COMMENTS OF THE UTILITY AIR REGULATORY GROUP on the UNITED STATES ENVIRONMENTAL PROTECTION AGENCY’S CARBON POLLUTION EMISSION GUIDELINES FOR EXISTING STATIONARY SOURCES: ELECTRIC UTILITY GENERATING UNITS; PROPOSED RULE 79 Fed. Reg. 34,830 (June 18, 2014) Docket ID No. EPA-HQ-OAR-2013-0602 Hunton & Williams LLP F. William Brownell Lauren E. Freeman Allison D. Wood Ted J. Murphy Craig S. Harrison Tauna M. Szymanski Aaron M. Flynn Andrew D. Knudsen 2200 Pennsylvania Ave., NW Washington, D.C. 20037 (202) 955-1500 Counsel for the Utility Air Regulatory Group Dated: December 1, 2014 TABLE OF CONTENTS Page I. Introduction ..........................................................................................................................2 II. Executive Summary .............................................................................................................5 III. The Proposed Guidelines Are Fundamentally Divorced from the Narrow and Exclusive Focus of Section 111 of the CAA on Stationary Sources in Listed Source Categories. ..................................................................................................17 A. B. IV. V. VI. Section 111 Authorizes Standards of Performance That Are Achievable for Individual Sources in a Source Category Based on Measures Those Sources Can Implement Themselves. .........................................18 1. The Text of Section 111.............................................................................19 2. CAA Context for Section 111 ....................................................................22 3. The History of EPA’s Implementation of Section 111 ..............................25 EPA’s Proposed Action Deviates So Far from the CAA as To Be Unrecognizable as an Exercise of Section 111 Authority. ....................................29 1. EPA’s Proposed Guidelines Impermissibly Base Standards of Performance on Measures that Go Beyond the Regulated Source. .......................................................................................................30 2. EPA Cannot Consider Reduced Utilization of Regulated Sources as BSER........................................................................................38 3. The Proposed Guidelines Are Based on Measures that States and EPA Cannot Enforce Against Regulated Sources. ......................................................................................................42 EPA Is Prohibited from Regulating Pollutants from a Source Category Already Regulated Under Section 112 of the CAA. ..........................................................52 A. EPA’s Interpretation of the Act Is Inconsistent With the Plain Language of Section 111(d), Which Contains No Ambiguity. ..............................52 B. EPA’s Proposed Construction Does Not Accord With Either Provision or With Legislative Intent. .....................................................................56 EPA Lacks the Sufficient Legal Prerequisites To Propose and Finalize the Proposed Guidelines. .........................................................................................................59 A. EPA’s Proposed NSPS for New Coal-Fired EGUs Is Unlawful and Cannot Support Section 111(d) Emission Guidelines. ..........................................59 B. The Proposed NSPS for New Sources Does Not Address the Same Type of Facilities as the Proposed Guidelines. ......................................................61 Definition of “Affected EGU” ...........................................................................................66 i VII. A. EPA’s Definition of “Affected EGU” in the Proposed Guidelines Must Retain the Exclusions for Subpart Da Units That Were Proposed for New Subpart Da Units. .....................................................................68 B. The “Broad Applicability” Approach to the Proposed NSPS Would Not Support Regulating a Broader Universe of Sources Under the Proposed Guidelines. .............................................................................................69 EPA Has Unlawfully Ignored the Federal Power Act’s Division Between State and Federal Authority and Arbitrarily and Capriciously Ignored the Realities of Wholesale Electric Markets, Regional Transmission Grids, and Bulk Power System Reliability. ..................................................................................73 A. B. The Proposed Guidelines Are Inconsistent With the FPA and Violate Section 310(a) of the CAA Because They Would Override the Division of Regulatory Authority Over the Generation, Transmission, Distribution, and Sale of Electric Energy Between the States and the Federal Government. ................................................................75 1. The Proposed Guidelines Unlawfully Usurp Authority Over the Generation of Electricity, Resource Planning, and Related Matters that the FPA and Tenth Amendment Reserve to the States. .................................................................................76 2. The Proposed Guidelines Usurp Regulatory Authority that the FPA Gives Exclusively to FERC. ........................................................81 a. The Proposed Guidelines Would Impermissibly Interfere With FERC’s Exclusive Authority Under Sections 205 and 206 of the FPA To Regulate Interstate Transmission and Wholesale Sale of Electric Energy...............................................................................82 b. The Proposed Guidelines Conflict With Section 202(a) of the FPA and Impermissibly Intrude on FERC’s Authority To Coordinate Regional Transmission. .................................................................................89 c. The Proposed Guidelines Unlawfully Intrude on FERC’s Authority Under Section 215 of the FPA To Ensure Reliability of the BPS. ..................................................92 EPA Has Arbitrarily and Capriciously Ignored the Realities of the FERC-Regulated Wholesale Electric Markets and Bulk Power System. ...................................................................................................................94 1. EPA Has Unreasonably Failed To Consult With FERC Regarding the Proposed Guidelines’ Technical Feasibility and Impacts on Reliability and Markets. ...................................................95 2. EPA’s Erroneous Assumption that States Control the Dispatch of EGUs Is Likely to Result in Serious Distortions and Unintended Consequences. .................................................................97 ii VIII. IX. 3. EPA Underestimates the Challenges of Energy Infrastructure Development and the Impact of Transmission Constraints. ................................................................................................98 4. EPA Relies on Assumptions that Independent Expert Analyses Demonstrate Are Faulty. ..........................................................101 5. EPA Has Not Accounted for the EPSA Decision’s Potential Impact on Demand-Side Resources. ........................................................109 6. EPA Inaccurately Describes RTO Capacity Markets and Overstates Their Potential To Advance the Proposed Guidelines’ Goals.....................................................................................110 If Lawful, The Proposed Guidelines Would Violate the Fifth Amendment Because Uncompensated Takings of Property Will Result. ............................................112 A. The Proposed Guidelines, if Lawful, Would Result in a Total Taking of Any EGU Forced To Shut Down To Ensure Compliance with the Guidelines. .............................................................................................114 B. The Proposed Guidelines, if Lawful, Would Result in a Partial Regulatory Taking of Any EGU That Must Significantly Curtail its Operation or Remaining Useful Life To Ensure Compliance with the Proposed Guidelines’ Emission Goals. ..........................................................115 1. The Character of the Proposed Guidelines Weighs in Favor of a Takings Claim. ..................................................................................116 2. The Proposed Guidelines Would Cause Severe Economic Impacts. ....................................................................................................119 3. The Proposed Guidelines Would Significantly Interfere With Reasonable Investment-Backed Expectations. ...............................120 EPA Usurps the Role of States and Treads Into Areas that Are Quintessentially Within the Purview of State Control in Violation of Section 111(d). .................................................................................................................122 A. State Primacy Under Section 111(d) and EPA’s Regulations Implementing Section 111(d) ..............................................................................122 B. The “Flexibility” EPA Claims Exists in the Proposed Guidelines Is Illusory. ................................................................................................................131 C. The Proposed Guidelines Conflict With State Laws and Fail To Take Account of Other CAA Requirements. .......................................................135 D. Selecting a Rate-Based or Mass-Based Goal .......................................................138 E. Types of State Plans .............................................................................................140 F. Timing for Plan Implementation and Achievement of State Emission Performance Goals ...............................................................................143 G. State Plan Approvability Criteria and Components of Approvable Plans .....................................................................................................................147 iii H. Process and Timing for Submittal of State Plans.................................................150 I. Key Considerations for States ..............................................................................152 J. 1. Affected Entities Other Than Affected EGUs .........................................153 2. Treatment of Existing State Programs .....................................................153 3. Incorporating RE and Demand-Side EE Measures Under a Rate-Based Approach ..............................................................................156 4. Quantification, Monitoring, and Verification of RE and Demand-Side EE Measures .....................................................................156 5. Reporting and Recordkeeping for Affected Entities Implementing RE and Demand-Side EE Measures .................................157 6. Treatment of Interstate Effects.................................................................158 7. Projecting Emission Performance ............................................................158 8. Potential Emission Reduction Measures Not Used To Set Proposed Goals ........................................................................................159 9. Consideration of a Facility’s “Remaining Useful Life” in Applying Standards of Performance ........................................................160 Multi-State Plan Considerations ..........................................................................163 X. EPA Cannot Require States To Regulate EGUs Simultaneously Under Sections 111(b) and 111(d) of the CAA. .........................................................................165 XI. New Source Review Issues ..............................................................................................170 XII. The Proposed Guidelines Are Not Justified by Their Purported Benefits When Those Benefits Are Characterized Properly. .........................................................177 XIII. A. Costs.....................................................................................................................178 B. Social Cost of Carbon ..........................................................................................181 C. Public Health Benefits .........................................................................................186 D. Summary ..............................................................................................................190 EPA’s Calculation of the State Goals Impermissibly Limits the States’ Discretion and Is Error-Ridden. .......................................................................................191 A. Despite EPA’s Claims, States Have Little to No Flexibility in Their Ability To Meet the State Goals Contained in the Proposed Guidelines. ...........................................................................................................191 B. EPA Should Not Include Hypothetical Generation From Units that Are Under Construction When Implementing Building Block 2 in State Goal Calculations. .......................................................................................196 C. EPA Should Not Penalize States that Have Invested in Nuclear Generation by Including “At Risk” and Under-Construction Nuclear Capacity in State Goals. .........................................................................198 D. EPA’s Goal Calculation Methodology Is Riddled With Errors. ..........................200 iv 1. EPA Applies the Building Blocks to Non-Affected Units.......................200 2. Errors Regarding 2012 Generation Data..................................................204 3. EPA Assumes Unrealistic Emission Rates for NGCC Units and Subpart Da Units. ..............................................................................209 XIV. EPA’s Building Blocks ....................................................................................................211 A. Building Block 1 Is Not Achievable. ...................................................................211 B. Building Block 2 Is Not Achievable. ...................................................................228 C. The Renewable Energy Component of Building Block 3 Is Not Achievable. ..........................................................................................................241 D. XV. 1. EPA’s Method for Calculating the Renewable Energy Targets for Building Block 3 Is Flawed. ..................................................242 2. EPA Has Not Provided All of the Data Needed To Evaluate the Proposed Guidelines in Violation of Section 307(d)(3) of the CAA. ..............................................................................................249 EPA Correctly Excluded Natural Gas Conversion and Co-Firing From Its Proposed BSER Determination. ............................................................251 Issues Raised by EPA’s October 30, 2014 Notice of Data Availability ..........................251 A. General Issues ......................................................................................................252 B. Building Block 2 Issues: New NGCC and Co-Firing as Components of BSER ..........................................................................................254 C. Issues Regarding an Alternative “Glide Path” to Compliance ............................258 D. Regionalized Approach to Building Blocks 2 and 3............................................262 E. Goal-Setting Methodology...................................................................................264 XVI. Monitoring, Recordkeeping, and Reporting Requirements for EGUs .............................267 A. Proposed CO2 Monitoring Methods.....................................................................271 1. The CEMS Option ...................................................................................271 2. The Heat Input and Fuel Factor Option ...................................................274 3. Monitoring of CO2 Mass Emissions From Integrated Equipment ................................................................................................275 B. Monitoring and Reporting of Electric Output ......................................................276 C. Monitoring and Reporting of Useful Thermal Output .........................................280 D. Monitoring Plan and QA/QC Testing ..................................................................282 E. Use of Specific Methods for Flow RATAs and Baseline Adjustments Following a Change in Method ......................................................283 F. EGU Recordkeeping Requirement ......................................................................286 XVII. Objection to Promulgation of Rules Not Proposed..........................................................287 v Comments of the Utility Air Regulatory Group on the United States Environmental Protection Agency’s Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units; Proposed Rule 79 Fed. Reg. 34,830 (June 18, 2014) Docket ID No. EPA-HQ-OAR-2013-0602 December 1, 2014 ______________________________________________________________________________ The Utility Air Regulatory Group (“UARG”) submits the following comments in response to the U.S. Environmental Protection Agency’s (“EPA” or “Agency”) proposed rule entitled “Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units,” 79 Fed. Reg. 34,830 (June 18, 2014) (“Proposed Guidelines”). Attached to these comments are several technical reports prepared in support of these comments. UARG is also filing separately its “Supplemental Materials for the Rulemaking Record,” which contains a variety of secondary materials to which these comments refer. 1 UARG is an ad hoc, unincorporated association of individual electric generating companies and industry groups. UARG’s purpose is to participate on behalf of its members collectively in EPA’s rulemakings and other Clean Air Act (“CAA” or “Act”) proceedings that affect the interests of electric generators and in litigation arising from those proceedings. 1 On October 16, 2014, UARG filed comments on a related EPA proposed rule entitled “Carbon Pollution Standards for Modified and Reconstructed Stationary Sources: Electric Utility Generating Units,” 79 Fed. Reg. 34,960 (June 18, 2014) (“Proposed Standards”). Docket ID No. EPA-HQ-OAR-2013-0603-0215 (“UARG Modified/Reconstructed Comments”). UARG also filed comments on a related EPA proposed rule entitled “Standards of Performance for Greenhouse Gas Emissions From New Stationary Sources: Electric Utility Generating Units, 79 Fed. Reg. 1430 (Jan. 8, 2014) (“January 2014 NSPS Proposal”). Docket ID No. EPA-HQ-OAR2013-0495-9666 (“UARG New Source Comments”). Both sets of comments are incorporated by reference and are included in UARG’s separately filed Supplemental Materials for the Rulemaking Record. 1 The electric generating companies that are members of UARG construct, own, and operate power plants, including fossil fuel-fired plants, and other facilities that generate electricity for residential, commercial, industrial, institutional, and government customers. Operation of fossil fuel-fired power plants results in emissions of carbon dioxide (“CO2”), the constituent greenhouse gas (“GHG”) that the Proposed Guidelines address. UARG therefore has a clear and significant interest in the present rulemaking as well as the other EPA rulemakings that are part of the Agency’s overall effort to regulate CO2 and other GHG emissions under the CAA, including GHG emissions from stationary sources. I. Introduction Imagine that EPA proposes regulations under a section of the CAA authorizing the Agency to develop standards of performance for tailpipe emissions from motor vehicles that burn fossil fuels. One might expect that these regulations would require vehicles to be equipped with emission control equipment (such as catalytic converters) or other equipment (such as onboard diagnostic computers) to limit each vehicle’s tailpipe emissions per mile. But what if EPA went further? Imagine that these regulations also attempted to reduce vehicle tailpipe emissions by requiring car owners to shift some of their travel to buses, or by requiring there to be more electric vehicle purchases, or by requiring individuals to reduce vehicle use altogether by working from home once a week. Can a “standard of performance” reasonably include measures like these? Would the CAA permit such requirements? To many, such broad requirements would seem entirely out of place and beyond the scope of EPA’s authority to limit air pollution from motor vehicles, despite the fact that these types of measures would indirectly reduce tailpipe emissions from vehicles. That is because they would have no effect on the emissions rate of the individual vehicles themselves, and they are completely beyond the control of the vehicle manufacturer. 