boiler or process heater is used as the This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

542. Page 542 of 604control device, then the vent stream shall be introducedinto the flame zone of the boiler or process heater; or (ii) A vapor recovery device (e.g., carbon adsorptionsystem or condenser) or other non-destructive controldevice that is designed and operated to reduce the masscontent of BTEX in the gases vented to the device asdetermined in accordance with the requirements of§63.1282(d); or (iii) A flare, as defined in §63.1271, that isdesigned and operated in accordance with the requirementsof §63.11(b). (2) The owner or operator shall operate each controldevice in accordance with the requirements specified inparagraphs (f)(2)(i) and (ii) of this section. (i) Each control device used to comply with thissubpart shall be operating at all times. An owner oroperator may vent more than one unit to a control deviceused to comply with this subpart. (ii) For each control device monitored in accordancewith the requirements of §63.1283(d), the owner or operatorshall demonstrate compliance according to the requirementsof either §63.1282(e) or (h). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

543. Page 543 of 604 (3) For each carbon adsorption system used as acontrol device to meet the requirements of paragraph (f)(1)of this section, the owner or operator shall manage thecarbon as required under (d)(5)(i) and (ii) of thissection. 29. Section 63.1282 is amended by: a. Revising paragraph (a) introductory text; b. Revising paragraph (a)(1)(ii); c. Revising paragraph (a)(2) introductory text; d. Revising paragraphs (a)(2)(i) and (ii); e. Adding paragraph (c); f. Revising paragraph (d) introductory text; g. Revising paragraphs (d)(1)(i) through (v); h. Revising paragraph (d)(2); i. Revising paragraph (d)(3) introductory text; j. Revising paragraph (d)(3)(i)(B); k. Revising paragraph (d)(3)(iv)(C)(1) l. Adding paragraphs (d)(3)(v) and (vi); m. Revising paragraph (d)(4) introductory text; n. Revising paragraph (d)(4)(i); o. Revising paragraph (d)(5); p. Revising paragraph (e) introductory text; This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

544. Page 544 of 604 q. Adding paragraphs (e)(2) through (e)(6); r. Revising paragraph (f) introductory text; s. Revising paragraph (f)(1); t. Revising paragraph (f)(2) introductory text; u. Revising paragraph (f)(2)(iii); v. Revising paragraph (f)(3); and w. Adding paragraphs (g) and (h) to read as follows:§63.1282 Test methods, compliance procedures, andcompliance demonstrations. (a) Determination of glycol dehydration unit flowrate,benzene emissions, or BTEX emissions. The procedures ofthis paragraph shall be used by an owner or operator todetermine glycol dehydration unit natural gas flowrate,benzene emissions, or BTEX emissions. (1) * * ** * * * * (ii) The owner or operator shall document, to theAdministrators satisfaction, the actual annual averagenatural gas flowrate to the glycol dehydration unit. (2) The determination of actual average benzene orBTEX emissions from a glycol dehydration unit shall be madeusing the procedures of either paragraph (a)(2)(i) or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

545. Page 545 of 604(a)(2)(ii) of this section. Emissions shall be determinedeither uncontrolled or with federally enforceable controlsin place. (i) The owner or operator shall determine actualaverage benzene or BTEX emissions using the model GRI-GLYCalcTM, Version 3.0 or higher, and the procedurespresented in the associated GRI-GLYCalcTM TechnicalReference Manual. Inputs to the model shall berepresentative of actual operating conditions of the glycoldehydration unit and may be determined using the proceduresdocumented in the Gas Research Institute (GRI) reportentitled “Atmospheric Rich/Lean Method for DeterminingGlycol Dehydrator Emissions” (GRI–95/0368.1); or (ii) The owner or operator shall determine an averagemass rate of benzene or BTEX emissions in kilograms perhour through direct measurement by performing three runs ofMethod 18 in 40 CFR part 60, appendix A (or an equivalentmethod), and averaging the results of the three runs.Annual emissions in kilograms per year shall be determinedby multiplying the mass rate by the number of hours theunit is operated per year. This result shall be convertedto megagrams per year. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

546. Page 546 of 604* * * * * (c) Test procedures and compliance demonstrations forsmall glycol dehydration units. This paragraph applies tothe test procedures for small dehydration units. (1) If the owner or operator is using a controldevice to comply with the emission limit in§63.1275(b)(1)(iii), the requirements of paragraph (d) ofthis section apply. Compliance is demonstrated using themethods specified in paragraph (e) of this section. (2) If no control device is used to comply with theemission limit in §63.1275(b)(1)(iii), the owner oroperator must determine the glycol dehydration unit BTEXemissions as specified in paragraphs (c)(2)(i) through(iii) of this section. Compliance is demonstrated if theBTEX emissions determined as specified in paragraphs(c)(2)(i) through (iii) are less than the emission limitcalculated using the equation in §63.1275(b)(1)(iii). (i) Method 1 or 1A, 40 CFR part 60, appendix A, asappropriate, shall be used for selection of the samplingsites at the outlet of the glycol dehydration unit processvent. Any references to particulate mentioned in Methods 1and 1A do not apply to this section. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

547. Page 547 of 604 (ii) The gas volumetric flowrate shall be determinedusing Method 2, 2A, 2C, or 2D, 40 CFR part 60, appendix A,as appropriate. (iii) The BTEX emissions from the outlet of theglycol dehydration unit process vent shall be determinedusing the procedures specified in paragraph (d)(3)(v) ofthis section. As an alternative, the mass rate of BTEX atthe outlet of the glycol dehydration unit process vent maybe calculated using the model GRI-GLYCalcTM , Version 3.0 orhigher, and the procedures presented in the associated GRI-GLYCalcTM Technical Reference Manual. Inputs to the modelshall be representative of actual operating conditions ofthe glycol dehydration unit and shall be determined usingthe procedures documented in the Gas Research Institute(GRI) report entitled “Atmospheric Rich/Lean Method forDetermining Glycol Dehydrator Emissions” (GRI–95/0368.1).When the BTEX mass rate is calculated for glycoldehydration units using the model GRI-GLYCalcTM, all BTEXmeasured by Method 18, 40 CFR part 60, appendix A, shall besummed. (d) Control device performance test procedures. Thisparagraph applies to the performance testing of control This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

548. Page 548 of 604devices. The owners or operators shall demonstrate that acontrol device achieves the performance requirements of§63.1281(d)(1), (e)(3)(ii), or (f)(1) using a performancetest as specified in paragraph (d)(3) of this section.Owners or operators using a condenser have the option touse a design analysis as specified in paragraph (d)(4) ofthis section. The owner or operator may elect to use thealternative procedures in paragraph (d)(5) of this sectionfor performance testing of a condenser used to controlemissions from a glycol dehydration unit process vent. Asan alternative to conducting a performance test under thissection for combustion control devices, a control devicethat can be demonstrated to meet the performancerequirements of §63.1281(d)(1), (e)(3)(ii), or (f)(1)through a performance test conducted by the manufacturer,as specified in paragraph (g) of this section, can be used. (1) * * * (i) Except as specified in paragraph (d)(2) of thissection, a flare, as defined in §63.1271, that is designedand operated in accordance with §63.11(b); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

549. Page 549 of 604 (ii) Except for control devices used for small glycoldehydration units, a boiler or process heater with a designheat input capacity of 44 megawatts or greater; (iii) Except for control devices used for small glycoldehydration units, a boiler or process heater into whichthe vent stream is introduced with the primary fuel or isused as the primary fuel; (iv) Except for control devices used for small glycoldehydration units, a boiler or process heater burninghazardous waste for which the owner or operator has eitherbeen issued a final permit under 40 CFR part 270 andcomplies with the requirements of 40 CFR part 266, subpartH, or has certified compliance with the interim statusrequirements of 40 CFR part 266, subpart H; (v) Except for control devices used for small glycoldehydration units, a hazardous waste incinerator for whichthe owner or operator has been issued a final permit under40 CFR part 270 and complies with the requirements of 40CFR part 264, subpart O, or has certified compliance withthe interim status requirements of 40 CFR part 265, subpartO.* * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