2 Although this imagined scenario seems fanciful, it is precisely what EPA proposes to do in the Proposed Guidelines. Rather than limit itself to emission control or other production process-related measures to lower the CO2 emissions rates of existing electric generating units (“EGUs”), EPA instead proposes to require electricity generation to be shifted from coal- and oil-fired EGUs to natural gas-fired EGUs (akin to requiring car owners to take the bus more), mandate the building of additional renewable energy (akin to requiring the purpose of more electric vehicles), and require programs that will result in customers using less electricity (akin to requiring drivers to work from home one day a week). This approach violates common sense and the CAA. Section 111 of the CAA authorizes EPA and states to promulgate standards of performance for new and existing sources within certain source categories. At its heart, section 111 is quite simple. It provides for the regulation of sources through standards that are based on what an individual source can do to reduce the source’s emissions at a given level of operations. Nothing in Building Blocks 2, 3, or 4 of the Proposed Guidelines would reduce the pounds per megawatt hour (“lb/MWh”) of CO2 emitted from any EGU. Efforts to require aggregate emission reductions by targeting entities outside the designated source category exceed the scope of section 111; a “standard of performance” cannot ask another source to operate more (or ask other entities to reduce demand for a product) so that the source in the designated source category must curtail its operations or not “perform” at all. EPA’s novel, first-time interpretation in the Proposed Guidelines of the “best system of emission reduction” (“BSER”) relies on a distension of the word “system” to broaden the scope of the section 111 program beyond the EGU itself. EPA claims that it may base a standard of performance on any “set of things” that leads to reduced emissions from the source category 3 overall, ranging from utilization limits at certain units to enforceable obligations for other entities that indirectly reduce utilization of some sources. This interpretation is unlawful. The plain language, the statutory context, and the regulatory history of section 111 are clear and unambiguous. A “system of emission reduction” must begin and end at the source itself. Although it is true that section 111(d), which governs existing sources, provides more flexibility than section 111(b), which governs new sources, that flexibility does not extend to the setting of the performance standard. Once an emission guideline has been established, section 111(d) does give states flexibility to determine how sources may comply with the standard or whether that standard should be modified for individual sources. This is how EPA’s 2005 Clean Air Mercury Rule (“CAMR”), which also established emission guidelines under section 111(d), would have worked. 2 Although CAMR authorized a flexible means of complying with the standards of performance in the form of an emissions trading program, the “system of emission reduction” that was used to set the emission guidelines themselves conformed to section 111’s focus on individual sources and was based on specific pollution control technology that could be installed at individual sources. 70 Fed. Reg. 28,606, 28,617-20, 28,621 (May 18, 2005). The Supreme Court has just recently reminded EPA of the limitations on its authority to regulate GHGs under the CAA. In Utility Air Regulatory Group v. EPA, 134 S. Ct. 2427 (2014) (“UARG v. EPA”), EPA’s Tailoring Rule, which was also aimed at controlling emissions of CO2 under the Act’s prevention of significant deterioration (“PSD”) program, was declared invalid because EPA had attempted to seize far more power than the CAA granted the Agency. In its ruling, the Court urged EPA to proceed with more caution—a lesson EPA should heed now in 2 The D.C. Circuit vacated CAMR for reasons having nothing to do with how the standards of performance were established or the flexible mechanisms that were permitted for complying with them. New Jersey v. EPA, 517 F.3d 574 (D.C. Cir. 2008). 4 these proceedings. Indeed, the Court stated that EPA cannot regulate GHGs in a manner that is “‘extreme,’ ‘counterintuitive,’ or contrary to ‘common sense.’” Id. at 2441 (quoting Massachusetts v. EPA, 549 U.S. 497, 531 (2007)). The Proposed Guidelines are all of these things. Further, the Court warned EPA against interpreting its statutory authority in a way that “would bring about an enormous and transformative expansion in [its] regulatory authority without clear congressional authorization.” Id. at 2444 (citing FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 160 (2000)). The Supreme Court further stated that “[w]hen an agency claims to discover in a long-extant statute an unheralded power to regulate ‘a significant portion of the American economy,’ … we typically greet its announcement with a measure of skepticism.” Id. (quoting Brown & Williamson, 529 U.S. at 159). Here, EPA has chosen an obscure and little used provision of the CAA as the basis for an unprecedented regulatory program that will undeniably alter the nation’s economy. UARG v. EPA makes clear that EPA’s choice is not only unwise but unlawful. EPA should withdraw the Proposed Guidelines. II. Executive Summary Many legal, technical, and policy-based shortcomings render the Proposed Guidelines arbitrary, capricious, and contrary to law. These issues, described in detail in these comments, are summarized here: 3 The Proposed Guidelines are fundamentally divorced from section 111’s narrow and exclusive focus on stationary sources. • Standards promulgated under section 111 must be source-based and reflect measures that the source’s owner can integrate into the design or operation of the source itself. 3 This Executive Summary is provided for the convenience of the reader. It does not summarize every point made within these comments and should not be a substitute for reading the comments in their entirety. 5 A standard cannot be based on actions taken beyond the source itself that somehow reduce the source’s utilization. Nor can it be based simply on requiring a source to reduce its operations. A “standard of performance” cannot mandate “nonperformance.” Moreover, requiring a source to operate at reduced capacity would effectively order a fundamental redesign of the facility, which EPA has no authority to do. • Standards of performance apply to sources within listed categories and cannot regulate categories or subcategories in the aggregate or in combination with other entities outside the source category. The Act narrowly confines the stationary sources that may be regulated under section 111 to any individual “building, structure, facility, or installation which emits or may emit any air pollutant.” This definition of “stationary source” does not extend to combinations of these facilities or to nonemitting entities such as states or consumers of electricity. • EPA’s Proposed Guidelines would unlawfully impose federally enforceable obligations on a broad, undefined class of “affected entities” beyond the regulated category of existing EGUs in order to effectuate the Agency’s policy goal of reducing aggregate CO2 emissions. • Section 111 of the CAA also requires that any standard of performance be “achievable” by the sources to which it applies based on applying an “adequately demonstrated” system of emission reduction. A standard cannot be “achievable” for a source if, in order to achieve the standard, the source must rely on measures taken by some other entity that it does not control, or must simply cease to operate. • The Agency’s past rulemakings pursuant to section 111 reflect the provision’s singular focus on individual sources. In over 40 years of EPA’s implementation of the CAA, every single standard of performance has been based on a “system of emission reduction” that is incorporated into the design or operation of individual sources. • The Agency’s radical redefinition of what measures may constitute a “system of emission reduction” is impermissible in light of the past four decades of contrary implementation and section 111’s exclusive focus on what individual sources can achieve. • The broad program of energy resource management and economy-wide demand reduction that EPA has proposed to create and administer in the Proposed Guidelines bears no resemblance to the source-focused regulatory program that Congress established in section 111. • The Proposed Guidelines would violate the D.C. Circuit’s holding in ASARCO Inc. v. EPA, 578 F.2d 319 (D.C. Cir. 1978). There the court invalidated EPA’s attempt to change the basic unit to which the section 111’s new source performance standards (“NSPS”) apply from a single building, structure, facility, or installation to a 6 combination of such units. The D.C. Circuit held that EPA “has no authority to rewrite the statute in this fashion.” EPA is prohibited from regulating pollutants from a source category already regulated under section 112 of the CAA. • Section 111(d) prohibits EPA from adopting emission guidelines for existing sources from a source category when the Agency has already regulated that source category under section 112. EPA listed coal- and oil-fired EGUs as a “source category” under section 112 in 2000, and regulated emissions from these sources in 2012 under the Mercury and Air Toxics Standards (“MATS”). • Section 111 is not ambiguous, and EPA therefore is not entitled to deference in interpreting that provision. The plain language of the Act prevents EPA from regulating EGUs under section 111(d). A drafting error in the 1990 CAA amendments does not vest the Agency with the authority it claims. EPA lacks the sufficient legal prerequisites for the Proposed Guidelines. • EPA may not finalize guidelines for existing sources unless new sources of the same type of entity are already subject to a section 111(b) performance standard. • EPA’s proposed NSPS for new coal-fired EGUs fails to meet the statutory requirements for a “standard of performance” under section 111(b) of the Act because partial carbon capture and storage is not “adequately demonstrated” and the proposed NSPS does not actually require a reduction in CO2 emissions. As a result, the NSPS does not satisfy the legal prerequisite for the Proposed Guidelines. • The Proposed Guidelines do not address the “same type” of facility that EPA’s proposed new source NSPS would address, and thus on this basis alone the Proposed Guidelines lack the legal prerequisite that a standard of performance for the same type of facility must have been proposed concurrently with, or before, a proposal for existing source emission guidelines under section 111(d). • EPA violates its section 111(d) Subpart B regulations by failing to identify the “designated facilities” subject to the Proposed Guidelines. EPA instead proposes to regulate a wide range of “affected entities” and states—sources that extend far beyond the coal- and gas-fired EGUs EPA would regulate under its new source NSPS proposal. • The section 111(b) proposals for modified and reconstructed facilities also do not provide a valid basis for the Proposed Guidelines because those proposed rules define the affected facility as a fossil fuel-fired EGU. The Proposed Guidelines would regulate sources far beyond a fossil fuel-fired EGU. 7 EPA’s proposed definition of “affected EGU” is broader than the scope of new sources covered by EPA’s proposed NSPS for new EGUs. • EPA’s definition of “affected EGU” in the Proposed Guidelines improperly includes EGUs that were excluded from the proposed new source NSPS. Specifically, EPA has unlawfully included EGUs that meet the new source NSPS’s “low fossil use criterion” and its “one-third sales criterion.” That action would expand the source category beyond EPA’s January 2014 NSPS Proposal, and it is therefore unlawful. • EPA should abandon the “Broad Applicability” approach to defining the category of sources it would regulate through the Proposed Guidelines. This approach would eliminate many of the applicability criteria that were proposed for new, modified, and reconstructed sources. The Proposed Guidelines are inconsistent with the Federal Power Act (“FPA”), violate section 310(a) of the CAA, unlawfully encroach on state authority in violation of the Tenth Amendment, and ignore the realities of wholesale electric markets, regional transmission grids, and bulk power system (“BPS”) reliability. • The Proposed Guidelines ignore the careful division of responsibilities embodied in the FPA and unlawfully intrude into the sphere that the FPA reserved for state regulation in violation of the Tenth Amendment of the United States Constitution. • The Proposed Guidelines violate section 310(a) of the CAA because they conflict and interfere with the Federal Energy Regulatory Commission’s (“FERC”) exclusive authority under the FPA to: (1) regulate interstate transmission and the wholesale sale of electric energy (sections 205 and 206 of the FPA); (2) promote the voluntary regional coordination of generation and transmission facilities (section 202(a) of the FPA); and (3) ensure reliability of the BPS (section 215 of the FPA). • EPA has ignored the realities of the FERC-regulated wholesale electric market and the BPS. EPA has engaged in only limited consultation with FERC, which had no role in EPA’s assumptions. • EPA has incorrectly assumed that states can exercise command and control authority over the dispatch of EGUs. EPA has ignored both the conflicts between “environmental dispatch” and FERC’s current dispatch policies and the difficulties of changing existing FERC-jurisdictional dispatch regimes to accommodate the Proposed Guidelines. • FERC commissioners, the North American Electric Reliability Corporation (“NERC”), regional transmission organizations (“RTOs”), and independent system operators (“ISOs”) have all noted that EPA has underestimated energy and electricity infrastructure constraints and have all expressed reliability concerns. Finding ways to finance new infrastructure may prove more challenging than EPA has assumed. 8 • The Proposed Guidelines do not account for the D.C. Circuit’s ruling in Electric Power Supply Ass’n v. FERC, 753 F.3d 216 (D.C. Cir. 2014), which limits the ability of demand-side resources to participate in FERC-jurisdictional markets. EPA also incorrectly assumes that all markets managed by regional transmission organizations are “forward capacity markets.” The Proposed Guidelines violate the Fifth Amendment because uncompensated takings of property will result. • The Proposed Guidelines are unconstitutional because they will result in the taking of property without just compensation. EPA acknowledges that numerous coal- and oilfired EGUs will need to shut down to comply with this rule, and many coal-fired EGUs that do remain open will have their operations so severely curtailed that it will result in a partial taking. • The premature retirement of coal- and oil-fired EGUs constitutes a total taking because it will completely deprive the owners of these EGUs of all economically beneficial use of their property. • The severe restriction in operations of coal- and oil-fired EGUs will also constitute a taking under the test set forth in Penn Central Transportation Co. v. City of New York, 438 U.S. 104, 124 (1978). The character of the Proposed Guidelines, their severe economic impacts, and their interference with reasonable investment-backed expectations all weigh in favor of finding a taking. EPA treads into areas that are quintessentially within the purview of state control in violation of section 111(d). • Section 111(d) gives states the primary authority to develop standards of performance for existing sources. EPA is limited to establishing procedures for the development and submission of state plans regarding the implementation of the standards of performance. The Proposed Guidelines, by prescribing non-negotiable emission standards, upend this statutory division of responsibility and would limit the significant discretion that states have to vary from EPA’s emission guidelines. Case law governing state implementation plans under section 110, which is informative in this context, confirm that EPA cannot second-guess state plans. • EPA is obligated to establish highly subcategorized emission guidelines within a broad source category like existing fossil fuel-fired EGUs, but it has failed to do so in the Proposed Guidelines. • EPA touts the “flexibility” the Proposed Guidelines would provide to states, but that flexibility is illusory. In most cases, there is no way for a state to meet its goal without adopting all four Building Blocks and implementing them in line with EPA’s assumptions. 9 • The Proposed Guidelines conflict with key state laws and constitutions. There is no evidence of congressional intent to authorize EPA to preempt these state laws. • The selection between a rate-based and mass-based emission goal does not provide any flexibility to states. EPA should confirm that states have authority to perform their own conversions to a mass-based goal and have the flexibility of demonstrating that different methods of making this calculation are appropriate, as was suggested in the preamble to the Proposed Guidelines. The lack of flexibility in Proposed Guidelines’ goals should have become apparent to EPA after it calculated draft massbased goals for each state. • EPA generally describes four different types of state plans, but does not provide adequate guidance as to what each type should contain. EPA should recognize that states are free to design their plans as they see fit, and EPA cannot limit state discretion. • The timeframes established in the Proposed Guidelines are too short. Even with the potential extensions EPA has proposed, the timeframe for developing state plans is inadequate. Once plans are approved, states will have very little time to implement the extensive measures required by the rule. Further, most states will be required to take significant action by 2020 in order to comply, obviating the Proposed Guidelines’ purported flexibility. At a minimum, EPA should eliminate the interim goals. • The Proposed Rule suggests achievement demonstrations will require substantial modeling. EPA is not authorized to impose such a requirement. Moreover, EPA is not authorized to require state plans to be “self-correcting” or to impose additional emission reduction requirements as punishment for failure to achieve an emission goal. • EPA cites no authority to restrict the ability of states to modify their plans. If a modified plan will achieve a state goal, the modification should be allowed. • Many of the “key considerations” listed in the Proposed Guidelines for states to consider in their plans raise significant concerns. o EPA admits that the Proposed Guidelines will be challenging to implement and leaves it to states to determine if entities other than affected EGUs can be included in the program. EPA must resolve the issue of whether entities other than affected EGUs can be included in the program itself. o EPA should give the most favorable treatment possible to energy efficiency programs adopted prior to the proposal of the Proposed Guidelines and should ensure adequate recognition and credit is given to states that have already undertaken other significant measures that limit CO2 emissions. 10 o EPA’s complicated proposed methods for addressing interstate effects demonstrates that the Proposed Guidelines are untethered to the requirements of section 111. o EPA should allow states broad flexibility to use emission reductions from new natural gas combined cycle (“NGCC”) units, including incremental emission reductions beyond what is required of new units under section 111(b), to comply with the Proposed Guidelines. o EPA cannot limit state consideration of remaining useful life, and EPA has failed to take into account requirements the Agency has imposed under the Regional Haze Program and MATS. EPA cannot require states to regulate EGUs simultaneously under section 111(b) and section 111(d) of the CAA. • EPA cannot compel states to regulate an EGU as both an existing and new source. Under the CAA, sources are defined as either new or existing. The terms are mutually exclusive and determine which provisions of the Act apply. • It would be permissible, however, for a state that employs the rate-based approach under section 111(d) to allow the megawatt hours generated by a newly constructed unit to be included in the denominator for a state’s rate. • EPA’s policy reasons for trying to force states to regulate sources under both section 111(b) and section 111(d) are invalid. They do not authorize EPA to rewrite statutory requirements. EPA has failed to address significant New Source Review (“NSR”) issues presented by the Proposed Guidelines. • The projects contemplated in Building Block 1 to improve heat rates at EGUs have been targeted by EPA and others as triggering NSR. • EPA asserts incorrectly that there will be few instances when an NSR permit will be required. Although UARG agrees that measures taken to implement Building Block 1 do not trigger the Act’s NSR requirements, Attachments A and B to these comments demonstrate that such projects have been challenged as NSR violations by EPA’s enforcement arm, by states, and by environmental organizations. • EPA should eliminate the threat of protracted NSR litigation and provide a clear statement that any upgrades necessary to implement Building Block 1 of EPA’s Clean Power Plan do not trigger NSR. 11 The Proposed Guidelines are not justified by their purported benefits when those benefits are characterized properly. • EPA never connects the Proposed Guidelines to any impact on climate change, raising questions about the purpose of the rule. The Proposed Guidelines would, if fully implemented, control only about 1 percent of global emissions and do nothing to mitigate climate change. • EPA has significantly underestimated the costs of the Proposed Guidelines. Because the Building Blocks are based on unreasonable assumptions, they will be much more costly to implement than EPA assumes. EPA has also failed to accurately estimate the impacts of broad disruptions to the economy, energy security, and the reliability of the U.S. power system. • EPA’s cost-benefit analysis is based on the Proposed Guidelines’ purported global benefits, but a cost-benefit analysis for domestic policy should examine domestic costs and benefits alone. Analyzing the costs and benefits of the Proposed Guidelines on a domestic basis reveals compliance costs will exceed the purported benefits of EPA’s proposal. • EPA has improperly excluded higher discount rates recommended by the Office of Management and Budget (“OMB”) to inflate projected benefits. A proper analysis would likely reduce any estimated benefits to marginal levels. • EPA has improperly included co-benefits from reductions of pollutants other than CO2, such as fine particulate matter and ozone, that are regulated to levels protective of public health under other provisions of the CAA. EPA’s calculation of the state goals impermissibly limits the states’ discretion and is errorridden. • The Agency’s goal calculation methodology contains numerous errors, inconsistencies, and oversights that substantially impact each state’s goals. EPA has, for instance, applied its BSER Building Blocks to sources that cannot be regulated under section 111(d). EPA’s error would force affected NGCC units to operate at capacity factors significantly above 70 percent in order to accommodate the expected generation that states cannot require from non-affected EGUs. • EPA has failed to account for the electricity that the U.S. imports from non-emitting sources in Canada. If the Proposed Guidelines do not recognize the benefit provided by this imported power, the rule could be construed as an unlawful trade barrier. • In calculating state goals, EPA improperly assumed that NGCC units that were under construction in 2012 have available generating capacity. EPA has no basis on which to determine what generating capacity from these units will be available for redispatch. These units should not be included in EPA’s application of Building Block 2. 12 • EPA’s goal calculation methodology is defective because it applies the measures identified as BSER to sources that do not qualify as “affected EGUs” for the purposes of the Proposed Guidelines. EPA can, and must, examine historical data for existing EGUs in order to determine which units would be exempt from regulation under the one-third sales exclusion before calculating each state’s goal. • EPA’s 2012 data regarding existing and under-construction EGUs contain numerous errors and invalid assumptions that affect the Agency’s state goal calculations. EPA has: (1) improperly used nameplate capacity to calculate generation available for redispatch from NGCC units; (2) double-counted some NGCC units as both existing and under-construction; (3) identified some under-construction NGCC capacity that does not appear to exist; and (4) improperly counted as “existing” some EGUs that retired or began operation in 2012 or shortly thereafter. • In calculating the state goals, EPA unreasonably assumes that each state’s coal-fired units will emit CO2 at an average rate that is 6 percent lower than their average rate in 2012, and that NGCC units will emit CO2 at an average rate that is the same as 2012, even after generation from those EGUs is redispatched under Building Block 2. EPA’s Proposed Guidelines will in fact likely increase EGU heat rates. EPA’s assessment of the Building Blocks is seriously flawed. • Building Block 1, which envisions a 6 percent overall heat rate improvement is unlawful, unachievable, and unsupported by the record. EPA would require efficiency improvements in equipment that is not subject to regulation. o EPA’s reliance on a 2009 Sargent & Lundy report is misguided. For instance, EPA does not account for the fact that the efficiency benefits associated with the measures identified in the report are highly variable by unit, are not cumulative, and are only temporary. EPA has also overlooked the negative impact on coal-fired EGU efficiency of EGUs’ obligations under the remaining components of its proposed BSER and under other CAA programs. o EPA’s “model” assessing potential heat rate improvements via “best practices” is unfounded, arbitrary, and misleading. EPA does not actually assess the feasibility of a 6 percent heat rate improvement requirement, and EPA has failed to account for unavoidable heat rate variability. o EPA’s assessment of available heat rate improvements through “equipment upgrades” is arbitrary. EPA erroneously assumed that the heat rate improvements from these upgrades are cumulative and that they provide consistent long-term benefits. o EPA’s analysis of heat rate improvements at 16 EGUs is deeply flawed. The reported improvements resulted almost exclusively from changes in continuous emission monitoring system (“CEMS”) reporting methodology and were not the result of purposeful efforts to improve unit efficiency. 13 • Building Block 2, which envisions a shift in generation from existing coal- and oil/gas-fired steam EGUs to existing NGCC units until those NGCC units reach a statewide maximum capacity factor of 70 percent, is unachievable. o EPA has miscalculated the amount of NGCC generating capacity in each state. o EPA has not accounted for factors limiting the ability of EGU owners to increase utilization of their NGCC units, including technical limitations, permit limits, or gas and transmission infrastructure constraints. o EPA has assumed that all NGCC units can meet a 70 percent capacity factor without assessing the differences between high- and low-capacity factor NGCC units. o EPA’s assessment of historic trends does not justify its claim that sufficient infrastructure exists to support Building Block 2. o EPA has failed to consider the effect of widespread redispatch to NGCC units on states’ ability to meet demand during periods of peak load. o EPA’s Building Block 2 is based on the faulty premise that generation from any NGCC unit located within a state may be redirected by a utility or state within the state’s borders to offset reduced generation from coal-fired units elsewhere in the state. In most states, electric generation is dispatched by an RTO or ISO, which calls up generation from EGUs within its system such that the lowest cost generating sources are dispatched first. Even where electricity dispatch is governed by individual utilities, similar limitations on in-state redispatch arise where a state is served by multiple utilities or where a utility spans multiple states. o EPA does not appear to understand how the natural gas system interfaces with the electric system and how that interface impacts reliability. • The renewable energy generation goals that EPA has calculated for each state in the Proposed Guidelines and applied in Building Block 3 are unachievable. o EPA’s proposed approach is arbitrary because it does not take each state’s actual renewable energy growth potential into account. o EPA’s own Integrated Planning Model (“IPM”) results indicate that renewable energy generation by the end of the compliance period cannot even come close to the renewable energy targets EPA used in its goal calculation. o EPA simplifies its analysis of renewable portfolio standards (“RPS”) in a manner that inflates renewable energy generation targets. For instance, EPA incorrectly assumes that each state’s most stringent primary goals apply to all in-state generation and treats all goals as mandatory. 14 o EPA’s proposed rule would require states to begin implementing the renewable energy portions of their plans in 2017, before many states have even received plan approval from EPA. • By including nuclear capacity that is under construction or “at risk” in state goals, EPA merely penalizes states that have invested in nuclear generation. • UARG agrees that natural gas conversion and co-firing are not BSER for coal-fired EGUs and should not be included in determining the proposed emission guidelines. EPA’s October 30, 2014 Notice of Data Availability (“NODA”) does not contain any specific EPA proposals or proposed regulatory language and for the most part cannot be included in any final guidelines. • The NODA contains virtually no “data,” and is rather a series of questions and musings about different policies the Agency could employ. The content in this document is more appropriate for an advance notice of proposed rulemaking. • EPA provided insufficient time to allow stakeholders to provide fully informed and well-considered comments on the NODA. The Proposed Guidelines are extraordinarily complex, and the issues raised in the NODA are broad and seem to contain some approaches that are either new or that contradict its previous positions. • EPA asks for comment on how Building Block 2 could be expanded to include new NGCCs and natural gas co-firing at existing coal-fired boilers. This is not formally proposed in the NODA and EPA cannot finalize this approach. Moreover, EPA has no authority under section 111(d) to impose federally enforceable obligations on new sources or to require the construction of a new source. The suggestion that natural gas co-firing can be designated as BSER is wrong as both a matter of fact and of rulemaking procedure. • If the final emission guidelines contain interim goals, which UARG does not support, then they should be flexible and allow states to comply using a reasonable “glide path.” Each state should have the authority to define its own glide path using criteria such as costs, energy impacts, and remaining useful life. Allowing the interim goals to be met using a glide path would be a logical outgrowth of the Proposed Guidelines. States should be allowed to determine the timing and to have a gradual phase-in of Building Blocks 1 and 2. • The regionalized approach to Building Block 3 and the regional approach to redispatch under Building Block 2 were not included in the Proposed Guidelines and are not actually proposed in the NODA. There is insufficient information such as technical analyses and the other items required by section 307(d)(3) of the CAA that would allow UARG to comment on these approaches in a meaningful way. • The NODA solicits comment on EPA’s goal-setting methodology but does not actually propose this change in the NODA, and EPA cannot finalize it as part of any 15 final emission guidelines. UARG opposes this approach for three reasons: (1) EPA has no authority to establish a standard of performance based on reduced utilization or retirement of a source; (2) if EPA wants to pursue such a dramatic change, it must undertake a thorough evaluation of costs and other implications, as required by section 307(d)(3), and allow UARG sufficient time to evaluate and comment on those studies; and (3) this is a nonsensical approach to EPA’s overarching goal of reducing fossil fuel-fired generation because it would have the counterintuitive effect of encouraging the construction of new fossil fuel-fired EGUs. EPA must revise the Proposed Guidelines to address problems with the monitoring, recordkeeping, and reporting provisions for EGUs. • EPA’s proposed regulatory language regarding EGU monitoring, recordkeeping, and reporting provisions in some cases is overly broad or duplicative. • UARG objects to the proposed elimination of one of the Part 75 monitoring options and the proposed imposition of additional measurement requirements on existing units. EPA has provided no justification for requiring EGUs to perform monitoring that is different from, let alone more stringent than, Part 75. The proposed requirement to measure stack diameter at three distinct locations may not even be possible without installing new sampling ports. UARG also objects to the Agency’s failure to make clear that use of Part 75 missing data substitution procedures is not appropriate in compliance calculations. • EPA proposes that affected EGUs also could report hourly net electric generation under Part 75 via an Emission Collection and Monitoring Plan System (“ECMPS”). It is not clear whether EPA is proposing that Part 75 be revised to simply provide a mechanism for such reporting in the event a state chooses to require or allow it, or whether EPA is suggesting that source reporting under Part 75 be required to report net electric generation. UARG would oppose the latter requirement. • For integrated equipment that is not monitored under Part 75 or an NSPS, UARG suggests that EPA include a provision allowing submission and approval of a petition for exemption from the compliance calculation of any de minimis CO2 emissions, or for an alternative means of accounting for any resulting emissions. Because use of such equipment, which either has very few emissions or operates very infrequently, is uncertain, the Agency should not spend resources specifying monitoring requirements in the rule. • If EPA intends sources to include partial unit operating hours in compliance calculations, the output values already being reported under Part 75 (in megawatts (“MW”) or steam load) will need some adjustment for use in EPA’s proposed compliance calculations. EPA should clarify whether it intends that partial hours be included in state calculations. And, if EPA anticipates that states rely on data reported to ECMPS, EPA must address the discrepancies. 