550. Page 550 of 604 (2) An owner or operator shall design and operate eachflare, as defined in §63.1271, in accordance with therequirements specified in §63.11(b) and the compliancedetermination shall be conducted using Method 22 of 40 CFRpart 60, appendix A, to determine visible emissions. (3) For a performance test conducted to demonstratethat a control device meets the requirements of§63.1281(d)(1), (e)(3)(ii), or (f)(1) the owner or operatorshall use the test methods and procedures specified inparagraphs (d)(3)(i) through (v) of this section. Theinitial and periodic performance tests shall be conductedaccording to the schedule specified in paragraph (d)(3)(vi)of this section. (i) * * ** * * * * (B) To determine compliance with the enclosedcombustion device total HAP concentration limit specifiedin §63.1281(d)(1)(i)(B), or the BTEX emission limitspecified in §63.1275(b)(1)(iii), the sampling site shallbe located at the outlet of the combustion device.* * * * * (iv) * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

551. Page 551 of 604* * * * * (C) * * ** * * * * (1) The emission rate correction factor for excessair, integrated sampling and analysis procedures of Method3A or 3B, 40 CFR part 60, appendix A, shall be used todetermine the oxygen concentration (%O2d). The samples shallbe taken during the same time that the samples are takenfor determining TOC concentration or total HAPconcentration.* * * * * (v) To determine compliance with the BTEX emissionlimit specified in §63.1281(f)(1) the owner or operatorshall use one of the following methods: Method 18, 40 CFRpart 60, appendix A; ASTM D6420–99 (2004), as specified in§63.772(a)(1)(ii); or any other method or data that havebeen validated according to the applicable procedures inMethod 301, 40 CFR part 63, appendix A. The followingprocedures shall be used to calculate BTEX emissions: (A) The minimum sampling time for each run shall be 1hour in which either an integrated sample or a minimum offour grab samples shall be taken. If grab sampling is used, This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

552. Page 552 of 60 04then the sam n mples sha all be ta aken at a approxima ately equ ualinte ervals in time, s n such as 1 15-minute interva e als durin the ngrun. . (B) The mass ra e ate of BT TEX (Eo) shall be compute using e edthe equation and pr ns rocedures specifi s ied in pa aragraphs s(d)( (3)(v)(B) )(1) and (2) of t this sect tion. (1) The followi e ing equat tion shal be use ll ed:Wherre:Eo= Mass rat of BTE at the outlet of the c te EX e control d device,dry basis, kilogram per hour k r.Coj= Concentration of sample compone ent j of the gas streamat tthe outle of the control device, dry bas et e l , sis, part per tsmilllion by volume. vMoj= Molecular weight of sam mple comp ponent j of the g gasstreeam at th outlet of the control device, gram/gra he t am-mole.Qo= Flowrate of gas stream a the ou e at utlet of the conttroldeviice, dry standard cubic m d meter per minute. r .K2= Constant 2.494× −6 (par t, ×10 rts per mmillion) (gram-moole perstanndard cub bic meter (kilog r) gram/gram (minut m) te/hour), where ,stanndard tem mperature (gram-m e mole per standard cubic m d meter)is 2 degree C. 20 esn = Number of compon o nents in sample. (2) Whe the BT en TEX mass rate is calculat ted, only BTEX ycomp pounds me easured b Method 18, 40 CFR part 60, app by d t pendixA, o ASTM D6420–99 (2004) a specif or D as fied in§63. .772(a)(1 1)(ii), s shall be summed u using the equatio e ons inpara agraph (d d)(3)(v)( (B)(1) of this se f ection. Th his document is a prepuublication n version n, signed d by EP PA Ad dministrat tor, Lisa P. Jackso on 07/2 on 28/2011. We have t taken step ps to ensure the accura o t acy of thi version but it is not th officia is n, he al ve ersion.

553. Page 553 of 604 (vi) The owner or operator shall conduct performancetests according to the schedule specified in paragraphs(d)(3)(vi)(A) and (B) of this section. (A) An initial performance test shall be conductedwithin 180 days after the compliance date that is specifiedfor each affected source in §63.1270(d)(3) and (4) exceptthat the initial performance test for existing combustioncontrol devices at existing major sources shall beconducted no later than 3 years after the date ofpublication of the final rule in the Federal Register. Ifthe owner or operator of an existing combustion controldevice at an existing major source chooses to replace suchdevice with a control device whose model is tested under§63.1282(g), then the newly installed device shall complywith all provisions of this subpart no later than 3 yearsafter the date of publication of the final rule in theFederal Register. The performance test results shall besubmitted in the Notification of Compliance Status Reportas required in §63.1285(d)(1)(ii). (B) Periodic performance tests shall be conducted forall control devices required to conduct initial performancetests except as specified in paragraphs (e)(3)(vi)(B)(1) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

554. Page 554 of 604and (2) of this section. The first periodic performancetest shall be conducted no later than 60 months after theinitial performance test required in paragraph(d)(3)(vi)(A) of this section. Subsequent periodicperformance tests shall be conducted at intervals no longerthan 60 months following the previous periodic performancetest or whenever a source desires to establish a newoperating limit. The periodic performance test results mustbe submitted in the next Periodic Report as specified in§63.1285(e)(2)(x). Combustion control devices meeting thecriteria in either paragraph (e)(3)(vi)(B)(1) or (2) ofthis section are not required to conduct periodicperformance tests. (1) A control device whose model is tested under, andmeets the criteria of, §63.1282(g), or (2) A combustion control device tested under§63.1282(d) that meets the outlet TOC or HAP performancelevel specified in §63.1281(d)(1)(i)(B) and thatestablishes a correlation between firebox or combustionchamber temperature and the TOC or HAP performance level.* * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

555. Page 555 of 604 (4) For a condenser design analysis conducted to meetthe requirements of §63.1281(d)(1), (e)(3)(ii), or (f)(1),the owner or operator shall meet the requirements specifiedin paragraphs (d)(4)(i) and (d)(4)(ii) of this section.Documentation of the design analysis shall be submitted asa part of the Notification of Compliance Status Report asrequired in §63.1285(d)(1)(i). (i) The condenser design analysis shall include ananalysis of the vent stream composition, constituentconcentrations, flowrate, relative humidity, andtemperature, and shall establish the design outlet organiccompound concentration level, design average temperature ofthe condenser exhaust vent stream, and the design averagetemperatures of the coolant fluid at the condenser inletand outlet. As an alternative to the condenser designanalysis, an owner or operator may elect to use theprocedures specified in paragraph (d)(5) of this section.* * * * * (5) As an alternative to the procedures in paragraph(d)(4)(i) of this section, an owner or operator may electto use the procedures documented in the GRI reportentitled, “Atmospheric Rich/Lean Method for Determining This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

556. Page 556 of 604Glycol Dehydrator Emissions,” (GRI–95/0368.1) as inputs forthe model GRI-GLYCalcTM, Version 3.0 or higher, to generatea condenser performance curve. (e) Compliance demonstration for control devicesperformance requirements. This paragraph applies to thedemonstration of compliance with the control deviceperformance requirements specified in §63.1281(d)(1),(e)(3)(ii), and (f)(1). Compliance shall be demonstratedusing the requirements in paragraphs (e)(1) through (3) ofthis section. As an alternative, an owner or operator thatinstalls a condenser as the control device to achieve therequirements specified in §63.1281(d)(1)(ii), (e)(3)(ii),or (f)(1) may demonstrate compliance according to paragraph(f) of this section. An owner or operator may switchbetween compliance with paragraph (e) of this section andcompliance with paragraph (f) of this section only after atleast 1 year of operation in compliance with the selectedapproach. Notification of such a change in the compliancemethod shall be reported in the next Periodic Report, asrequired in §63.1285(e), following the change.* * * * * (2) The owner or operator shall calculate the daily This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

557. Page 557 of 604average of the applicable monitored parameter in accordancewith §63.1283(d)(4) Except that the inlet gas flowrate tothe control device shall not be averaged. (3) Compliance is achieved when the daily average ofthe monitoring parameter value calculated under paragraph(e)(2) of this section is either equal to or greater thanthe minimum or equal to or less than the maximum monitoringvalue established under paragraph (e)(1) of this section.For inlet gas flowrate, compliance with the operatingparameter limit is achieved when the value is equal to orless than the value established under §63.1282(g). (4) Except for periods of monitoring systemmalfunctions, repairs associated with monitoring systemmalfunctions, and required monitoring system qualityassurance or quality control activities (including, asapplicable, system accuracy audits and required zero andspan adjustments), the CMS required in §63.1283(d) must beoperated at all times the affected source is operating. Amonitoring system malfunction is any sudden, infrequent,not reasonably preventable failure of the monitoring systemto provide valid data. Monitoring system failures that arecaused in part by poor maintenance or careless operation This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