16 III. • EPA has not justified the proposed burdensome requirement to comply with the American National Standards Institute (“ANSI”) 2010 C12.20 standard for electricity meters in this rulemaking. • UARG agrees that if EPA requires the monitoring of net electric output, the Agency must allow EGUs to use existing equipment and methods for metering station service, and to apportion common station service to individual units using unit generation. • Combined heat and power (“CHP”) facilities should not be required to employ any specific technology or quality assurance (“QA”) procedures for equipment used to measure useful thermal output. Appropriate technologies and procedure vary according to application, and any gains in accuracy from employing different technologies or procedures likely would be insignificant relative to other uncertainties in EPA’s calculation of creditable output. • UARG does not believe that there is any benefit to including equipment used to measure electric and energy output in the Part 75 monitoring plan or to requiring additional quality assurance and quality control (“QA/QC”) for that equipment. EGUs already have sufficient incentives to ensure that their equipment for measuring output is accurate, and EPA has not provided any information to suggest that additional QA is warranted. • EGUs should be allowed to choose which version of the flow method is best for the particular application and should not be required to perform 3-D testing where it is not needed. With respect to EPA’s suggestion that it might develop “adjustment factors” that would be applied if an EGU changed flow methods, EPA has not provided sufficient information for UARG to comment on the reasonableness of such a requirement. As a result, UARG opposes it. To proceed with that approach, EPA would need to issue a more complete rulemaking proposal. • UARG opposes the proposed ten-year record retention period for supporting information, unless EPA can show a special need as required under Paperwork Reduction Act regulations. UARG also opposes the proposed on-site requirement if it means that electronic records that can be accessed “onsite” (even if they are physically located offsite) would not be sufficient. The Proposed Guidelines Are Fundamentally Divorced from the Narrow and Exclusive Focus of Section 111 of the CAA on Stationary Sources in Listed Source Categories. Section 111 of the CAA speaks with absolute clarity as to the subject and nature of standards of performance for new and existing sources. A standard of performance under section 111 must be achievable for individual sources based on measures that the source’s owner can integrate into the design or production process of the source itself. Section 111 has existed, with 17 only minor changes, for over 40 years, and in that time EPA has applied it with total consistency on this fundamental point. A standard of performance cannot be based on actions taken beyond the source itself that somehow reduce the source’s utilization. Nor can it be based simply on requiring a source to reduce its operations. Yet EPA is proposing a rule that departs so far from the Act’s clear directions as to be unrecognizable as a section 111 rulemaking. The Agency’s proposed approach has inverted the lawful order of standard-setting under section 111. Instead of defining a category of regulated sources, identifying the BSER that any individual source can incorporate into its design, and then allowing states to determine what standards are achievable for sources based on that BSER and other factors, EPA has started at the end by defining an inflexible emissions goal for each state and then requiring states to impose whatever obligations are necessary to achieve that goal. To effectuate this unprecedented approach, EPA proposes to base its “emission guidelines” on actions that reach beyond individual regulated sources and impose obligations on entities with no emissions at all, or to simply mandate that regulated sources curtail operations or cease operations altogether. EPA’s proposed approach has no basis in the statute. Indeed, it defies reason and is flatly inconsistent with the CAA and with the Agency’s own 40-year history of consistently contrary implementation. EPA has exceeded its authority and must withdraw the Proposed Guidelines. A. Section 111 Authorizes Standards of Performance That Are Achievable for Individual Sources in a Source Category Based on Measures Those Sources Can Implement Themselves. To fully appreciate how dramatically EPA has departed from its authority in this rulemaking, one must first understand the nature and limits of the CAA’s section 111 regulatory program as it has been defined by the text of that provision, by its context within the remainder of the CAA, and by 44 years of consistent EPA implementation. This program begins and ends 18 with the regulated source itself. It provides for the regulation of individual emission sources through performance standards that are based on what design or process changes an individual source’s owner can integrate into its facility. The nature of any standard of performance under section 111 is that, at a given level of operations, the regulated source emits less when subject to the standard than when not. 1. The Text of Section 111 Section 111’s regulatory program is narrow, limited, and ultimately quite simple. Under section 111(b), EPA is to publish “a list of categories of stationary sources” that “cause[], or contribute[] significantly to, air pollution which may reasonably be anticipated to endanger public health or welfare.” CAA § 111(b)(1)(A). Once EPA lists a source category, the Agency “establish[es] Federal standards of performance for new sources within such category.” Id. § 111(b)(1)(B). Meanwhile, under section 111(d), EPA is also directed to establish a procedure . . . under which each State shall submit to the Administrator a plan which . . . establishes standards of performance for any existing source for any air pollutant . . . to which a standard of performance under [section 111] would apply if such existing source were a new source. Id. § 111(d)(1). In other words, after EPA promulgates an NSPS for new sources in a category under section 111(b), EPA may require states to develop plans adopting standards of performance that apply to existing sources in that category under section 111(d). Under EPA’s Subpart B regulations, which implement section 111(d), the Agency publishes emission guidelines to inform the states’ development and submittal of these state plans. See 40 C.F.R. §§ 60.20-60.29. The statute adopts the same definition of “standard of performance” for both new and existing sources: a standard for emissions of air pollutants which reflects the degree of emission limitation achievable through the application of the best system of emission 19 reduction which (taking into account the cost of achieving such reduction and any nonair quality health and environmental impact and energy requirements) the Administrator determines has been adequately demonstrated. CAA § 111(a)(1). Thus, the Administrator’s analysis for establishing standards of performance for new sources will inform a state’s development of a plan establishing performance standards for existing sources. The actual standards for existing sources in a category will typically differ, however, from the new source standards developed by EPA due to the disparate costs, energy impacts, and other factors associated with regulating each. See 40 Fed. Reg. 53,340, 53,341 (Nov. 17, 1975) (promulgating and explaining Subpart B). In addition, the CAA directs states (and EPA, in the case of a federal plan) to consider, “among other factors, the remaining useful life of the existing source to which such standard applies” when adopting standards of performance for existing sources. CAA § 111(d)(1) (emphasis added). On its face, section 111 provides only for standards that regulate the emissions performance of individual stationary sources in a category listed under section 111(b). The plain text of this section makes it clear that standards of performance apply to sources within listed categories as individual facilities and do not regulate categories or subcategories as a whole: NSPS apply only to “new sources within [a listed] category,” while state standards under section 111(d) apply to “any existing source . . . to which a [NSPS] . . . would apply if such existing source were a new source.” Id. § 111(b)(1)(B), (d)(1). Further, the Act narrowly confines the stationary sources that may be regulated under section 111 to any individual “building, structure, facility, or installation which emits or may emit any air pollutant.” Id. § 111(a)(3). This definition notably does not extend to combinations of these facilities or to other non-emitting entities. See ASARCO, 578 F.2d at 319. In addition, section 111(d) explicitly directs states (and EPA, in the case of a federal plan) to consider the “remaining useful life” of existing sources 20 when applying any standard of performance, further demonstrating that this section focuses solely on what individual sources can do to improve their performance at reasonable cost rather than what the entire source category (or even other entities) can do collectively to reduce overall emissions. CAA § 111(d)(1), (d)(2). Even the titles used in these provisions indicate their narrow focus on individual sources. Section 111 is titled “[s]tandards of performance for new stationary sources,” and section 111(d) is designated “[s]tandards of performance for existing sources; remaining useful life of source.” Id. § 111, 111(d). Importantly, section 111 also requires that any standard of performance be “achievable” by the sources to which it applies based on application of an “adequately demonstrated” system of emission reduction. Id. § 111(a)(1). The achievability requirement clearly indicates that Congress intended standards of performance to be based on systems of emission reduction that are within the control of (and thus, incorporated into the design or production process of) an individual source. A standard cannot be “achievable” for a source if the source must rely on measures taken by some other entity that it does not control, or if the source must not operate, in order to achieve the standard. Moreover, a “standard of performance” by its terms assumes that the source will continue to “perform.” A “standard of performance” cannot mandate “nonperformance.” This focus on measures incorporated into individual sources pervades other parts of section 111 as well. For example, section 111(h) authorizes EPA to promulgate a design, equipment, work practice, or operational standard in cases where “it is not feasible to prescribe or enforce a standard of performance,” and defines exactly when Congress considered it “not 21 feasible” to establish a standard of performance. 4 Id. § 111(h)(1). One such situation is where the regulated pollutant “cannot be emitted through a conveyance designed and constructed to emit or capture such pollutant.” Id. § 111(h)(2)(A). This provision equates a “standard of performance” with the use of a conveyance at the regulated source to capture and to control a portion of the source’s emissions. 2. CAA Context for Section 111 The CAA’s other provisions confirm the narrow scope of what and how EPA may regulate under section 111. Nothing in the remainder of the Act authorizes EPA to regulate emissions from stationary sources by basing standards of performance on measures that are not implemented by the regulated source itself. Likewise, there is no provision of the CAA under which the Agency may base a standard of performance on reduced operations. Standards of performance cannot be based on enforceable limitations on hours of operation or on electricity production. 4 EPA invites comment on whether it can include in its BSER determination (and allow state plans to require) “design, equipment, work practice, or operational standard[s]” under its section 111(d) program. 79 Fed. Reg. at 34,926-27. If EPA finalizes the Proposed Guidelines as a conventional section 111(d) program, it may not. Congress specifically excluded these methods from the definition of “standard of performance” under section 111(a)(1) and relegated them under section 111(h) as an “alternative” in instances where standards of performance are not “feasible.” CAA § 111(h)(1) (“[I]f in the judgment of the Administrator, it is not feasible to prescribe or enforce a standard of performance, [s]he may instead promulgate a design, equipment, work practice, or operational standard, or combination thereof . . . .”) (emphasis added). If EPA decides to promulgate work practice or operational standard-based emission guidelines for existing fossil fuel-fired EGUs at some point in the future (after notice-andcomment rulemaking in a separate proceeding), then it must make the section 111(h) findings that are preconditions to setting a section 111(h) standard and then adhere to the requirements of section 111(h), which mandate that a section 111(h) standard be feasible, reflect the best technological system of continuous emission reduction, be cost-effective, and adequately demonstrated. See id. § 111(h). 22 The CAA’s other programs establishing emission standards for new and existing sources focus solely on achieving emission reductions at individual sources. These programs regulate emissions from sources based on what other individual sources can or already have achieved. For example, emission standards for hazardous air pollutants must be based on the maximum achievable control technology and reflect the application of “measures, processes, methods, systems or techniques” directly to individual sources. 5 Id. § 112(d)(2). Standards for visibilityimpairing pollutants must reflect the “best available retrofit technology . . . for controlling emissions from [each eligible] source,” considering the costs, existing control technology, and remaining useful life for that source. Id. § 169A(b)(2)(A). And under the CAA’s PSD program, new and modified sources must implement the “best available control technology” (“BACT”), which the permitting authority must identify on a case-by-case basis for each source and which must reflect “application of production processes and available methods, systems, and techniques” at the source. Id. §§ 165(a)(4), 169(3). Although a source may accept operational limits in its preconstruction permit under the PSD provisions in order to avoid qualifying as a “major emitting facility” and having to implement BACT, under no circumstances may the BACT standard itself be based on reduced utilization of the unit. See id. § 169(3), 40 C.F.R. § 52.21(b)(12). The BACT program is particularly relevant as evidence of section 111’s scope because Congress explicitly tied the BACT emission standards to section 111. Standards of performance 5 These measures include those that: “(A) reduce the volume of, or eliminate emissions of, such pollutants through process changes, substitution of materials or other modifications, (B) enclose systems or processes to eliminate emissions, (C) collect, capture or treat such pollutants when released from a process, stack, storage or fugitive emissions point, (D) are design, equipment, work practice, or operational standards (including requirements for operator training or certification) . . ., or (E) are a combination of the above.” CAA § 112(d)(2). Notably, all of these measures can be implemented at an individual source. 23 under section 111 provide a regulatory floor for BACT standards. CAA § 169(3). But if a standard of performance relies on a “system of emission reduction” that goes beyond the source itself, it cannot meaningfully inform a BACT standard for sources in that category. In contrast, in the few regulatory programs where Congress did authorize emission control measures that go beyond a specific source for the purpose of meeting aggregate emission reduction goals, it spoke clearly and precisely. For example, when Congress took action in the 1990 CAA Amendments to cap sulfur dioxide emissions and establish a program for emissions allowances and trading, it added an entirely new title to the Act (Title IV) spelling out the requirements and implementation procedures for that program in great detail. See id. §§ 401416. Unlike Title I of the CAA, in which section 111 is found, Congress’s statement of purpose in Title IV specifically includes “encourag[ing] energy conservation, use of renewable and clean alternative technologies, and pollution prevention as a long-range strategy, consistent with the provisions of this subchapter, for reducing air pollution.” Id. § 401(b). Congress also gave EPA explicit instructions on how to credit sources for compliance with emission requirements based on avoided emissions from renewable energy and energy conservation. Id. § 404(f). Section 111 contains no such instructions. Likewise, the CAA’s provisions regarding implementation of the national ambient air quality standards (“NAAQS”) authorize states to take broad measures to attain and maintain compliance with those standards, and even contemplate that some existing stationary sources may need to reduce operations or retire in order for an area to achieve those standards. See id. § 110. But in establishing the NAAQS program, Congress explicitly authorized EPA to promulgate health-based standards for ambient concentrations of certain pollutants at a level “requisite to protect the public health” with “an adequate margin of safety.” Id. § 109(b). And 24 Congress directed states to enact any “emission limitations and other control measures, means, or techniques (including economic incentives such as fees, marketable permits, and auctions of emissions rights) . . . as may be necessary or appropriate to meet” these standards, without singling out for regulation any specific sources of the regulated pollutants. Id. § 110(a)(2)(A). In contrast, section 111 regulates emissions from specific sources, not overall concentrations or amounts of pollutants. Its technology-based standards of performance are driven by what emission limitation those sources can achieve during their performance at reasonable cost, using adequately demonstrated measures, rather than by requirements dictated by public health or welfare. 3. The History of EPA’s Implementation of Section 111 EPA’s long and consistent history of implementing section 111 confirms the plain language of the statute and has also given shape to this regulatory program. The Agency’s past rulemakings reflect the program’s singular focus on individual sources. In fact, in the 44-year history of the Act, EPA has limited the scope of section 111 to the regulated source in every rulemaking it has undertaken. First, the Agency’s 1975 Subpart B regulations—in which EPA established a procedural framework for states to adopt standards of performance for existing sources—share the statute’s exclusive focus on standards that are achievable by individual sources. Subpart B directs EPA to publish a “guideline document containing information pertinent to control of the designated pollutant form [sic] designated facilities [i.e., existing sources subject to regulation under section 111(d)].” 40 C.F.R. § 60.22(a) (emphasis added). Echoing the statutory text, emission guidelines under Subpart B must “reflect[] the application of the best system of emission reduction (considering the cost of such reduction) that has been adequately demonstrated for designated facilities.” Id. § 60.22(b)(5) (emphasis added). Acknowledging section 111’s 25 statutory command to consider the “remaining useful life” of regulated existing sources, Subpart B also notes that states may tailor standards of performance for individual designated facilities to account for “[u]nreasonable cost of control resulting from plant age, location, or basic process design,” “[p]hysical impossibility of installing necessary control equipment,” or “[o]ther factors specific to the facility (or class of facilities) that make application of a less stringent standard or final compliance time significantly more reasonable.” Id. § 60.24(f). In adopting this provision, EPA noted that such flexibility to depart from emission guidelines was necessary because “controls cannot be included in the design of an existing facility and because physical limitations may make installation of particular control systems impossible or unreasonably expensive in some cases.” 40 Fed. Reg. at 53,344. This discretion reflects Subpart B’s focus on identifying what emission reductions individual existing sources can achieve themselves. Subpart B also specifies that compliance with any state standard of performance for existing sources will be shown through a series of “[i]ncrements of progress,” which are “steps to achieve compliance which must be taken by an owner or operator of a designated facility.” 40 C.F.R. § 60.21(h). These increments of progress include awarding contracts, initiating on-site construction or installation, and completing on-site construction or installation of emission control equipment or process changes. Thus, Subpart B makes clear that compliance with standards of performance is achieved through on-site measures taken by regulated sources. Second, out of the nearly 100 NSPS and emission guidelines EPA has promulgated and subsequently revised since 1970, every single standard of performance has been based on a “system of emission reduction” that is incorporated into the design or operation of individual sources. For example, when the Agency promulgated and later revised NSPS for kraft pulp mills, it never considered basing the standard of performance on requiring increased use of 26 recycled paper to reduce kraft pulp mill operations, even though such a measure arguably would have reduced emissions from kraft pulp mills. See 43 Fed. Reg. 7568 (Feb. 23, 1978); 79 Fed. Reg. 18,952 (Apr. 4, 2014). Likewise, on each past occasion in which EPA has promulgated or revised NSPS for the categories of EGUs at issue in this proposed rulemaking—a timeframe spanning from 1971 to 2012—the Agency’s analysis has focused entirely on the use of emission control technology or low-sulfur fuels at individual sources. 6 Despite pervasively regulating EGUs under section 111 over the past forty years, EPA has never before even considered basing a standard of performance for emissions from these sources on reduced utilization or shifting generation to lower-emitting sources. Nor has EPA’s focus on individual sources changed in recent years. In an NSPS rulemaking that took place just weeks after the Proposed Guidelines were published, EPA once again reaffirmed that standards of performance “apply to sources” and must be “based on the BSER achievable at that source.” 79 Fed. Reg. 36,880, 36,885 (June 30, 2014) (emphasis added). Because of the nature of section 111(d), EPA has conducted very few rulemakings under that provision. But on the few occasions that EPA has issued emission guidelines for existing sources, it has maintained the same focus on measures that the regulated source can incorporate into its design or otherwise implement by itself. Adopting the same approach is not only appropriate but required, given that the CAA defines “standard of performance” identically for new and existing sources. Since 1970, EPA has only published valid emission guidelines under section 111(d) for five source categories, and in all five of these rulemakings the emission 6 See, e.g., 36 Fed. Reg. 24,876 (Dec. 23, 1971) (promulgating Subpart D); 44 Fed. Reg. 33,580 (June 11, 1979) (promulgating Subpart Da); 44 Fed. Reg. 52,792 (Sept. 10, 1979) (promulgating Subpart GG); 63 Fed. Reg. 49,442 (Sept. 16, 1998) (amending Subpart Da); 71 Fed. Reg. 38,482 (July 6, 2006) (promulgating Subpart KKKK); 77 Fed. Reg. 9304 (Feb. 16, 2012) (amending Subpart Da). 27 guidelines were based on the application of pollution control technology or other process controls at individual sources. 7 Even EPA’s short-lived CAMR under section 111(d) did not adopt a broader approach to establishing standards of performance. Although CAMR did authorize an emissions trading program as a tool for compliance with standards of performance, the “system of emission reduction” that was used to set the emission guidelines themselves was limited to pollution control technology that could be installed at individual sources. 70 Fed. Reg. at 28,617-20, 28,621 (final guideline was “based on the level of Hg emissions reductions that will be achievable by the combined use of co-benefit (CAIR) and Hg-specific controls”). In sum, the plain text of section 111 of the CAA establishes a program that is focused on reducing the rate of emissions from new and existing stationary sources through the application of systems that can be integrated into the design or operation of the source itself. The overall context of the CAA and EPA’s unwavering commitment to this approach outside of the present rulemaking confirm this focus. The notion that a source must be able to comply with a standard of performance through measures under its own control is so fundamental to section 111 that it pervades every aspect of that provision’s language and history of implementation. 7 41 Fed. Reg. 19,585 (May 12, 1976) (guidelines for phosphate fertilizer plants based on “spray cross-flow packed scrubbers”); 41 Fed. Reg. 48,706 (Nov. 4, 1976) (guidelines for sulfuric acid production units based on “fiber mist eliminators”); 43 Fed. Reg. 7597 (Feb. 23, 1978) (guidelines for kraft pulp mills based on process controls and two-stage black liquor oxidation system); 45 Fed. Reg. 26,294 (Apr. 17, 1980) (guidelines for primary aluminum plants based on “effective collection of emissions, followed by efficient fluoride removal by dry scrubbers or by wet scrubbers”); 61 Fed. Reg. 9905, 9907 (Mar. 12, 1996) (guidelines for municipal solid waste landfills based on “(1) [a] well-designed and well-operated gas collection system and (2) a control device capable of reducing [non-methane organic compounds] in the collected gas by 98 weight-percent”). 28 B. EPA’s Proposed Action Deviates So Far from the CAA as To Be Unrecognizable as an Exercise of Section 111 Authority. The broad program of energy resource management and economy-wide demand reduction that EPA proposes to create and administer in this rule bears no resemblance to the source-focused regulatory program that Congress established in section 111 and that EPA has consistently implemented over the past four decades. In the Proposed Guidelines, the Agency has defied precedent and logic by rewriting some parts of the Act and completely ignoring others in order to embark on a “multiyear voyage of discovery” that far exceeds EPA’s lawful authority. See UARG v. EPA, 134 S. Ct. at 2446. Indeed, as noted in the introduction to this section of UARG’s comments, EPA’s proposal would invert the Act’s regulatory process under section 111 for controlling emissions from existing sources. In the lawful course of regulation, reducing emissions from existing sources under section 111(d) would proceed such that: (1) EPA specifies a source category and identifies achievable source-based performance measures that can reduce sources’ emissions through design or process changes; (2) the state determines which measures are appropriate for particular sources based on cost, remaining useful life, and other factors; (3) individual regulated sources implement those measures based on standards of performance promulgated by states; and (4) emission reductions result from individual sources operating at a reduced emission rate. Yet EPA’s Proposed Guidelines would begin at the end, by mandating a specific level of CO2 emissions from each state at the outset, and only deviate farther from the Act from there. EPA’s approach is that: (1) the Agency specifies a permissible CO2 emissions rate for each state expressed in lb/MWh; (2) the state determines what entities—whether EGUs or other commercial, institutional, corporate, or governmental entities—it will subject to enforceable obligations; and (3) the state requires a broad and undefined range of emission-reducing or 29 operational measures at those entities as necessary to achieve the desired emissions level. This backwards approach plainly exceeds the Agency’s authority under the CAA. Nearly every aspect of the Proposed Guidelines is inconsistent with the Act and with EPA’s binding legislative rules. In particular, the Agency’s radical redefinition of what measures may constitute a “system of emission reduction” is plainly impermissible in light of section 111’s exclusive focus on what individual sources can achieve and the past four decades of contrary implementation. Likewise, EPA’s claim that it may apply BSER to a state as a whole rather than to individual sources in a source category lacks any legal merit. And, finally, the fact that the Proposed Guidelines’ required emission reductions cannot be enforced without dramatically expanding the universe of entities subject to obligations under section 111 underscores the extent to which EPA has strayed from the scope of regulation permitted under the CAA. Together, these defects reflect the Agency’s disregard for the statutory and administrative bounds of its own authority. The Proposed Guidelines “would bring about an enormous and transformative expansion in EPA’s regulatory authority without clear congressional authorization,” and therefore it must be withdrawn. Id. at 2444. 1. EPA’s Proposed Guidelines Impermissibly Base Standards of Performance on Measures that Go Beyond the Regulated Source. Perhaps the most fundamental departure from the law in EPA’s Proposed Guidelines is the Agency’s assertion that the “best system of emission reduction” for the sources in a designated source category under section 111 may include measures that would (either directly or indirectly) reduce a source’s utilization and that are not within the control of individual sources. Of the four “building blocks” that make up EPA’s proposed statewide BSER for existing EGUs, only Building Block 1 (heat rate improvements at coal-fired EGUs) falls within the scope of measures contemplated in the Act and could therefore provide the foundation for a 30 legally defensible emission guideline. The others—mandating redispatch of generation from coal-fired units to NGCC units, displacing generation from affected EGUs with generation from renewable energy sources, and reducing electricity demand through energy efficiency measures—all impermissibly rely on measures that go beyond the boundaries of individual affected EGUs and that are not within the control of individual EGU owners and operators. The measures in Building Blocks 2, 3, and 4 are all based on implementation of statewide energy policies that would confiscate the available generating capacity of existing EGUs. Indeed, the Proposed Guidelines impose undefined regulatory obligations on a broad swath of unspecified “affected entities” in addition to the fossil fuel-fired EGUs that are the source category for this rule. Many of these “affected entities” do not emit CO2 in any appreciable amount. This “beyond-the-source” approach would allow EPA to restructure every aspect of the states’ electric power markets and regulate any electricity user—effectively, to administer the entire national economy—for the purposes of reducing demand for and generation by sources in the listed source category (i.e., existing fossil fuel-fired EGUs). This is impermissible and would transform section 111 into something untethered to its statutory language and unrecognizable to the Congress that created it. The genesis of this novel claim to regulatory power lies in the Agency’s attempt to rewrite what constitutes a “system of emission reduction” for the purposes of section 111. As noted above, under that section, a standard of performance must “reflect[] the degree of emission limitation achievable through the application of the best system of emission reduction” that has been adequately demonstrated for sources in the regulated category. CAA § 111(a)(1) (emphasis added). In this rulemaking, however, EPA “claims to [have] discover[ed] in [this] long-extant statute an unheralded power to regulate ‘a significant portion of the American economy,’” a 31 practice that courts have looked upon skeptically in the past. UARG v. EPA, 134 S. Ct. at 2444. EPA asserts here that because the word “system” is not explicitly defined in the CAA, the Agency may freely apply that word’s abstract dictionary definition: “‘a set of things working together as parts of a mechanism or interconnecting network; a complex whole.’” Legal Memorandum for Proposed Carbon Pollution Emission Guidelines for Existing Electric Utility Generating Units at 51 (undated) (emphasis added), Docket ID No. EPA-HQ-OAR-2013-06020419 (“EPA Legal Memorandum”). EPA applies this definition in the abstract to conclude that a “system of emission reduction” can be “virtually any ‘set of things’ that reduce emissions,” including anything from “add-on controls . . . to measures that replace production or generation at the affected sources.” Id. at 51-52 (emphasis added). The Agency even claims that it may require “reduced utilization” of a source as part of a “system of emission reduction,” presumably including a prohibition on utilizing a specific regulated source. Id. at 79; 79 Fed. Reg. at 34,889. The breadth of EPA’s unprecedented assertion of authority is staggering, particularly in light of section 111’s singular focus on regulating individual sources of emissions. Under EPA’s asserted definition of a “system of emission reduction,” standards of performance for sources in a regulated source category would become a mere pretext for imposing a wide range of demandreducing obligations on countless entities across the entire nation. Under EPA’s interpretation, EPA would be able to require any “affected entity” to implement any “set of things” that the Agency believes would reduce the operation of sources in the listed source category—and hence emissions from that category—no matter how far removed the required actions are from the source itself. In the context of EGUs, these measures could include anything from capping annual generation from certain EGUs, to ordering the shutdown of some EGUs or other affected entities 32 altogether, to regulating how individuals use electricity or consume other goods and services that require electricity. Applying this “beyond-the-source” approach to other source categories—as this rulemaking suggests EPA is likely to attempt—leads to similarly expansive outcomes. For example, this approach could lead EPA to adopt standards of performance for kraft pulp mills that are based on efforts to reduce demand for new paper, such as requiring office buildings to implement paper recycling programs or encouraging credit card companies to provide paperless billing to customers. Or it could allow EPA to adopt standards of performance for oil and gas production facilities that are based on reducing demand for gasoline, such as mandating carpooling programs or requiring additional state investments in public transportation infrastructure. These beyond-the-source measures are inconsistent with the regulatory program Congress provided for in section 111 of the CAA. Although the dictionary definition of “system,” if considered in the abstract and devoid of any statutory or historical context, might theoretically embrace statewide, regional, or even national reduction programs, the word as used in section 111 can refer only to reductions resulting from measures that are incorporated into the source itself. EPA’s claim that nothing in the language or context of section 111 limits the Agency’s expansive redefinition of “system of emission reduction” is patently false, and it suggests that EPA has ignored the statute and its own past rulemakings. See EPA Legal Memorandum at 51-52. Indeed, section 111’s clear focus on measures that are achievable by the regulated source itself pervades every aspect of that provision’s language, its statutory context, and EPA’s legislative rulemakings implementing that provision. As described above in Section III.A, section 111 is a narrow program designed to improve the emissions performance of new and existing sources in specific categories based on the application of achievable measures 33 implemented in the design or production process of the source at reasonable cost. EPA completely ignored these factors when addressing the meaning of the word “system” as it is used in section 111, and in developing its proposed BSER. Building Blocks 2, 3, and 4 of the proposed BSER are based on beyond-the-source measures that cannot be achieved by individual sources, and accordingly, they cannot be used to support a rulemaking under section 111 of the CAA. As noted above, section 111 requires that emission guidelines and standards of performance be “achievable” by “any existing source” in the regulated category—not by only some sources, by a state, or by the category in the aggregate or in combination with other “entities.” An emission guideline that requires regulated sources to obtain emission reductions offsite in order to comply with an otherwise unachievable standard would unambiguously contravene the plain text of the CAA. Yet EPA failed to acknowledge how the achievability requirement limits the systems of emission reduction that may provide the basis for section 111 regulation. Building Blocks 2, 3, and 4 of the proposed BSER require measures that go beyond individual existing EGUs. The owner of an individual coal-fired utility boiler cannot control the dispatch of NGCC units relative to other fossil fuel-fired EGUs. Nor can it make changes in the design or production process of its boiler to generate renewable energy or to lead consumers to use less electricity. Therefore, Building Blocks 2, 3, and 4 cannot support an achievable emission guideline. Indeed, when EPA has attempted in the past to group individual sources together for the purposes of applying standards of performance, these attempts have been soundly rejected by the courts. See ASARCO, 578 F.2d at 327. In ASARCO, the D.C. Circuit invalidated EPA’s attempt to “change the basic unit to which the NSPSs apply from a single building, structure, facility, or 34 installation—the unit prescribed in the statute—to a combination of such units,” holding that EPA “has no authority to rewrite the statute in this fashion.” Id. (emphasis in original). The Agency must be “guided by a reasoned application of the terms of the statute it is charged to enforce, not by an abstract ‘dictionary’ definition.” Id. at 324 n.17. If the Agency is concerned about the need for flexibility in order for sources to comply with a standard, the solution is to alter the stringency of the standard rather than the scope of the source to which that standard applies. Id. at 328-29. Moreover, EPA failed to acknowledge the other ways in which the language and overall statutory context of section 111 render its beyond-the-source approach unreasonable. It is a “fundamental canon of statutory construction that the words of a statute must be read in their context and with a view to their place in the overall statutory scheme.” UARG v. EPA, 134 S. Ct. at 2441 (internal quotation marks omitted). Yet EPA myopically limits its discussion of the “context in which . . . ‘system of emission reduction’ is found” to section 111(d)(1) of the CAA in isolation, and states that nothing in that subsection constrains the Agency from applying a beyond-the-source regulatory approach. EPA Legal Memorandum at 52. As an initial matter, this conclusion is simply inaccurate. Section 111(d)(1) specifies that its standards of performance must apply to “any existing source” and directs states to account for the remaining useful life of each regulated source when applying these standards to the source, demonstrating that section 111(d) is concerned with achievable measures that individual sources can take to reduce their rate of emissions. But more importantly, it is unreasonable for EPA to limit the context of this statutory provision to a single paragraph in section 111, or even to a single section of the CAA, particularly when that term is used repeatedly throughout the Act. As discussed above, section 111 unambiguously limits standards of performance, and thus the systems of 35 emission reduction on which they are based, to individual sources. Likewise, the CAA’s other similar programs for regulating emissions from stationary sources—including programs for hazardous air pollutants, visibility-impairing pollutants, and prevention of significant deterioration of air quality—are limited to measures incorporated into the design or production processes of individual sources. Nowhere in the Proposed Guidelines does EPA acknowledge that its claimed authority to base standards on beyond-the-source measures fundamentally conflicts with its entire 44-year history of rulemaking under section 111. EPA’s Subpart B regulations speak exclusively in terms of requirements that apply to, and are implemented by, individual designated facilities. Emission guidelines must be based on “the application of the best system of emission reduction . . . that has been adequately demonstrated for designated facilities,” and EPA committed itself to address “different sizes, types, and classes” of existing sources by “specify[ing] different emission guidelines or compliance times” when “costs of control, physical limitations, geographical location, or similar factors” warrant the application of different guidelines. 