558. Page 558 of 604are not malfunctions. Monitoring system repairs arerequired to be completed in response to monitoring systemmalfunctions and to return the monitoring system tooperation as expeditiously as practicable. (5) Data recorded during monitoring systemmalfunctions, repairs associated with monitoring systemmalfunctions, or required monitoring system qualityassurance or control activities may not be used incalculations used to report emissions or operating levels.All the data collected during all other required datacollection periods must be used in assessing the operationof the control device and associated control system. (6) Except for periods of monitoring systemmalfunctions, repairs associated with monitoring systemmalfunctions, and required quality monitoring systemquality assurance or quality control activities (including,as applicable, system accuracy audits and required zero andspan adjustments), failure to collect required data is adeviation of the monitoring requirements. (f) Compliance demonstration with percent reduction oremission limit performance requirements—condensers. Thisparagraph applies to the demonstration of compliance with This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

559. Page 559 of 604the performance requirements specified in§63.1281(d)(1)(ii), (e)(3) or (f)(1) for condensers.Compliance shall be demonstrated using the procedures inparagraphs (f)(1) through (f)(3) of this section. (1) The owner or operator shall establish a site-specific condenser performance curve according to theprocedures specified in §63.1283(d)(5)(ii). For sourcesrequired to meet the BTEX limit in accordance with§63.1281(e) or (f)(1) the owner or operator shall identifythe minimum percent reduction necessary to meet the BTEXlimit. (2) Compliance with the percent reduction requirementin §63.1281(d)(1)(ii), (e)(3), or (f)(1) shall bedemonstrated by the procedures in paragraphs (f)(2)(i)through (iii) of this section.* * * * * (iii) Except as provided in paragraphs (f)(2)(iii)(A),(B), and (D) of this section, at the end of each operatingday the owner or operator shall calculate the 30-dayaverage HAP, or BTEX, emission reduction, as appropriate,from the condenser efficiencies as determined in paragraph(f)(2)(ii) of this section for the preceding 30 operating This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

560. Page 560 of 604days. If the owner or operator uses a combination ofprocess modifications and a condenser in accordance withthe requirements of §63.1281(e), the 30-day average HAPemission, or BTEX, emission reduction, shall be calculatedusing the emission reduction achieved through processmodifications and the condenser efficiency as determined inparagraph (f)(2)(ii) of this section, both for thepreceding 30 operating days. (A) After the compliance date specified in§63.1270(d), an owner or operator of a facility that storesnatural gas that has less than 30 days of data fordetermining the average HAP, or BTEX, emission reduction,as appropriate, shall calculate the cumulative average atthe end of the withdrawal season, each season, until 30days of condenser operating data are accumulated. For afacility that does not store natural gas, the owner oroperator that has less than 30 days of data for determiningaverage HAP, or BTEX, emission reduction, as appropriate,shall calculate the cumulative average at the end of thecalendar year, each year, until 30 days of condenseroperating data are accumulated. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

561. Page 561 of 604 (B) After the compliance date specified in§63.1270(d), for an owner or operator that has less than 30days of data for determining the average HAP, or BTEX,emission reduction, as appropriate, compliance is achievedif the average HAP, or BTEX, emission reduction, asappropriate, calculated in paragraph (f)(2)(iii)(A) of thissection is equal to or greater than 95.0 percent.* * * * * (3) Compliance is achieved based on the applicablecriteria in paragraphs (f)(3)(i) or (ii) of this section. (i) For sources meeting the HAP emission reductionspecified in §63.1281(d)(1)(ii) or (e)(3) if the averageHAP emission reduction calculated in paragraph (f)(2)(iii)of this section is equal to or greater than 95.0 percent. (ii) For sources required to meet the BTEX limit under§63.1281(e)(3) or (f)(1), compliance is achieved if theaverage BTEX emission reduction calculated in paragraph(f)(2)(iii) of this section is equal to or greater than theminimum percent reduction identified in paragraph (f)(1) ofthis section.* * * * * (g) Performance testing for combustion control devices This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

562. Page 562 of 604- manufacturers’ performance test. (1) This paragraph applies to the performance testingof a combustion control device conducted by the devicemanufacturer. The manufacturer shall demonstrate that aspecific model of control device achieves the performancerequirements in (g)(7) of this section by conducting aperformance test as specified in paragraphs (g)(2) through(6) of this section. (2) Performance testing shall consist of three one-hour (or longer) test runs for each of the four followingfiring rate settings making a total of 12 test runs pertest. Propene (propylene) gas shall be used for the testingfuel. All fuel analyses shall be performed by anindependent third-party laboratory (not affiliated with thecontrol device manufacturer or fuel supplier). (i) 90 - 100 percent of maximum design rate (fixedrate). (ii) 70 - 100 - 70 percent (ramp up, ramp down). Beginthe test at 70 percent of the maximum design rate. Withinthe first 5 minutes, ramp the firing rate to 100 percent ofthe maximum design rate. Hold at 100 percent for 5 minutes.In the 10-15 minute time range, ramp back down to 70 This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

563. Page 563 of 604percent of the maximum design rate. Repeat three more timesfor a total of 60 minutes of sampling. (iii) 30 - 70 - 30 percent (ramp up, ramp down). Beginthe test at 30 percent of the maximum design rate. Withinthe first 5 minutes, ramp the firing rate to 70 percent ofthe maximum design rate. Hold at 70 percent for 5 minutes.In the 10-15 minute time range, ramp back down to 30percent of the maximum design rate. Repeat three more timesfor a total of 60 minutes of sampling. (iv) 0 - 30 - 0 percent (ramp up, ramp down). Beginthe test at 0 percent of the maximum design rate. Withinthe first 5 minutes, ramp the firing rate to 100 percent ofthe maximum design rate. Hold at 30 percent for 5 minutes.In the 10-15 minute time range, ramp back down to 0 percentof the maximum design rate. Repeat three more times for atotal of 60 minutes of sampling. (3) All models employing multiple enclosures shall betested simultaneously and with all burners operational.Results shall be reported for the each enclosureindividually and for the average of the emissions from allinterconnected combustion enclosures/chambers. Controldevice operating data shall be collected continuously This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

564. Page 564 of 604throughout the performance test using an electronic DataAcquisition System and strip chart. Data shall be submittedwith the test report in accordance with paragraph (8)(iii)of this section. (4) Inlet testing shall be conducted as specified inparagraphs (4)(i) through (iii) of this section. (i) The fuel flow metering system shall be located inaccordance with Method 2A, 40 CFR part 60, appendix A-1,(or other approved procedure) to measure fuel flow rate atthe control device inlet location. The fitting for fillingfuel sample containers shall be located a minimum of 8 pipediameters upstream of any inlet fuel flow monitoring meter. (ii) Inlet flow rate shall be determined using Method2A, 40 CFR part 60, appendix A-1. Record the start and stopreading for each 60-minute THC test. Record the gaspressure and temperature at 5-minute intervals throughouteach 60-minute THC test. (iii) Inlet fuel sampling shall be conducted inaccordance with the criteria in paragraphs (g)(4)(iii)(A)and (B) of this section. (A) At the inlet fuel sampling location, securelyconnect a Silonite-coated stainless steel evacuated This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

565. Page 565 of 604canister fitted with a flow controller sufficient to fillthe canister over a 1 hour period. Filling shall beconducted as specified in the following: (1) Open the canister sampling valve at the beginningof the total hydrocarbon (THC) test, and close the canisterat the end of the THC test. (2) Fill one canister for each THC test run. (3) Label the canisters individually and record on achain of custody form. (B) Each fuel sample shall be analyzed using thefollowing methods. The results shall be included in thetest report. (1) Hydrocarbon compounds containing between one andfive atoms of carbon plus benzene using ASTM D1945-03. (2) Hydrogen (H2), carbon monoxide (CO), carbon dioxide(CO2), nitrogen (N2), oxygen (O2) using ASTM D1945-03. (3) Carbonyl sulfide, carbon disulfide plus mercaptansusing ASTM D5504. (4) Higher heating value using ASTM D3588-98 or ASTMD4891-89. (5) Outlet testing shall be conducted in accordancewith the criteria in paragraphs (g)(5)(i) through (v) of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