40 C.F.R. § 60.22(b)(5) (emphasis added). In addition, every single NSPS or emission guideline that EPA has ever issued—spanning over four decades, multiple comprehensive amendments to the CAA, and nearly 100 subparts within EPA’s regulations—has been based on measures that individual sources can implement through design or process measures at the source itself. The Agency’s only reference to precedent is to note that some of its emission guidelines for municipal waste combustors, which were promulgated jointly under sections 111 and 129 of the CAA, have included provisions concerning emission rate averaging, tradable emission credits, and waste management plans. EPA Legal Memorandum at 63. These rules do not support the position EPA has adopted in the Proposed Guidelines. Contrary to the Agency’s 36 implication, none of the emission guidelines for municipal waste combustors were based on measures similar to Building Blocks 2, 3, and 4. Rather, the emission guidelines for existing hospital/medical/infectious waste incinerators were based on “good combustion” and wet scrubbers of varying efficiency levels, 62 Fed. Reg. 48,348, 48,371 (Sept. 15, 1997), while the emission guideline for existing commercial and industrial solid waste incinerators was based on “the use of a wet scrubbing system, or other equivalent emission control technology,” 65 Fed. Reg. 75,338, 75,344 (Dec. 1, 2000). Critically, the emission rate averaging and tradable emission credits that EPA included in its emission guidelines for municipal waste combustors were merely flexible compliance tools and were not used to determine the stringency of the standard itself, as EPA seeks to do here. Thus, these rulemakings only serve to reinforce—not to contradict—section 111’s narrow focus on measures that individual sources can implement (and have implemented in the past). This focus is unsurprising in joint rulemaking under sections 111 and 129, given that standards of performance under both provisions are required to reflect the “degree of reduction in emissions of air pollutants” that the Administrator determines is “achievable” for individual sources. Compare CAA § 111(a)(1) with CAA § 129(a)(2) (emphases added). EPA also points to the now-defunct emission guidelines under CAMR as an example of regulations under section 111 that EPA claims are based on a beyond-the-source emission trading program. EPA Legal Memorandum at 63 n.51. But like the emission guidelines for municipal waste combustors described above, CAMR’s trading program was merely a tool for compliance: the actual standards were set based on the application of pollution control technology at individual EGUs. The systems of emission reduction that were used to set CAMR’s emission guidelines were (1) the co-benefit mercury reductions of installing scrubbers 37 and selective catalytic reduction (“SCR”) at individual units under the Clean Air Interstate Rule (“CAIR”) (for the first phase of CAMR) and (2) the installation of mercury-specific pollution control technologies such as activated carbon injection (for the second phase). 70 Fed. Reg. at 28,617-20, 28,621. Thus, EPA’s reasoning in CAMR does not support a beyond-the-source approach to standard-setting under section 111. It confirms the source-focused nature of section 111, and further demonstrates that, after setting achievable emission standards for individual sources, EPA may allow states to adopt flexible programs (such as emissions trading) to comply with standards that those individual sources are otherwise capable of achieving without resorting to those flexible programs. Section 111 could not be more clear that its standards of performance for new and existing sources must be achievable by individual sources using measures that are implemented into the sources themselves. When compared to the narrow regulatory program that Congress created in section 111 and to which EPA has given shape through its consistent administrative implementation, the authority that the Agency now claims under its novel redefinition of “system of emission reduction” is plainly unlawful. EPA’s beyond-the-source approach “would bring about an enormous and transformative expansion in EPA’s regulatory authority without clear congressional authorization”—indeed, despite clear congressional language and regulatory history contradicting that claim of authority. See UARG v. EPA, 134 S. Ct. at 2444. Therefore, EPA must withdraw the Proposed Guidelines. 2. EPA Cannot Consider Reduced Utilization of Regulated Sources as BSER. These same fundamental flaws extend to EPA’s proposed “alternative approach to BSER,” under which BSER is, “in addition to [B]uilding [B]lock 1, the reduction of affected fossil fuel-fired EGUs’ mass emissions achievable through reductions in generation of specified 38 amounts from those EGUs.” 79 Fed. Reg. at 34,889 (emphasis added); EPA Legal Memorandum at 15. Under this approach, “the measures in [B]uilding [B]locks 2, 3, and 4 would not be components of the system of emission reduction but instead would serve as bases for quantifying the reduced generation (and therefore emissions) at affected EGUs, and assuring that . . . the reduced generation can be achieved while the demand for electricity services can continue to be met in a reliable and affordable manner.” 79 Fed. Reg. at 34,889. EPA has no authority to establish a standard of performance based on reduced utilization of a source. A standard of performance regulates a source’s emission performance—i.e., its emissions at a given level of operation—and not its total emissions of a pollutant. Any standard of performance under section 111 must be based on measures that can reduce an individual source’s emission capacity when operating. This is evident from the fact that Congress clearly distinguished between an “emission standard” or “emission limitation” and a “standard of performance.” CAA § 302(k), (l). The former may limit the “quantity . . . of emissions of air pollutants,” while a standard of performance must require “continuous emission reduction,” which requires imposing a limit on the rate at which the source emits an air pollutant (i.e., improving its emissions “performance”). See id. In addition, section 111(a)(1) further specifies that a “standard of performance” for the purpose of section 111 must be a “standard for emissions of air pollutants which reflects the degree of emission limitation achievable through the application of” BSER. Id. § 111(a)(1) (emphases added). Thus, the standard of performance must set forth the degree of emission limitation—i.e., the relative intensity or rate of emissions— for a source that is achievable by applying a system of emission reduction to that source. Further, the performance standard is a standard “for emissions of air pollutants,” and not one for operation of the source. There is no room in this language for a reading that Congress 39 authorized EPA to tell companies how much to operate their sources (or not to operate their sources at all). Indeed, even with respect to measures to control emissions applied to a source, Congress explicitly prohibited EPA from requiring “any particular technological system of continuous emission reduction to comply with any [NSPS]”—a requirement that would be far less intrusive than telling companies how much to run (or not to run) their sources. Id. § 111(b)(5). Tellingly, EPA has never proposed reduced utilization or operation of the source as a system of reduction under section 111—despite the fact that reducing utilization or operation of the source would always result in fewer total emissions from the source simply because the source does not have any emissions during periods of non-operation. The reductions in generation that EPA contemplates under the alternative BSER approach would not improve the emission performance of affected EGUs at any given level of operation, and thus cannot be the basis of a standard of performance. Limiting operations does not yield “continuous emission reduction.” It limits emissions only during times when a source is not running. Likewise, a standard of performance cannot be “achievable” if it can be met only by reducing utilization of the source. A source does not “achieve” a specified emissions performance level by curtailing operations; it merely ceases to “perform” at all. Congress knew how to address intermittent control strategies such as shutting down or reducing utilization of regulated sources—and when it did, it explicitly rejected their use. For example, in section 123 of the CAA, Congress provided that “dispersion techniques”—defined to include “intermittent . . . control of air pollutants varying with atmospheric conditions”—“shall not . . . affect[] in any manner” the “degree of emission limitation required for control of any air pollutant under an applicable implementation plan.” Id. § 123(a), (b). In other words, Congress rejected control strategies that vary utilization based on atmospheric conditions in favor of 40 strategies that reduce the rate of emissions. The “standard of performance” language of section 111 is inconsistent with the concept of “intermittent control” strategies, and Congress’s explicit rejection of these strategies in the context of implementation plans counsels against any reading of section 111(d) to authorize their use. An emission standard based on reducing utilization of EGUs would constitute the unlawful confiscation of generating capacity from affected EGUs. Moreover, requiring a 100 MW EGU to operate as a 50 MW EGU would effectively “order a fundamental redesign of the facility,” which EPA has no authority to do when regulating stationary sources. See UARG v. EPA, 134 S. Ct. at 2448 (discussing BACT standards). EPA’s references to other provisions of the CAA that may require some sources to reduce operations, such as the NAAQS program and the residual risk provisions of section 112, are inapposite. As these comments explain, those provisions are based on requirements driven by public health and welfare standards, whereas Congress explicitly required that section 111’s technology-based standards of performance be achievable by individual sources in a listed source category. 8 EPA Legal Memorandum at 8182. Notably, EPA has never before claimed this authority in a CAA § 111 rulemaking. As with the other beyond-the-source measures that EPA attempts to impose in the Proposed Guidelines, EPA has never even considered capping a source’s operations as part of a system of emission reduction in any of its previous NSPS or emission guideline rulemakings. If it did, the result would be bizarre, unreasonable, and unlawful. EPA is claiming authority to effectively cap the size of any new source in a category by basing its NSPS on operational limits at new sources in 8 Likewise, EPA’s reference to commitments to retire individual EGUs in NSR settlement agreements is inexplicable and irrelevant. EPA Legal Memorandum at 83. The fact that some EGUs may voluntarily decide to retire or limit operations in lieu of implementing NSR requirements bears no relation to whether EPA may require such measures as the basis for a standard of performance that applies to all sources in a source category. 41 that category. This is dissimilar from PSD review, in which a source may accept permit limits on its operations to avoid becoming subject to a stringent BACT standard as its emission rate. Under EPA’s alternative BSER proposal, the standard itself would be an operational limit, and sources would have no way to avoid becoming subject to that limit. Further, EPA’s “alternative BSER” approach is defective in that it still relies on the beyond-the-source measures that constitute Building Blocks 2, 3, and 4 of its proposed BSER. Under the alternative BSER, EPA relies on Building Blocks 2, 3, and 4 to “determine . . . the amount of the generation reduction component of the BSER” that is achievable for affected EGUs, by assuming that the beyond-the-source measures supporting those building blocks provide available means to ensure that energy demand will still be met. 79 Fed. Reg. at 34,889. But, as discussed above, the measures in Building Blocks 2, 3, and 4 are beyond the control of individual affected EGUs, and therefore cannot render a standard of performance achievable for individual sources. 3. The Proposed Guidelines Are Based on Measures that States and EPA Cannot Enforce Against Regulated Sources. EPA’s impermissible redefinition of a “system of emission reduction” is not the Agency’s only departure from the regulatory program Congress established in section 111. Indeed, in its attempt to accommodate this dramatic expansion of regulatory authority, the Agency is forced to deviate even further from the statute and its own past rulemakings in numerous other ways that also highlight the unlawfulness of the Proposed Guidelines. In particular, EPA’s proposal attempts to impose federally enforceable obligations on a broad, undefined class of “affected entities” beyond the regulated category of existing EGUs in order to effectuate the Agency’s policy goal of reducing aggregate CO2 emissions by reducing the operation of existing EGUs. This approach is untethered from any regulatory program under the 42 CAA. Nothing in the Act authorizes the states or EPA to impose obligations on any entity under section 111 other than a source in a listed source category. In fact, nothing in the Act gives EPA the authority to establish and enforce standards of performance based on its pursuit of aggregate emission reductions from a designated source category at all—a fact that should have led the Agency to abandon this misguided, ultra vires rulemaking. In this rulemaking, EPA unlawfully employs beyond-the-source measures to effectuate an “aggregate emission reduction” approach to standard-setting, in which the Agency defines an aggregate amount of emissions that it desires to remove from the entire source category’s emissions inventory, and then imposes a broad range of requirements on whatever entities are necessary to obtain those reductions. That this aggregate emission reduction approach requires emission reductions that cannot possibly be implemented through standards of performance that apply to existing EGUs alone is apparently legally irrelevant in EPA’s view. The aggregate emission reductions associated with Building Blocks 2, 3, and 4 cannot be expressed as part of a standard of performance that applies to regulated sources because they do not reduce the emission rate of individual EGUs and cannot be implemented by individual sources. The owner of an individual existing EGU cannot implement measures at that EGU that increase the share of a state’s overall fossil fuel-fired generation that is provided by NGCC units. Nor can it implement measures at that EGU that increase generation from renewable energy sources. And it cannot implement measures at the EGU that cause electricity consumers to use energy more efficiently, or ensure that any efficiency improvements actually lead to reduced energy consumption. Likewise, EPA and states have no authority under section 111 to base a standard of performance on the reduced utilization of a source, and even if they had such authority, the 43 aggregate emission reductions associated with that reduced utilization would not be reflected in a “standard of performance” as an improvement in the EGU’s emission rate. At the same time, nothing in section 111 authorizes EPA or the states to impose legally enforceable obligations on other entities in order to effectuate the aggregate emission reductions that EPA has concluded could be obtained under Building Blocks 2, 3, and 4. To the contrary, the statute could not be more clear that the only emission control obligations that it authorizes EPA and states to impose are standards of performance that apply to new and existing sources in listed source categories. Section 111 provides that legal requirements will apply only to regulated sources themselves. Under section 111(b), EPA may “establish[] Federal standards of performance for new sources” within “a category of stationary sources in a list under subparagraph (A).” CAA § 111(b)(1)(B) (emphasis added). And under section 111(d), states may submit plans that “establish[] standards of performance for any existing source for any air pollutant . . . to which a standard of performance under this section would apply if such existing source were a new source.” Id. § 111(d)(1)(A) (emphasis added). Nowhere else in that section does Congress authorize the imposition of binding legal obligations for entities other than the regulated source. Accordingly, only a narrow set of entities may be subject to requirements under section 111(d). First, the entity must be a “stationary source,” meaning a “building, structure, facility, or installation which emits, or may emit any air pollutant.” Id. § 111(a)(3). Second, that stationary source must be an existing source that falls within the “category of sources” that EPA has listed as eligible for regulation because that category “causes, or contributes significantly to, air pollution which may reasonably be anticipated to endanger public health or welfare.” Id. § 111(b)(1)(A). And finally, EPA must have promulgated a standard of performance for new 44 sources in that category that applies to the relevant pollutant. 9 Id. § 111(d)(1)(A)(ii). If an entity is not a stationary source of pollutants in a category for which the Agency has promulgated an endangerment finding and an applicable NSPS, then neither EPA nor the states have authority to impose any obligation on that entity under section 111(d). Thus, the statute unambiguously does not authorize EPA to impose obligations on the types of entities that the Agency would need to regulate in order to implement the beyond-thesource measures underlying the Proposed Guidelines. For that reason, “the need to rewrite clear provisions of the statute should have alerted EPA that it had taken a wrong interpretive turn” when it adopted its overly broad redefinition of “system of emission reduction.” UARG v. EPA, 134 S. Ct. at 2446. But instead, the Agency proposes to take another wrong turn away from the regulatory program established in section 111 and to direct states to impose legal obligations on “affected entities” other than EGUs to obtain the aggregate emission reductions that cannot be achieved by adopting proper standards of performance for existing EGUs. 79 Fed. Reg. at 34,901-03. EPA’s proposed definition of affected entity is circular and astonishingly broad. It includes any “entity with obligations under this subpart for the purpose of meeting the emissions performance goal requirements in these emission guidelines.” 79 Fed. Reg. at 34,956, Proposed 40 C.F.R. § 60.5820. In other words, under EPA’s so-called “portfolio approach,” the Proposed Guidelines would allow a state to impose enforceable requirements upon any entity that the state believes could directly or indirectly reduce energy demand, and thus emissions, from existing 9 As discussed in more detail in Section IV, regulation under section 111(d) also is permissible only when the pollutant in question is not listed as a criteria air pollutant and regulated under the NAAQS program and when the source category is not regulated under section 112’s hazardous air pollutant program. CAA § 111(d)(1)(A). 45 fossil fuel-fired EGUs. By EPA’s own admission, the standards of performance applicable to individual affected EGUs—i.e., the standards actually contemplated by section 111—“would not, on their own, assure, or be required to assure, achievement of the emission performance level” that is determined to represent the application of BSER to affected sources. 