566. Page 566 of 604this section. (i) Sampling and flowrate measured in accordance withthe following: (A) The outlet sampling location shall be a minimum of4 equivalent stack diameters downstream from the highestpeak flame or any other flow disturbance, and a minimum ofone equivalent stack diameter upstream of the exit or anyother flow disturbance. A minimum of two sample ports shallbe used. (B) Flow rate shall be measured using Method 1, 40 CFRpart 60, Appendix 1, for determining flow measurementtraverse point location; and Method 2, 40 CFR part 60,Appendix 1, shall be used to measure duct velocity. If lowflow conditions are encountered (i.e., velocity pressuredifferentials less than 0.05 inches of water) during theperformance test, a more sensitive manometer shall be usedto obtain an accurate flow profile. (ii) Molecular weight shall be determined as specifiedin paragraphs (g)(4)(iii)(B), (5)(ii)(A) and (B) of thissection. (A) An integrated bag sample shall be collected duringthe Method 4, 40 CFR part 60, Appendix A, moisture test. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

567. Page 567 of 604Analyze the bag sample using a gas chromatograph-thermalconductivity detector (GC-TCD) analysis meeting thefollowing criteria: (1) Collect the integrated sample throughout theentire test, and collect representative volumes from eachtraverse location. (2) The sampling line shall be purged with stack gasbefore opening the valve and beginning to fill the bag. (3) The bag contents shall be kneaded or otherwisevigorously mixed prior to the GC analysis. (4) The GC-TCD calibration procedure in Method 3C, 40CFR part 60, Appendix A, shall be modified by using EPAAlt-045 as follows: For the initial calibration, triplicateinjections of any single concentration must agree within 5percent of their mean to be valid. The calibration responsefactor for a single concentration re-check must be within10 percent of the original calibration response factor forthat concentration. If this criterion is not met, theinitial calibration using at least three concentrationlevels shall be repeated. (B) Report the molecular weight of: O2, CO2, methane(CH4), and N2 and include in the test report submitted under This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

568. Page 568 of 604§63.775(d)(iii). Moisture shall be determined using Method4, 40 CFR part 60, Appendix A. Traverse both ports with theMethod 4, 40 CFR part 60, Appendix A, sampling train duringeach test run. Ambient air shall not be introduced into theMethod 3C, 40 CFR part 60, Appendix A, integrated bagsample during the port change. (iv) Carbon monoxide shall be determined using Method10, 40 CFR part 60, Appendix A. The test shall be run atthe same time and with the sample points used for the EPAMethod 25A, 40 CFR part 60, Appendix A, testing. Aninstrument range of 0-10 per million by volume-dry (ppmvd)shall be used. (v) Visible emissions shall be determined using Method22, 40 CFR part 60, Appendix A. The test shall be performedcontinuously during each test run. A digital colorphotograph of the exhaust point, taken from the position ofthe observer and annotated with date and time, will betaken once per test run and the four photos included in thetest report. (6) Total hydrocarbons (THC) shall be determined asspecified by the following criteria: (i) Conduct THC sampling using Method 25A, 40 CFR part This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

569. Page 569 of 60460, Appendix A, except the option for locating the probe inthe center 10 percent of the stack shall not be allowed.The THC probe must be traversed to 16.7 percent, 50percent, and 83.3 percent of the stack diameter during thetesting. (ii) A valid test shall consist of three Method 25A,40 CFR part 60, Appendix A, tests, each no less than 60minutes in duration. (iii) A 0-10 parts per million by volume-wet (ppmvw)(as propane) measurement range is preferred; as analternative a 0-30 ppmvw (as carbon) measurement range maybe used. (iv) Calibration gases will be propane in air and becertified through EPA Protocol 1 – “EPA TraceabilityProtocol for Assay and Certification of Gaseous CalibrationStandards,” September 1997, as amended August 25, 1999,EPA–600/R–97/121 (or more recent if updated since 1999). (v) THC measurements shall be reported in terms ofppmvw as propane.(vi) THC results shall be corrected to 3 percent CO2, asmeasured by Method 3C, 40 CFR part 60, Appendix A. (vii) Subtraction of methane/ethane from the THC data This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

570. Page 570 of 604is not allowed in determining results. (7) Performance test criteria: (i) The control device model tested must meet thecriteria in paragraphs (g)(7)(i)(A) through (C) of thissection: (A) Method 22, 40 CFR part 60, Appendix A, resultsunder paragraph (g)(5)(v) of this section with noindication of visible emissions, and (B) Average Method 25A, 40 CFR part 60, Appendix A,results under paragraph (g)(6) of this section equal to orless than 10.0 ppmvw THC as propane corrected to 3.0percent CO2, and (C) Average CO emissions determined under paragraph(g)(5)(iv) of this section equal to or less than 10 partsppmvd, corrected to 3.0 percent CO2. (ii) The manufacturer shall determine a maximum inletgas flow rate which shall not be exceeded for each controldevice model to achieve the criteria in paragraph (g)(7)(i)of this section. (iii) A control device meeting the criteria inparagraph (g)(7)(i)(A) through (C) of this section willhave demonstrated a destruction efficiency of 98.0 percent This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

571. Page 571 of 604for HAP regulated under this subpart. (8) The owner or operator of a combustion controldevice model tested under this section shall submit theinformation listed in paragraphs (g)(8)(i) through (iii) inthe test report required under §63.775(d)(1)(iii). (i) Full schematic of the control device anddimensions of the device components. (ii) Design net heating value (minimum and maximum) ofthe device. (iii) Test fuel gas flow range (in both mass andvolume). Include the minimum and maximum allowable inletgas flow rate. (iv) Air/stream injection/assist ranges, if used.(v) The test parameter ranges listed in paragraphs(g)(8)(vi)(A) through (O) of this section, as applicablefor the tested model. (A) Fuel gas delivery pressure and temperature. (B) Fuel gas moisture range. (C) Purge gas usage range. (D) Condensate (liquid fuel) separation range. (E) Combustion zone temperature range. This isrequired for all devices that measure this parameter. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

572. Page 572 of 604 (F) Excess combustion air range. (G) Flame arrestor(s). (H) Burner manifold pressure. (I) Pilot flame sensor. (J) Pilot flame design fuel and fuel usage. (K) Tip velocity range. (L) Momentum flux ratio. (M) Exit temperature range. (N) Exit flow rate. (O) Wind velocity and direction. (vi) The test report shall include all calibrationquality assurance/quality control data, calibration gasvalues, gas cylinder certification, and strip chartsannotated with test times and calibration values. (h) Compliance demonstration for combustion controldevices - manufacturers’ performance test. This paragraphapplies to the demonstration of compliance for a combustioncontrol device tested under the provisions in paragraph (g)of this section. Owners or operators shall demonstrate thata control device achieves the performance requirements of§63.1281(d)(1), (e)(3)(ii) or (f)(1), by installing adevice tested under paragraph (g) of this section and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

573. Page 573 of 604complying with the following criteria: (1) The inlet gas flow rate shall meet the rangespecified by the manufacturer. Flow rate shall be measuredas specified in §63.1283(d)(3)(i)(H)(1). (2) A pilot flame shall be present at all times ofoperation. The pilot flame shall be monitored in accordancewith §63.1283 (d)(3)(i)(H)(2). (3) Devices shall be operated with no visibleemissions, except for periods not to exceed a total of 5minutes during any 2 consecutive hours. A visible emissionstest using Method 22, 40 CFR part 60, Appendix A, shall beperformed monthly. The observation period shall be 2 hoursand shall be used according to Method 22. (4) Compliance with the operating parameter limit isachieved when the following criteria are met: (i) The inlet gas flow rate monitored under paragraph(h)(1) of this section is equal to or below the maximumestablished by the manufacturer; and (ii) The pilot flame is present at all times; and (iii) During the visible emissions test performedunder paragraph (h)(3) of this section the duration ofvisible emissions does not exceed a total of 5 minutes This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