79 Fed. Reg. at 34,901. Instead, “the state plan would include measures enforceable against other entities that support reduced generation by, and therefore CO2 emission reductions from, the affected EGUs.” Id. These measures “would be federally enforceable because they would be included in the state plan.” Id. The Agency suggests that affected entities could include “electric distribution utilit[ies],” “private or public third-party entit[ies],” or “a state agency, authority or entity.” Id. at 34,917. There is virtually no limit on the types of “affected entities” that EPA or states might choose to regulate under section 111, potentially including “retail stores, offices, apartment buildings, shopping centers, schools, and churches.” UARG v. EPA, 134 S. Ct. at 2446. The Supreme Court recently rejected a similar attempt by EPA to “assert[] newfound authority” over such entities and then “decide, on an ongoing basis . . . how many of those sources to regulate.” Id. EPA’s attempt to authorize regulation of these “affected entities” under a program to establish performance standards for a source category of EGUs has no basis in the CAA. Much like the Agency’s novel redefinition of “system of emission reduction,” EPA has never applied this “portfolio approach” in any of its previous NSPS or emission guideline rulemakings over the past four decades. The Agency’s primary argument for this departure is that “[t]here is no specific language in CAA section 111(d) or elsewhere in the Act that prohibits states from including measures other than performance standards and implementation and enforcement measures” in state plans for existing sources. 79 Fed. Reg. at 34,903 (emphasis added). But this 46 argument erroneously assumes that, under the CAA, any action that Congress did not explicitly prohibit is permitted. In fact, just the opposite is true. Congress must “speak clearly if it wishes to assign to an agency decisions of vast ‘economic and political significance.’” UARG v. EPA, 134 S. Ct. at 2444. A reading of the CAA that gives EPA and states the authority to potentially impose obligations on any entity that uses electricity would unquestionably bear vast economic and political significance. As EPA freely admits, “the terms of CAA section 111(d)(1) do not explicitly address whether, in addition to emission limits on affected EGUs, state plans may include other measures for achieving the emission performance level.” 79 Fed. Reg. at 34,902. Without such explicit authorization, EPA may not seize regulatory authority over a swath of affected entities so broad that it eclipses the affected sources on which section 111’s narrow regulatory program is focused. As EPA also acknowledges, its proposal to include enforceable requirements for affected entities as part of the emission guidelines raises significant practical enforcement concerns that are not addressed in the Proposed Guidelines. See id. at 34,909 (“[A] portfolio approach may result in enforceable state plan obligations accruing to a diverse range of affected entities beyond affected EGUs, and . . . there may be challenges to practically enforcing against some such entities in the event of noncompliance.”). In order to receive EPA approval, a state plan must contain “enforceable measures that reduce EGU CO2 emissions.” Id. Once a state plan is approved by the Agency, its provisions—including its requirements for affected entities—would then generally become federally enforceable by EPA and by private individuals in citizen suits. See CAA §§ 113, 304(a). Yet it is unclear what kind of “enforceable” obligations a state could impose on affected entities in order to implement Building Blocks 2, 3, and 4, or how those 47 requirements could be enforced in the event they are not implemented. Accordingly, it may well prove impossible for states to submit plans that are “approvable” under EPA’s criteria. For example, EPA suggests that in order to implement Building Blocks 2, 3, and 4, “affected entities” may include “a state agency, authority or entity,” such as a state environmental management agency or public utilities commission. 79 Fed. Reg. at 34,917. In some states, these may be the only entities with authority to implement the redispatch of generation from existing coal-fired units to NGCC units or to coordinate investments in renewable energy generation or demand-side energy efficiency campaigns. These state agencies, however, are immune from suit under the CAA’s citizen suit provision. Sierra Club v. Korleski, 681 F.3d 342, 351 (6th Cir. 2012) (“The text and structure of the CAA make plain that [section 304(a)(1)] does not permit citizen suits against state regulators qua regulators.”). Thus, it would not be possible to enforce emission-reducing measures that are the responsibility of state agencies. In such a case, under the plain terms of EPA’s proposal, the state’s plan submission would not be approvable unless the state adopts legislation expressly opening its agencies to citizen suits seeking to enforce these obligations, which would plainly exceed the scope of measures EPA may require in a state plan. Even where “affected entities” are not state agencies that are immune from CAA citizen suits, it is unclear what would constitute a “violation” warranting an enforcement action or how a court would enjoin that violation. For example, a state might assign responsibility for implementing Building Block 3 (increased renewable energy generation) to a third party, such as an electric distribution utility, and require that utility to invest a certain amount of money each year into new renewable energy development. Even assuming arguendo that such an investment requirement is legally enforceable, is there any guarantee that those investments will result in the 48 required reduction in EGU CO2 emissions to render the state plan approvable? Can the utility demonstrate compliance with that obligation by simply making the required investments, or must it also show that the contemplated amount of renewable energy was produced as a result? Would the utility “violate” its obligation if new renewable energy capacity failed to materialize, or if that renewable capacity was not utilized enough (e.g., due to variations in weather) to generate the contemplated amount of renewable energy? If so, how would a court enjoin that violation— by requiring greater investments, or by requiring increased renewable generation? Would the owners of renewable energy projects receiving investments under the state plan also be “affected entities” subject to enforceable obligations requiring a minimum amount of generation each year? If so, would it be a violation of that requirement if a solar or wind-powered energy source did not meet its generation target because of insufficient sun or wind? The same or similar questions arise in the context of Building Blocks 2 and 4. EPA has not answered any of these essential implementation questions. Nor can the CAA or EPA’s past rulemakings under section 111 provide answers to any of these questions. That is because the types of aggregate emission reduction measures and the broad universe of affected entities that EPA asserts authority over in the Proposed Guidelines are so far removed from section 111’s narrow source-focused regulatory program as to be completely unrecognizable as a program under section 111(d) of the CAA. The CAA “do[es] not explicitly address” any of these issues, 79 Fed. Reg. at 34,902, for the same reason that section 111 does not explicitly address mobile source regulation or endangered species protection: because such language would be nonsensical in a statutory provision that has nothing to say about those issues. Section 111, on its face, is singularly concerned with design or process changes that individual new and existing stationary sources within specific categories can implement to reduce their own 49 emissions. That EPA is now grappling with the complex question of how to impose enforceable obligations on other affected entities in order to reduce aggregate emissions by reducing utilization of affected sources simply demonstrates that EPA has strayed far off the CAA’s regulatory path. If these issues were not enough, the extent to which EPA has deviated from the statute is also evident from the fact that if a state does not submit an “approvable” plan, EPA would have no authority to promulgate a federal plan that includes the measures it would require of states (i.e., beyond-the-source measures contemplated under Building Blocks 2, 3, and 4). In this regard, EPA cannot adopt a plan that imposes a federally enforceable energy resource development and dispatch program upon states. The Agency has no authority over independent system operators, state public utility commissions, or other entities responsible for managing the dispatch of EGUs to meet load requirements. EPA has no inherent police powers and lacks authority to require states to exercise their police powers. Notably, even under section 110 of the CAA—a provision that, in contrast to section 111, gives the states and EPA broad authority to require aggregate emission reductions from diverse sources—EPA has no power under a federal implementation plan to require that states enact specific statutes and regulations or to administer and enforce programs contained in EPA regulations. Dist. of Columbia v. Train, 521 F.2d 971, 983, 984 (D.C. Cir. 1975) (“Had Congress intended to adopt the novel approach of empowering a federal agency to order unconsenting states to enact state statutes and regulations, thereby converting state legislatures into arms of the EPA, it most likely would have made that intent clear in the statute.”), vacated as moot, EPA v. Brown, 431 U.S. 99 (1977); Maryland v. EPA, 530 F.2d 215, 228 (4th Cir. 1975) (CAA does not authorize EPA “to require Maryland to establish the programs [described in EPA’s regulations] and furnish legal authority for the 50 administration thereof”), vacated as moot, Brown, 431 U.S. at 103-04. Indeed, if any provision of the CAA did purport to authorize EPA to require that states implement and enforce new statutes or regulations, it would present serious constitutional problems. See Maryland, 530 F.2d at 225 (observing that “how an Act of Congress may be construed to permit an agency of the United States to direct a state legislature to legislate is difficult to understand”). Those same constitutional concerns are implicated where, as here, EPA would require state plans to include measures that can only be implemented and enforced (whether by the state or by EPA) if the state enacts new legislation. And importantly, as noted above, section 111 gives EPA no authority to base a standard of performance on reduced utilization or operation of a source. In sum, section 111 of the CAA creates a narrow regulatory program in which EPA— or in the case of existing sources, states—adopts standards of performance for stationary emission sources in specified categories that reduce each source’s emissions at a given operational level based on the implementation of achievable measures that can be incorporated into the design or production process of the source itself. EPA has consistently applied the program in this manner for over forty years. By contrast, the Proposed Guidelines require states to enact measures that reduce aggregate emissions from each state’s fleet of existing EGUs overall, based on measures that are beyond the control of individual EGUs and that can be achieved only by imposing obligations on other “entities” the state identifies that produce or consume electricity or by requiring existing EGUs to curtail operations altogether. The contrast could not be more clear. EPA’s Proposed Guidelines are inconsistent with the plain language of section 111(d), as well as “the design and structure of the statute as a whole,” and must be withdrawn. UARG v. EPA, 134 S. Ct. at 2442 (internal quotation marks omitted). 51 IV. EPA Is Prohibited from Regulating Pollutants from a Source Category Already Regulated Under Section 112 of the CAA. The Proposed Guidelines contravene the express terms of section 111(d), which prohibits EPA from adopting emission guidelines for existing sources from a source category when the Agency has already regulated that source category under section 112. CAA § 111(d)(1)(A)(i). EPA listed coal- and oil-fired EGUs as a “source category” under section 112 in 2000, 65 Fed. Reg. 79,825, 79,826, 79,831 (Dec. 20, 2000), and regulated emissions from these sources in 2012 under the MATS rule, 77 Fed. Reg. 9304 (Feb. 16, 2012). Because EPA has already promulgated a national emission standard for existing coal-fired power plants under section 112, it may not promulgate the emission guidelines it has proposed here. A. EPA’s Interpretation of the Act Is Inconsistent With the Plain Language of Section 111(d), Which Contains No Ambiguity. EPA argues in its Legal Memorandum that accompanies the Proposed Guidelines that section 111(d) is “ambiguous” about whether the Agency can regulate sources under both section 112 and section 111(d), and that this ambiguity allows EPA to interpret the statute in any reasonable manner. EPA Legal Memorandum at 23 (“The two versions conflict with each other and thus render the Section 112 Exclusion ambiguous. Under these circumstances, the EPA may reasonably construe the Section 112 Exclusion to authorize the regulation of GHGs under section 111(d).”). Courts grant agencies deference in interpreting the meaning of ambiguous statutes. See, e.g., Chevron, U.S.A., Inc. v. Natural Res. Defense Council, 467 U.S. 837 (1984). But there is no ambiguity in section 111(d). Section 111(d) says plainly that EPA has no authority to regulate a source category already regulated under section 112. The basis of EPA’s argument that section 111(d) is ambiguous is the existence in the Statutes at Large of two amendments to section 111(d) in the 1990 amendments to the Clean Air 52 Act. One amendment was a technical amendment not codified in the U.S. Code because it was obvious to the codifier that the substantive amendment that was enacted superseded and rendered unnecessary the technical amendment. The U.S. Code therefore notes that the clerical entry “could not be executed.” Revisor’s Note, 42 U.S.C. § 7411. The technical amendment included in the Statutes at Large sought to update a cross-reference to section 112(b)(1)(A) after section 112(b)(1)(A) was eliminated by another 1990 amendment to the Act. This non-substantive, ministerial revision was included in a section in the Statutes at Large entitled “Conforming Amendments.” Clean Air Act Amendments, Pub. L. No. 101-549, § 302(a), 104 Stat. 2399, 2574 (1990). But this revision was superseded and became moot when the second, substantive amendment to section 111(d) passed nearly two months later. H.R. 3030 (containing the substantive provision) passed on May 23, 1990, while S. 1630 (containing the ministerial crossreference) passed on April 3, 1990. See H.R. REP. NO. 101-490, at 454 (1990), reprinted in 2 A LEGISLATIVE HISTORY OF THE CLEAN AIR ACT AMENDMENTS OF 1990 (“Leg. History”), at 3021, 3478 (1993) (report to accompany H.R. 3030); S. 1630, 101st Cong. § 112(e)(4)-(5)(D) (as passed by Senate, Apr. 3, 1990), reprinted in 3 Leg. History, at 4119, 4431-33. Although the technical amendment had a role to play prior to the substantive amendment to section 111(d) that was eventually adopted, once that amendment was adopted, inclusion of the technical amendment in the Statutes at Large was simply a “drafting error,” as EPA has properly recognized. 70 Fed. Reg. 15,994, 16,031 (Mar. 29, 2005). As EPA stated, it was “hard to conceive that Congress would have adopted section 112(n)(1)(A), yet retained the [clerical] amendment to section 111(d).” Id. at 16,031. It does not conflict with and need not be reconciled with the substantive amendment which, in “a literal reading,” EPA admits, means that the Agency “c[an] not regulate any air pollutant from a source category regulated under section 53 112.” EPA Legal Memorandum at 26 (emphasis added). The Supreme Court has interpreted the provision in the same way: “EPA may not employ § 7411(d) if existing stationary sources of the pollutant in question are regulated under . . . the ‘hazardous air pollutants’ program, § 7412.” Am. Elec. Power Co. v. Connecticut, 131 S. Ct. 2527, 2537 n.7 (2011). And yet, in an effort to expand its regulatory reach and to invoke judicial deference in doing so, the Agency attempts to fabricate ambiguity in the statute. The D.C. Circuit has explained that it will not “read[] too much into a mere clerical error” and notes that incorrect conforming amendments (or the lack thereof) are common and can often be “the result of a scrivener’s error.” Am. Petroleum Inst. v. SEC, 714 F.3d 1329, 1336, 1337 (D.C. Cir. 2013). Such errors should thus not be interpreted as giving rise to an “ambiguity.” Id. at 1336. EPA is not required to reconcile obvious mistakes—editorial vestiges that were never executed in the U.S. Code—with substantive, intentional amendments to the Act. EPA should simply disregard obvious mistakes in the Statutes at Large that, because they were mistakes, “could not be executed” in the U.S. Code. Even if there were ambiguity in the presence of the two provisions in the Statutes at Large, however, agencies have deference only to interpret the ambiguous meaning of a statute, not to divine what the text of a statute actually is. In this case, if there is any uncertainty at all (which there is not), it centers on what the language of the statute actually is: why Congress neglected to reconcile both a substantive and a superseded clerical provision before they were entered in the Statutes at Large. The meaning of each of the substantive and clerical amendments is clear, not ambiguous. EPA inexplicably argues that both amendments should be read together even when there is no doubt that Congress did not intend for a superseded technical amendment to remain in the statute. Disregarding this obvious intent, the Agency insists it 54 should “give some effect to both amendments.” EPA Legal Memorandum at 26. But even if two provisions in a statute are irreconcilable, “Chevron is not a license for an agency to repair a statute that does not make sense.” Scialabba v. Cuellar de Osorio, 134 S. Ct. 2191, 2214 (2014) (Roberts, C.J., concurring). In other words, “[d]irect conflict is not ambiguity, and the resolution of such a conflict is not statutory construction but legislative choice.” Id. Courts will grant agencies deference only “because we presume that Congress intended to assign responsibility to resolve the ambiguity to the agency.” Id. If it is clear that Congress did not assign such responsibility to an agency, but merely made a drafting mistake (which it does with some frequency), it is equally clear that such mistakes do not confer to agencies the license to rewrite the statute. UARG v. EPA, 134 S. Ct. at 2446 (“[A]n agency may not rewrite clear statutory terms to suit its own sense of how the statute should operate.”). EPA in fact cites no authority for the deference it implies it should receive to reconcile the “conflict” in section 111(d) it purports to have discovered. In the litigation challenging CAMR, however, which also attempted to regulate coal-fired EGUs under section 111(d), EPA cited Citizens to Save Spencer County v. EPA, 600 F.2d 844 (D.C. Cir. 1979), to support its creative interpretation of section 111(d). Final Br. of Resp’t EPA at 102-03, New Jersey v. EPA, 517 F.3d 574 (D.C. Cir. 2008) (No. 05-1097). This is a very different situation from Citizens to Save Spencer County, where two different statutory provisions had different, mutually exclusive dates for regulatory action. Those provisions could not be reconciled and the Agency was found to have reasonably “pursue[d] a middle course.” Citizens to Save Spencer Cnty., 600 F.2d at 871. By contrast, the amendments at issue here can easily be reconciled, first and foremost by recognizing that the clerical amendment was inserted in the Statutes at Large by mistake and, therefore, should be disregarded. Even if one desires to infuse the clerical amendment with 55 substantive intent, as discussed below, the two provisions are most clearly read as prohibiting the duplicative regulation of sources under both section 111(d) and section 112. B. EPA’s Proposed Construction Does Not Accord With Either Provision or With Legislative Intent. Even if one were to attempt to treat the clerical amendment as substantive, as EPA does, the two provisions can be read in harmony. One can easily give effect to both provisions, which when read together simply say that no source category and no pollutant that are regulated under section 112 can be regulated under section 111. In other words, reading both amendments together, EPA is prohibited from regulating under section 111(d) any pollutant or any source category that is regulated under section 112. This construction is the only one that construes both provisions substantively and that is also faithful to the plain text of both. The substantive amendment in the Statutes at Large reads: “Regulation of Existing Sources.