574. Page 574 of 604during the observation period. Devices failing the visibleemissions test shall follow the requirements in paragraphs(h)(4)(iii)(A) and (B) of this section. (A) Following the first failure, the fuel nozzle(s)and burner tubes shall be replaced. (B) If, following replacement of the fuel nozzle(s)and burner tubes as specified in paragraph (h)(4)(iii)(A),the visible emissions test is not passed in the nextscheduled test, either a performance test shall beperformed under paragraph (d) of this section, or thedevice shall be replaced with another control device whosemodel was tested, and meets, the requirements in paragraph(g) of this section. 30. Section 63.1283 is amended by: a. Revising paragraph (b); b. Revising paragraph (d)(1) introductory text; c. Revising paragraph (d)(1)(ii) and adding paragraphs(d)(1)(iii) and (iv); d. Revising paragraph (d)(2)(i) and (d)(2)(ii); e. Revising paragraphs (d)(3)(i)(A) and (B); f. Revising paragraphs (d)(3)(i)(D) and (E); g. Revising paragraphs (d)(3)(i)(F)(1) and (2); This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

575. Page 575 of 604 h. Revising paragraph (d)(3)(i)(G); i. Adding paragraph (d)(3)(i)(H); j. Revising paragraph (d)(4); k. Revising paragraph (d)(5)(i) introductory text; l. Revising paragraphs (d)(5)(i)(A) and (B); m. Adding paragraph (d)(5)(i)(C); n. Revising paragraphs (d)(5)(ii)(A) through (C); o. Revising paragraph (d)(6) introductory text; p. Revising paragraphs (d)(6)(ii); q. Revising paragraph (d)(7)(v); r. Revising paragraph (d)(8)(i)(A); and s. Revising paragraph (d)(8)(ii) to read as follows:§63.1283 Inspection and monitoring requirements.* * * * * (b) The owner or operator of a control device whosemodel was tested under 63.1282(g) shall develop aninspection and maintenance plan for each control device. Ata minimum, the plan shall contain the control devicemanufacturer’s recommendations for ensuring properoperation of the device. Semi-annual inspections shall beconducted for each control device with maintenance and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

576. Page 576 of 604replacement of control device components made in accordancewith the plan. (d) Control device monitoring requirements. (1) Foreach control device except as provided for in paragraph(d)(2) of this section, the owner or operator shall installand operate a continuous parameter monitoring system inaccordance with the requirements of paragraphs (d)(3)through (9) of this section. Owners or operators thatinstall and operate a flare in accordance with§63.1281(d)(1)(iii) or (f)(1)(iii) are exempt from therequirements of paragraphs (d)(4) and (5) of this section.The continuous monitoring system shall be designed andoperated so that a determination can be made on whether thecontrol device is achieving the applicable performancerequirements of §63.1281(d), (e)(3), or (f)(1). Eachcontinuous parameter monitoring system shall meet thefollowing specifications and requirements:* * * * * (ii) A site-specific monitoring plan must be preparedthat addresses the monitoring system design, datacollection, and the quality assurance and quality controlelements outlined in paragraph (d) of this section and in This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

577. Page 577 of 604§63.8(d). Each CPMS must be installed, calibrated,operated, and maintained in accordance with the proceduresin your approved site-specific monitoring plan. Using theprocess described in §63.8(f)(4), you may request approvalof monitoring system quality assurance and quality controlprocedures alternative to those specified in paragraphs(d)(1)(ii)(A) through (E) of this section in your site-specific monitoring plan. (A) The performance criteria and designspecifications for the monitoring system equipment,including the sample interface, detector signal analyzer,and data acquisition and calculations; (B) Sampling interface (e.g., thermocouple) locationsuch that the monitoring system will provide representativemeasurements; (C) Equipment performance checks, system accuracyaudits, or other audit procedures; (D) Ongoing operation and maintenance procedures inaccordance with provisions in §63.8(c)(1) and (c)(3); and (E) Ongoing reporting and recordkeeping procedures inaccordance with provisions in §63.10(c), (e)(1), and(e)(2)(i). This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

578. Page 578 of 604 (iii) The owner or operator must conduct the CPMSequipment performance checks, system accuracy audits, orother audit procedures specified in the site-specificmonitoring plan at least once every 12 months. (iv) The owner or operator must conduct a performanceevaluation of each CPMS in accordance with the site-specific monitoring plan. (2) * * * (i) Except for control devices for small glycoldehydration units, a boiler or process heater in which allvent streams are introduced with the primary fuel or areused as the primary fuel; (ii) Except for control devices for small glycoldehydration units, a boiler or process heater with a designheat input capacity equal to or greater than 44 megawatts. (3) * * * (i) * * * (A) For a thermal vapor incinerator that demonstratesduring the performance test conducted under §63.1282(d)that combustion zone temperature is an accurate indicatorof performance, a temperature monitoring device equippedwith a continuous recorder. The monitoring device shall This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

579. Page 579 of 604have a minimum accuracy of ±1 percent of the temperaturebeing monitored in degrees C, or ±2.5 degrees C, whichevervalue is greater. The temperature sensor shall be installedat a location representative of the combustion zonetemperature. (B) For a catalytic vapor incinerator, a temperaturemonitoring device equipped with a continuous recorder. Thedevice shall be capable of monitoring temperatures at twolocations and have a minimum accuracy of ±1 percent of thetemperatures being monitored in degrees C, or ±2.5 degreesC, whichever value is greater. One temperature sensor shallbe installed in the vent stream at the nearest feasiblepoint to the catalyst bed inlet and a second temperaturesensor shall be installed in the vent stream at the nearestfeasible point to the catalyst bed outlet.* * * * * (D) For a boiler or process heater, a temperaturemonitoring device equipped with a continuous recorder. Thetemperature monitoring device shall have a minimum accuracyof ±1 percent of the temperature being monitored in degreesC, or ±2.5 degrees C, whichever value is greater. The This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

580. Page 580 of 604temperature sensor shall be installed at a locationrepresentative of the combustion zone temperature. (E) For a condenser, a temperature monitoring deviceequipped with a continuous recorder. The temperaturemonitoring device shall have a minimum accuracy of ±1percent of the temperature being monitored in degrees C, or±2.8 degrees C, whichever value is greater. The temperaturesensor shall be installed at a location in the exhaust ventstream from the condenser. (F) * * * (1) A continuous parameter monitoring system tomeasure and record the average total regeneration streammass flow or volumetric flow during each carbon bedregeneration cycle. The flow sensor must have a measurementsensitivity of 5 percent of the flow rate or 10 cubic feetper minute, whichever is greater. The mechanicalconnections for leakage must be checked at least everymonth, and a visual inspection must be performed at leastevery 3 months of all components of the flow CPMS forphysical and operational integrity and all electricalconnections for oxidation and galvanic corrosion if yourflow CPMS is not equipped with a redundant flow sensor; and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

581. Page 581 of 604 (2) A continuous parameter monitoring system tomeasure and record the average carbon bed temperature forthe duration of the carbon bed steaming cycle and tomeasure the actual carbon bed temperature afterregeneration and within 15 minutes of completing thecooling cycle. The temperature monitoring device shall havea minimum accuracy of ±1 percent of the temperature beingmonitored in degrees C, or ±2.5 degrees C, whichever valueis greater. (G) For a nonregenerative-type carbon adsorptionsystem, the owner or operator shall monitor the designcarbon replacement interval established using a performancetest performed in accordance with §63.1282(d)(3) and shallbe based on the total carbon working capacity of thecontrol device and source operating schedule. (H) For a control device whose model is tested under§63.1282(g): (1) A continuous monitoring system that measures gasflow rate at the inlet to the control device. Themonitoring instrument shall have an accuracy of plus orminus 2 percent or better. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

582. Page 582 of 604 (2) A heat sensing monitoring device equipped with acontinuous recorder that indicates the continuous ignitionof the pilot flame.* * * * * (4) Using the data recorded by the monitoring system,except for inlet gas flowrate, the owner or operator mustcalculate the daily average value for each monitoredoperating parameter for each operating day. If theemissions unit operation is continuous, the operating dayis a 24-hour period. If the emissions unit operation is notcontinuous, the operating day is the total number of hoursof control device operation per 24-hour period. Valid datapoints must be available for 75 percent of the operatinghours in an operating day to compute the daily average. (5) * * * (i) The owner or operator shall establish a minimumoperating parameter value or a maximum operating parametervalue, as appropriate for the control device, to define theconditions at which the control device must be operated tocontinuously achieve the applicable performancerequirements of §63.1281(d)(1), (e)(3)(ii), or (f)(1). Each This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

583. Page 583 of 604minimum or maximum operating parameter value shall beestablished as follows: (A) If the owner or operator conducts performancetests in accordance with the requirements of §63.1282(d)(3)to demonstrate that the control device achieves theapplicable performance requirements specified in§63.1281(d)(1), (e)(3)(ii), or (f)(1), then the minimumoperating parameter value or the maximum operatingparameter value shall be established based on valuesmeasured during the performance test and supplemented, asnecessary, by a condenser design analysis or control devicemanufacturers recommendations or a combination of both. (B) If the owner or operator uses a condenser designanalysis in accordance with the requirements of§63.1282(d)(4) to demonstrate that the control deviceachieves the applicable performance requirements specifiedin §63.1281(d)(1), (e)(3)(ii), or (f)(1), then the minimumoperating parameter value or the maximum operatingparameter value shall be established based on the condenserdesign analysis and may be supplemented by the condensermanufacturers recommendations. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

584. Page 584 of 604 (C) If the owner or operator operates a control devicewhere the performance test requirement was met under§63.1282(g) to demonstrate that the control device achievesthe applicable performance requirements specified in§63.1281(d)(1), (e)(3)(ii) or (f)(1), then the maximuminlet gas flow rate shall be established based on theperformance test and supplemented, as necessary, by themanufacturer recommendations. (ii) * * * (A) If the owner or operator conducts a performancetest in accordance with the requirements of §63.1282(d)(3)to demonstrate that the condenser achieves the applicableperformance requirements in §63.1281(d)(1), (e)(3)(ii), or(f)(1), then the condenser performance curve shall be basedon values measured during the performance test andsupplemented as necessary by control device designanalysis, or control device manufacturers recommendations,or a combination or both. (B) If the owner or operator uses a control devicedesign analysis in accordance with the requirements of§63.1282(d)(4)(i) to demonstrate that the condenserachieves the applicable performance requirements specified This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

585. Page 585 of 604in §63.1281(d)(1), (e)(3)(ii), or (f)(1), then thecondenser performance curve shall be based on the condenserdesign analysis and may be supplemented by the controldevice manufacturers recommendations. (C) As an alternative to paragraph (d)(5)(ii)(B) ofthis section, the owner or operator may elect to use theprocedures documented in the GRI report entitled,“Atmospheric Rich/Lean Method for Determining GlycolDehydrator Emissions” (GRI–95/0368.1) as inputs for themodel GRI-GLYCalcTM, Version 3.0 or higher, to generate acondenser performance curve. (6) An excursion for a given control device isdetermined to have occurred when the monitoring data orlack of monitoring data result in any one of the criteriaspecified in paragraphs (d)(6)(i) through (d)(6)(v) of thissection being met. When multiple operating parameters aremonitored for the same control device and during the sameoperating day, and more than one of these operatingparameters meets an excursion criterion specified inparagraphs (d)(6)(i) through (d)(6)(iv) of this section,then a single excursion is determined to have occurred forthe control device for that operating day. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

586. Page 586 of 604* * * * * (ii) For sources meeting §63.1281(d)(1)(ii), anexcursion occurs when average condenser efficiencycalculated according to the requirements specified in§63.1282(f)(2)(iii) is less than 95.0 percent, as specifiedin §63.1282(f)(3). For sources meeting §63.1281(f)(1), anexcursion occurs when the 30-day average condenserefficiency calculated according to the requirements of§63.1282(f)(2)(iii) is less than the identified 30-dayrequired percent reduction.* * * * * (v) For control device whose model is tested under§63.1282(g) an excursion occurs when: (A) The inlet gas flow rate exceeds the maximumestablished during the test conducted under §63.1282(g). (B) Failure of the monthly visible emissions testconducted under §63.1282(h)(3) occurs. (8) * * * (i) * * * (A) During a malfunction when the affected facility isoperated during such period in accordance with §63.6(e)(1);or This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

587. Page 587 of 604* * * * * (ii) For each control device, or combinations ofcontrol devices, installed on the same emissions unit, oneexcused excursion is allowed per semiannual period for anyreason. The initial semiannual period is the 6-monthreporting period addressed by the first Periodic Reportsubmitted by the owner or operator in accordance with§63.1285(e) of this subpart.* * * * * 31. Section 63.1284 is amended by: a. Revising paragraph (b)(3) introductory text; b. Deleting and reserving paragraph (b)(3)(ii); c. Revising paragraph (b)(4)(ii) introductory text; d. Adding paragraph (b)(4)(ii)(A) through (C); e. Adding paragraph (b)(7)(ix); and f. Adding paragraph (f), (g) and (h) to read asfollows:§63.1284 Recordkeeping requirements.* * * * * (b) * * ** * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

588. Page 588 of 604 (3) Records specified in §63.10(c) for each monitoringsystem operated by the owner or operator in accordance withthe requirements of §63.1283(d). Notwithstanding theprevious sentence, monitoring data recorded during periodsidentified in paragraphs (b)(3)(i) through (iv) of thissection shall not be included in any average or percentleak rate computed under this subpart. Records shall bekept of the times and durations of all such periods and anyother periods during process or control device operationwhen monitors are not operating or failed to collectrequired data.* * * * * (ii) [Reserved]* * * * * (4) * * ** * * * * (ii) Records of the daily average value of eachcontinuously monitored parameter for each operating daydetermined according to the procedures specified in§63.1283(d)(4) of this subpart, except as specified inparagraphs (b)(4)(ii)(A) through (C) of this section. (A) For flares, the records required in paragraph (e) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

589. Page 589 of 604of this section. (B)_For condensers installed to comply with §63.1275,records of the annual 30-day rolling average condenserefficiency determined under §63.1282(f) shall be kept inaddition to the daily averages. (C) For a control device whose model is tested under§63.1282(g), the records required in paragraph (g) of thissection.* * * * * (7) * * * (ix) Records identifying the carbon replacementschedule under §63.1281(d)(5) and records of each carbonreplacement.* * * * * (f) The owner or operator of an affected sourcesubject to this subpart shall maintain records of theoccurrence and duration of each malfunction of operation(i.e., process equipment) or the air pollution controlequipment and monitoring equipment. The owner or operatorshall maintain records of actions taken during periods ofmalfunction to minimize emissions in accordance with§63.1274(a), including corrective actions to restore This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

590. Page 590 of 604malfunctioning process and air pollution control andmonitoring equipment to its normal or usual manner ofoperation. (g) Record the following when using a control devicewhose model is tested under §63.1282(g) to comply with§63.1281(d), (e)(3)(ii) and (f)(1): (1) All visible emission readings and flowratemeasurements made during the compliance determinationrequired by §63.1282(h); and (2) All hourly records and other recorded periods whenthe pilot flame is absent. (h) The date the semi-annual maintenance inspectionrequired under §63.1283(b) is performed. Include a list ofany modifications or repairs made to the control deviceduring the inspection and other maintenance performed suchas cleaning of the fuel nozzles. 32. Section 63.1285 is amended by: a. Revising paragraph (b)(1) introductory text; b. Adding paragraphs (b)(1)(i) and (ii); c. Revising paragraph (b)(6); d. Deleting and reserving paragraph (b)(7); e. Revising paragraph (d)(1) introductory text; This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

591. Page 591 of 604 f. Revising paragraphs (d)(1)(i); g. Revising paragraph (d)(1)(ii) introductory text; h. Revising paragraph (d)(2) introductory text; i. Revising paragraph (d)(2)(ii); j. Revising paragraph (d)(4)(ii); k. Adding paragraph (d)(4)(iv); l. Revising paragraph (d)(10); m. Adding paragraphs (d)(11) and (d)(12); n. Revising paragraph (e)(2) introductory text; o. Revising paragraphs (e)(2)(ii)(B); p. Adding paragraphs (e)(2)(ii)(D) and (E); q. Adding paragraphs (e)(2)(x),(xi) and (xii); and r. Adding paragraph (g) to read as follows:§63.1285 Reporting requirements.* * * * * (b) * * * (1) The initial notifications required for existingaffected sources under §63.9(b)(2) shall be submitted asprovided in paragraphs (b)(1)(i) and (ii) of this section. (i) Except as otherwise provided in paragraph (ii),the initial notification shall be submitted by 1 year afteran affected source becomes subject to the provisions of This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

592. Page 592 of 604this subpart or by June 17, 2000, whichever is later.Affected sources that are major sources on or before June17, 2000 and plan to be area sources by June 17, 2002 shallinclude in this notification a brief, nonbindingdescription of a schedule for the action(s) that areplanned to achieve area source status. (ii) An affected source identified under§63.1270(d)(3) shall submit an initial notificationrequired for existing affected sources under §63.9(b)(2)within 1 year after the affected source becomes subject tothe provisions of this subpart or by one year afterpublication of the final rule in the Federal Register,whichever is later. An affected source identified under§63.1270(d)(3) that plans to be an area source by threeyears after publication of the final rule in the FederalRegister, shall include in this notification a brief,nonbinding description of a schedule for the action(s) thatare planned to achieve area source status.* * * * * (6) If there was a malfunction during the reportingperiod, the Periodic Report specified in paragraph (e) ofthis section shall include the number, duration, and a This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

593. Page 593 of 604brief description for each type of malfunction whichoccurred during the reporting period and which caused ormay have caused any applicable emission limitation to beexceeded. The report must also include a description ofactions taken by an owner or operator during a malfunctionof an affected source to minimize emissions in accordancewith §63.1274(h), including actions taken to correct amalfunction. (7) [Reserved]* * * * * (d) * * * (1) If a closed-vent system and a control device otherthan a flare are used to comply with §63.1274, the owner oroperator shall submit the information in paragraph(d)(1)(iii) of this section and the information in eitherparagraph (d)(1)(i) or(ii) of this section. (i) The condenser design analysis documentationspecified in §63.1282(d)(4) of this subpart if the owner oroperator elects to prepare a design analysis; or (ii) If the owner or operator is required to conduct aperformance test, the performance test results includingthe information specified in paragraphs (d)(1)(ii)(A) and This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

594. Page 594 of 604(B) of this section. Results of a performance testconducted prior to the compliance date of this subpart canbe used provided that the test was conducted using themethods specified in §63.1282(d)(3), and that the testconditions are representative of current operatingconditions. If the owner or operator operates a combustioncontrol device model tested under §63.1282(g), anelectronic copy of the performance test results shall besubmitted via email to Oil_and_Gas_PT@EPA.GOV.* * * * * (2) If a closed-vent system and a flare are used tocomply with §63.1274, the owner or operator shall submitperformance test results including the information inparagraphs (d)(2)(i) and (ii) of this section. The owneror operator shall also submit the information in paragraph(d)(2)(iii) of this section.* * * * * (ii) A statement of whether a flame was present at thepilot light over the full period of the compliancedetermination.* * * * * (4) * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

595. Page 595 of 604* * * * * (ii) An explanation of the rationale for why the owneror operator selected each of the operating parameter valuesestablished in §63.1283(d)(5) of this subpart. Thisexplanation shall include any data and calculations used todevelop the value, and a description of why the chosenvalue indicates that the control device is operating inaccordance with the applicable requirements of§63.1281(d)(1), (e)(3)(ii), or (f)(1).* * * * * (iv) For each carbon adsorber, the predeterminedcarbon replacement schedule as required in§63.1281(d)(5)(i).* * * * * (10) The owner or operator shall submit the analysisprepared under §63.1281(e)(2) to demonstrate that theconditions by which the facility will be operated toachieve the HAP emission reduction of 95.0 percent, or theBTEX limit in §63.1275(b)(1)(iii) through processmodifications or a combination of process modifications andone or more control devices. (11) If the owner or operator installs a combustion This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

596. Page 596 of 604control device model tested under the procedures in§63.1282(g), the data listed under §63.1282(g)(8). (12) For each combustion control device model testedunder §63.1282(g), the information listed in paragraphs(d)(12)(i) through (vi) of this section. (i) Name, address and telephone number of the controldevice manufacturer. (ii) Control device model number. (iii) Control device serial number. (iv) Date of control device certification test. (v) Manufacturer’s HAP destruction efficiency rating. (vi) Control device operating parameters, maximumallowable inlet gas flowrate.* * * * * (e) * * ** * * * * (2) The owner or operator shall include theinformation specified in paragraphs (e)(2)(i) through (xii)of this section, as applicable.* * * * * (ii) * * ** * * * * This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

597. Page 597 of 604 (B) For each excursion caused when the 30-day averagecondenser control efficiency is less than the value, asspecified in §63.1283(d)(6)(ii), the report must includethe 30-day average values of the condenser controlefficiency, and the date and duration of the period thatthe excursion occurred.* * * * * (D) For each excursion caused when the maximum inletgas flow rate identified under §63.1282(g) is exceeded, thereport must include the values of the inlet gas identifiedand the date and duration of the period that the excursionoccurred. (E) For each excursion caused when visible emissionsdetermined under §63.1282(h) exceed the maximum allowableduration, the report must include the date and duration ofthe period that the excursion occurred.* * * * * (x) The results of any periodic test as required in§63.1282(d)(3) conducted during the reporting period. (xi) For each carbon adsorber used to meet the controldevice requirements of §63.1281(d)(1), records of each This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

598. Page 598 of 604carbon replacement that occurred during the reportingperiod. (xii) For combustion control device inspectionsconducted in accordance with §63.1283(b) the recordsspecified in §63.1284(h).* * * * * (g) Electronic Reporting. (1) As of January 1, 2012,and within 60 days after the date of completing eachperformance test, as defined in §63.2 and as required inthis subpart, you must submit performance test data, exceptopacity data, electronically to the EPA’s Central DataExchange (CDX) by using the Electronic Reporting Tool (ERT)(see http://www.epa.gov/ttn/chief/ert/ert tool.html/). Onlydata collected using test methods compatible with ERT aresubject to this requirement to be submitted electronicallyinto the EPA’s WebFIRE database. (2) All reports required by this subpart not subjectto the requirements in paragraphs (g)(1) of this sectionmust be sent to the Administrator at the appropriateaddress listed in §63.13. If acceptable to both theAdministrator and the owner or operator of a source, thesereports may be submitted on electronic media. The This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

599. Page 599 of 604Administrator retains the right to require submittal ofreports subject to paragraph (g)(1) of this section inpaper format.* * * * * 33. Section 63.1287 is amended by revising paragraph(a) to read as follows: (a) If, in the judgment of the Administrator, analternative means of emission limitation will achieve areduction in HAP emissions at least equivalent to thereduction in HAP emissions from that source achieved underthe applicable requirements in §§63.1274 through 63.1281,the Administrator will publish a notice in the FederalRegister permitting the use of the alternative means forpurposes of compliance with that requirement. The noticemay condition the permission on requirements related to theoperation and maintenance of the alternative means. 34. Appendix to Subpart HHH of Part 63 – Table isamended by revising Table 2 to read as follows:Appendix to Subpart HHH of Part 63—Tables* * * * * Table 2 to Subpart HHH of Part 63—Applicability of 40 CFR Part 63 General Provisions to Subpart HHHGeneral provisions Applicable Explanationreference to This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

600. Page 600 of 604 subpart HHH§63.1(a)(1) ..... Yes.§63.1(a)(2) ..... Yes.§63.1(a)(3) ..... Yes.§63.1(a)(4) ..... Yes.§63.1(a)(5) ..... No ..........Section reserved.§63.1(a)(6) ..... Yes.through (a)(8)§63.1(a)(9) ..... No ..........Section reserved.§63.1(a)(10) .... Yes.§63.1(a)(11) .... Yes.§63.1(a)(12) .... Yes.through (a)(14)§63.1(b)(1) ..... No ..........Subpart HHH specifies applicability.§63.1(b)(2) ..... Yes.§63.1(b)(3) ..... No.§63.1(c)(1) ..... No ..........Subpart HHH specifies applicability.§63.1(c)(2) ..... No.§63.1(c)(3) ..... No ..........Section reserved.§63.1(c)(4) ..... Yes.§63.1(c)(5) ..... Yes.§63.1(d) ........ No ..........Section reserved.§63.1(e) ........ Yes.§63.2 ........... Yes .........Except definition of major source is unique for this source category and there are additional definitions in subpart HHH.§63.3(a) through Yes.(c)§63.4(a)(1) ..... Yes.through (a)(3)§63.4(a)(4) ..... No ..........Section reserved.§63.4(a)(5) ..... Yes.§63.4(b) ........ Yes.§63.4(c) ........ Yes.§63.5(a)(1) ..... Yes.§63.5(a)(2) ..... No ..........Preconstruction review required only for major sources that commence construction after promulgation of the standard.§63.5(b)(1) ..... Yes. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

601. Page 601 of 604§63.5(b)(2) ..... No ..........Section reserved.§63.5(b)(3) ..... Yes.§63.5(b)(4) ..... Yes.§63.5(b)(5) ..... Yes.§63.5(b)(6) ..... Yes.§63.5(c) ........ No ..........Section reserved.§63.5(d)(1) ..... Yes.§63.5(d)(2) ..... Yes.§63.5(d)(3) ..... Yes.§63.5(d)(4) ..... Yes.§63.5(e) ........ Yes.§63.5(f)(1) ..... Yes.§63.5(f)(2) ..... Yes.§63.6(a) ........ Yes.§63.6(b)(1) ..... Yes.§63.6(b)(2) ..... Yes.§63.6(b)(3) ..... Yes.§63.6(b)(4) ..... Yes.§63.6(b)(5) ..... Yes.§63.6(b)(6) ..... No ..........Section reserved.§63.6(b)(7) ..... Yes.§63.6(c)(1) ..... Yes.§63.6(c)(2) ..... Yes.§63.6(c)(3) and No ..........Section reserved.(c)(4)§63.6(c)(5) ..... Yes.§63.6(d) ........ No ..........Section reserved.§63.6(e) ........ Yes.§63.6(e) ........ Yes .........Except as otherwise specified.§63.6(e)(1)(i) .. No ..........See §63.1274(h) for general duty requirement.§63.6(e)(1)(ii) No.§63.6(e)(1)(iii) Yes.§63.6(e)(2) ..... Yes.§63.6(e)(3) ..... No.§63.6(f)(1) ..... No.§63.6(f)(2) ..... Yes.§63.6(f)(3) ..... Yes.§63.6(g) ........ Yes.§63.6(h) ........ No ..........Subpart HHH does not contain opacity or visible emission standards.§63.6(i)(1) ..... Yes.through (i)(14) This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

602. Page 602 of 604§63.6(i)(15) .... No ..........Section reserved.§63.6(i)(16) .... Yes.§63.6(j) ........ Yes.§63.7(a)(1) ..... Yes.§63.7(a)(2) ..... Yes .........But the performance test results must be submitted within 180 days after the compliance date.§63.7(a)(3) ..... Yes.§63.7(b) ........ Yes.§63.7(c) ........ Yes.§63.7(d) ........ Yes.§63.7(e)(1) ..... No.§63.7(e)(2) ..... Yes.§63.7(e)(3) ..... Yes.§63.7(e)(4) ..... Yes.§63.7(f) ........ Yes.§63.7(g) ........ Yes.§63.7(h) ........ Yes.§63.8(a)(1) ..... Yes.§63.8(a)(2) ..... Yes.§63.8(a)(3) ..... No ..........Section reserved.§63.8(a)(4) ..... Yes.§63.8(b)(1) ..... Yes.§63.8(b)(2) ..... Yes.§63.8(b)(3) ..... Yes.§63.8(c)(1) ..... Yes.63.8(c)(1)(i). No§63.8(c)(1)(ii). Yes§63.8(c)(1)(iii) Pending§63.8(c)(2) ..... Yes.§63.8(c)(3) ..... Yes.§63.8(c)(4) ..... No.§63.8(c)(5) ..... Yes.through (c)(8)§63.8(d) ........ Yes.§63.8(d)(3) ..... Yes .........Except for last sentence, which refers to an SSM plan. SSM plans are not required.§63.8(e) ........ Yes .........Subpart HHH does not specifically require continuous emissions monitor performance evaluations, however, the Administrator can request that one be conducted. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

603. Page 603 of 604§63.8(f)(1) ..... Yes .........through (f)(5)§63.8(f)(6) ..... No ..........Subpart HHH does not require continuous emissions monitoring.§63.8(g) ........ No ..........Subpart HHH specifies continuous monitoring system data reduction requirements.§63.9(a) ........ Yes.§63.9(b)(1) ..... Yes.§63.9(b)(2) ..... Yes .........Existing sources are given 1 year (rather than 120 days) to submit this notification.§63.9(b)(3) ..... Yes.§63.9(b)(4) ..... Yes.§63.9(b)(5) ..... Yes.§63.9(c) ........ Yes.§63.9(d) ........ Yes.§63.9(e) ........ Yes.§63.9(f) ........ No.§63.9(g) ........ Yes.§63.9(h)(1) ..... Yes.through (h)(3)§63.9(h)(4) ..... No ..........Section reserved.§63.9(h)(5) and . Yes.(h)(6)§63.9(i) ........ Yes.§63.9(j) ........ Yes.§63.10(a) ....... Yes.§63.10(b)(1) .... Yes .........Section 63.1284(b)(1) requires sources to maintain the most recent 12 months of data on-site and allows offsite storage for the remaining 4 years of data.§63.10(b)(2) .... Yes.§63.10(b)(2)(i) . No ..........§63.10(b)(2)(ii) No ..........See §63.1284(f) for recordkeeping of occurrence, duration, and actions taken during malfunction§63.10(b)(2)(iii) Yes.§63.10(b)(2)(iv) No.through (b)(2)(v)§63.10(b)(2)(vi) Yes.through This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

604. Page 604 of 604(b)(2)(xiv)§63.10(b)(3) .... No.§63.10(c)(1) .... Yes.§63.10(c)(2) .... No ..........Sections reserved.through (c)(4)§63.10(c)(5) .... Yes.through (c)(8)§63.10(c)(9) .... No ..........Section reserved.§63.10(c)(10) and No ..........See §63.1284(f)for recordkeeping(c)(11) of malfunctions§63.10(c)(12) ... Yes.through (c)(14)§63.10(c)(15) ... No.§63.10(d)(1) .... Yes.§63.10(d)(2) .... Yes.§63.10(d)(3) .... Yes.§63.10(d)(4) .... Yes.§63.10(d)(5) .... No ..........See §63.1285(b)(6) for reporting of malfunctions.§63.10(e)(1) .... Yes.§63.10(e)(2) .... Yes.§63.10(e)(3)(i) . Yes .........Subpart HHH requires major sources to submit Periodic Reports semi-annually.§63.10(e)(3)(i)(A) Yes.§63.10(e)(3)(i)(B) Yes.§63.10(e)(3)(i)(C) No ..........Subpart HHH does not require quarterly reporting for excess emissions.§63.10(e)(3)(ii) Yes.through(e)(3)(viii)§63.10(f) ....... Yes.§63.11(a) and (b) Yes.§63.11(c), (d), . Yes.and (e)§63.12(a) through Yes.(c)§63.13(a) through Yes.(c)§63.14(a) and (b) Yes.§63.15(a) and (b) Yes. This document is a prepublication version, signed by EPA Administrator, Lisa P. Jackson on 07/28/2011. We have taken steps to ensure the accuracy of this version, but it is not the official version.

SlideShare for iOS

EPA Proposed Amendments to Air Regulations for the Oil and Natural Gas Industry (Full Copy)

by Marcellus Drilling News

on Jul 29, 2011

Statistics

Views

Actions

1 Embed 402

Accessibility

Categories

Upload Details

Usage Rights

Report content

EPA Proposed Amendments to Air Regulations for the Oil and Natural Gas Industry (Full Copy) Document Transcript