—Section 111(d)(1)(A)(i) of the Clean Air Act . . . is amended by striking ‘or 112(b)(1)(A)’ and inserting ‘or emitted from a source category which is regulated under section 112.’” Clean Air Act Amendments, Pub. L. No. 101-549, § 108(g), 104 Stat. 2399, 2467 (1990). The clerical amendment reads: “Conforming Amendments. (a) Section 111(d)(1) of the Clean Air Act is amended by striking ‘112(b)(1)(A)’ and inserting in lieu thereof ‘112(b).’” Id. § 302(a), 104 Stat. at 2574. In the CAMR rulemaking, the Agency adopted an interpretation of the two amendments in its attempt “to give some effect to both.” EPA Legal Memorandum at 26. It proposes to do so again here, though with a slightly different hurdle to maneuver. In CAMR, EPA sought to regulate a hazardous air pollutant under section 111(d), and as part of that effort, “delisted” the source category of coal- and oil-fired EGUs from regulation under section 112. In 2005, EPA interpreted the two amendments together to mean: 56 Where a source category is being regulated under section 112, a section 111(d) standard of performance cannot be established to address any HAP listed under section 112(b) that may be emitted from that particular source category. Thus, if EPA is regulating source category X under section 112, section 111(d) could not be used to regulate any HAP emissions from that particular source category. 70 Fed. Reg. at 16,031-32. In this interpretation, rather than giving effect to both amendments, the Agency in fact limited the reach of each and constrained its authority to regulate to a world much smaller than either the clerical or substantive amendment alone contemplated. 10 EPA thus gave effect to neither provision and proposes the same interpretation now. EPA ignores the statutory text, despite being required “to give effect, if possible, to every word Congress used.” Reiter v. Sonotone Corp., 442 U.S. 330, 339 (1979). EPA did not then and does not now explain how altering the intent of both amendments by limiting their individual reaches “gave effect” to each amendment. EPA’s interpretation expressly contravenes the first amendment (under section 108(g) of the Statutes at Large, 104 Stat. at 2467) because it would allow EPA to regulate the same source category it is regulating under section 112. EPA’s approach would also expressly contravene the second amendment (under section 302(a) of the Statutes at Large, 104 Stat. at 2574) because it would allow the Agency to use section 111(d) to regulate hazardous air pollutants from a source category that is not already subject to section 112 standards. But perceived ambiguity or conflict “gives no license to a court or agency to indulge in unrestrained and fanciful flights of constructional imagination to arrive at artful but artificially consistent interpretations.” Citizens to Save Spencer Cnty., 600 F.2d at 870. 10 The substantive amendment addresses source categories, without regard to pollutant (“Universe A”). The clerical amendment, read substantively, addresses pollutants, without regard to source category (“Universe B”). By artificially merging the two, EPA created not Universe A + B, but Universe C, a potential regulated universe much smaller than either Universe A or Universe B, thereby ignoring the intents of both amendments rather than choosing to give effect to both or at least one. 57 Besides limiting the universe of sources and gases that EPA can regulate under section 111(d) beyond what either the House or Senate intended with their amendments alone, the Agency’s interpretation also unreasonably constrains the plain meaning of the substantive amendment in the mistaken interest of giving some meaning to the clerical amendment, which again EPA admits was likely a non-substantive “drafting error.” 70 Fed. Reg. at 16,031. Although EPA argued that its 2005 interpretation “gives effect to the House’s desire to increase the scope of EPA’s authority under section 111(d) and to avoid duplicative regulation of [hazardous air pollutants] for a particular source category,” id. at 16,032 (citing 136 Cong. Rec. H12911, 12934 (daily ed. Oct. 26, 1990)), it also acknowledged the interpretation “does not give full effect to the [substantive amendment], because a literal reading . . . would mean that EPA could not regulate [hazardous air pollutants] or non-[hazardous air pollutants] emitted from a source category regulated under section 112,” id. It is not surprising that Congress would have confirmed and clarified the limited reach of section 111(d). Prior to the 1990 Amendments, section 111(d) looked solely to whether the pollutant at issue was already regulated under section 108 or section 112. With the 1990 amendments, section 112 was transformed from a little used provision into a comprehensive control technology program. See EPA, “Summary of the Clean Air Act,” http://www2.epa.gov/ laws-regulations/summary-clean-air-act; H.R. REP. NO. 101-490, at 151 (1990), reprinted in 2 Leg. History, at 3175; S. REP. NO. 101-228, at 148 (1990), reprinted in 5 Leg. History, at 8338, 8488. Congress fundamentally changed section 112 so that those standards now “apply to sources in a category . . . rather than to pollutants individually.” S. REP. NO. 101-228, at 148 (1990), reprinted in 5 Leg. History at 8488. To avoid onerous duplication, Congress then amended section 111, which is also technology-based and also applies to source categories rather 58 than to pollutants. If an entity is regulated under section 112, it cannot also be regulated under section 111(d). For all the reasons discussed above, EPA’s interpretation of section 111(d) does not accord with either the plain statutory language or its legislative history. EPA lacks authority to regulate coal- and oil-fired EGUs under section 111(d). EPA must withdraw the Proposed Guidelines because they violate both the text and Congress’s clear intent in the 1990 Amendments. V. EPA Lacks the Sufficient Legal Prerequisites To Propose and Finalize the Proposed Guidelines. Section 111(d) makes clear that EPA may not finalize guidelines for an existing source unless new sources of the same type of entity are already subject to a section 111(b) performance standard. CAA § 111(d)(1). EPA’s Subpart B regulations spell this requirement out in more detail and extend it to the proposal stage: Only “[c]oncurrently upon or after” standards have been proposed for “a designated pollutant from [new] facilities” may EPA propose draft emission guidelines for the same pollutant for “designated facilities.” 40 C.F.R. § 60.22(a). A “designated facility” is an “existing facility” of the “same type” that is subject to a standard of performance under section 111(b). Id. § 60.21(b); 40 Fed. Reg. at 53,341. Thus, in order for EPA to lawfully propose and finalize section 111(d) emission guidelines for EGUs, it first must have lawfully proposed and finalized standards of performance for new EGUs of the same type. The legal prerequisites for lawful emission guidelines governing CO2 emissions from EGUs have not been met, and the Proposed Guidelines are therefore unlawful and may not be finalized. A. EPA’s Proposed NSPS for New Coal-Fired EGUs Is Unlawful and Cannot Support Section 111(d) Emission Guidelines. Because EPA’s proposed NSPS for new coal-fired EGUs fails to meet the statutory requirements for a “standard of performance” under section 111(b) of the Act, it cannot satisfy 59 the legal prerequisite for the Proposed Guidelines. The CAA requires a “standard of performance” to be “achievable” and based on a “best system of emission reduction” that must be “adequately demonstrated” for sources within a category or subcategory. CAA § 111(a)(1). The BSER must be a design or process measure that is available for each source type to which the standard will apply. EPA’s proposed NSPS for new Subpart Da fossil-fired units is based on the use of carbon capture and storage (“CCS”), a technology that, as UARG exhaustively demonstrated in its May 2014 comments to EPA on that proposed rule, is neither available nor adequately demonstrated. See UARG New Source Comments at 18-85. Furthermore, an NSPS must reduce emissions from a source by incorporating a system of continuous emission reduction into the design of the source itself. EPA’s proposed NSPS for new Subpart Da sources failed this most basic test; the proposed Subpart Da standard would not reduce emissions. The proposed NSPS for new Subpart Da sources, which is based on one element of CCS, would mandate no reduction in the CO2 generated by a source. Instead, it merely requires that a new source separate some percentage of the CO2 from its flue gas stream, without imposing any requirements regarding the CO2’s ultimate fate. The proposed NSPS for new Subpart Da units is a “carbon separation” system, not a “carbon capture and sequestration” system, and therefore it does not reduce emissions. The proposed 1,100 lb CO2/MWh standard merely reflects the amount of unseparated CO2 a source would be limited to releasing from its stack. Yet the total CO2 generated by the source would not change. The remainder of the CO2 generated by the source and not emitted from the stack would be sent off-site with the expectation that it would be sequestered. But it would not violate the proposed NSPS if 100 percent of the CO2 separated by the capture technology subsequently were released to the atmosphere. Thus, the proposed NSPS for Subpart Da units would not reduce emissions. 60 Moreover, in the proposed NSPS for new EGUs, EPA failed to show that the proposed separation-based standard is achievable for sources within the category as a whole because its limited discussion of achievability considers only integrated gasification combined cycle (“IGCC”) units, which make up only one type of source covered by the proposed standard. A standard based solely on “IGCC with a single-stage [water gas shift] reactor and a two-stage acid gas removal system,” 79 Fed. Reg. at 1470, a form of pre-combustion CO2 separation, cannot support a standard for pulverized coal (“PC”) or circulating fluidized bed units, which require post-combustion separation. EPA did not analyze or even assert that the proposed separation standard is achievable for coal-fired units using post-combustion capture methods. Any standard that applies to PC units based on a finding that it is achievable with a technology that only IGCC units can use is contrary to law. Because there is no lawful proposed standard of performance for new coal-fired EGUs, EPA lacks the legal prerequisite to propose or promulgate the Proposed Guidelines for those sources. A standard of performance under section 111(b) of the Act must be “achievable” and “adequately demonstrated” for sources in a source category. Because EPA’s January 2014 NSPS Proposal for new sources is neither, the section 111(b) new source rule does not meet the requirements to be a standard of performance for new coal-fired units. Without an effective standard for new units, EPA cannot propose a standard for existing units. See CAA § 111(d)(1); 40 C.F.R. § 60.22(a). B. The Proposed NSPS for New Sources Does Not Address the Same Type of Facilities as the Proposed Guidelines. Even if EPA could demonstrate that it proposed a lawful NSPS for new EGUs, which it cannot do, its Proposed Guidelines do not address the “same type” of facility that the proposed NSPS for new EGUs would address. On this basis alone the Proposed Guidelines lack the legal 61 prerequisite that a standard of performance for the same type of facility will have been proposed concurrently with or before the section 111(d) proposal. Id. Under the Agency’s own Subpart B regulations, emissions guidelines may apply only to designated facilities—the same type of facility regulated under the new source NSPS rule. In this context, that would mean that EPA could propose emission guidelines only for fossil fuel-fired EGUs that meet the threshold requirements set forth in the new source NSPS proposal. In the Proposed Guidelines, however, EPA never defines what the regulated “designated facility” would be. In an attempt to accomplish its novel, systems-based approach to BSER, EPA dodges the question of what the “designated facility” is and instead has introduced a new term: “affected entity.” See, e.g., 79 Fed. Reg. at 34,951, Proposed 40 C.F.R. § 60.5740(a)(3) (State plans must identify “the state emission performance level for affected entities that will be achieved through implementation of the plan.”); id., Proposed 40 C.F.R. § 60.5740(a)(3)(ii)(B) (“For a mass-based CO2 emission performance level, the identified level of performance must represent the total tons of CO2 that will be emitted by affected entities during each plan performance period.”) (emphasis added). EPA defines “[a]ffected [e]ntities” to mean “any of the following: An affected EGU, or another entity with obligations under this subpart for the purpose of meeting the emissions performance goal requirements in these emission guidelines.” Id. at 34,956, Proposed 40 C.F.R. § 60.5820 (emphasis added). The CAA and EPA’s Subpart B regulations prohibit EPA from expanding the universe of covered sources under section 111(d) beyond those regulated as new sources under section 111(b). CAA § 111(d)(1); 40 C.F.R. § 60.22(a). The Subpart B regulations limit EPA to proposing and finalizing emission guidelines for “designated facilities,” not “affected entities,” which EPA has defined to be a world much larger 62 than “designated facilities.” See, e.g., 40 C.F.R. § 60.21(e) (“Emission guideline means a guideline set forth in subpart C of this part, or in a final guideline document published under § 60.22(a), which reflects the degree of emission reduction achievable through the application of the best system of emission reduction which (taking into account the cost of such reduction) the Administrator has determined has been adequately demonstrated for designated facilities.”) (second emphasis added); id. § 60.22 (a) (“Concurrently upon or after proposal of standards of performance for the control of a designated pollutant from affected facilities, the Administrator will publish a draft guideline document containing information pertinent to control of the designated pollutant [from] designated facilities.”) (emphasis added); id. § 60.22(b)(5) (“The Administrator will specify different emission guidelines or compliance times or both for different sizes, types, and classes of designated facilities when costs of control, physical limitations, geographical location, or similar factors make subcategorization appropriate.”) (emphasis added). EPA insists it is “following the requirements of” the Subpart B regulations and is not amending or reopening them, 79 Fed. Reg. at 34,844, and yet it fails to adhere to the most basic of the Subpart B requirements: identifying the regulated source category, which by law must be identical to the source category regulated under the section 111(b) NSPS. In fact, the very first non-definitional sentence of the Subpart B regulations states that EPA’s guideline document must contain “information pertinent to control of the designated pollutant [from] designated facilities.” 40 C.F.R. § 60.22(a). Indeed, the Proposed Guidelines would regulate sources that extend far beyond the coaland gas-fired EGUs EPA would regulate under its January 2014 NSPS Proposal. EPA does at times use the term “affected EGU” but such term is not used interchangeably or as a synonym with the term “affected entity.” It is clear that EPA intends for the section 111(d) emission 63 guidelines to apply to “affected entities,” not only to “affected EGUs.” See, e.g., 79 Fed. Reg. at 34,953, Proposed 40 C.F.R. § 60.5780(a) (“Your state plan shall include emission standard(s) that are quantifiable, verifiable, non-duplicative, permanent, and enforceable with respect to each affected entity.”) (emphasis added); id. at 34,953-54, Proposed C.F.R. § 60.5780(f) (“An emission standard is enforceable against an affected entity if: (1) A technically accurate limitation or requirement and the time period for the limitation or requirement is specified; (2) Compliance requirements are clearly defined; (3) The affected entities responsible for compliance and liable for violations can be identified; (4) Each compliance activity or measure is enforceable as a practical matter; and (5) The Administrator and the state maintain the ability to enforce violations and secure appropriate corrective actions . . . .”) (emphases added). Only after EPA discusses the requirement that state plans apply standards to “affected entities,” EPA inserts a separate section in its proposed regulatory language discussing the “Applicability of State Plans to Affected EGUs.” Id. at 34,954. States must address emissions from affected EGUs in their state plans, but it is abundantly clear that EPA’s design envisions emissionsrelated requirements under section 111(d) applying to a much broader range of sources than the sources that would be regulated under its section 111(b) NSPS proposal. EPA explains in the preamble that state plans may apply “enforceable measures” that “are implemented by the state or by another entity” in addition to EGUs. Id. at 34,837 (emphasis added); see also id. at 34,853 (“Based on requests from stakeholders, the EPA is proposing that states be authorized to submit state plans that do not impose legal responsibility on the affected EGUs for the entirety of the emission performance level, but instead, by adopting what this preamble refers to as a ‘portfolio approach,’ impose requirements on other affected entities (e.g., renewable energy and demandside energy efficiency measures) that would reduce CO2 emissions from the affected EGUs.”) 64 (emphasis added). Thus, under EPA’s Proposed Guidelines, the “designated facility” would be EGUs, the state (the entity to which EPA is in fact applying a numeric rate-based emissions goal), and other entities, including any “private or public third-party entity.” Id. at 34,917. Moreover, as discussed in Section IX.A, EPA has no authority, under either section 111(d) or the Subpart B regulations, to adopt nonnegotiable, binding “emission goals” for states as it proposes to do. Subpart B does not even mention the term “emission goal.” Both Subpart B and section 111(d) speak in terms of “standards of performance.” Under section 111(a)(1), the standard of performance must reflect the degree of “emission limitation” that is achievable with the BSER. Under section 111(d), states are to establish standards of performance applicable to sources. EPA’s convoluted approach disregards both the statute and its own implementing regulations and would apply to states a binding emission limitation, and states would then be free to apply standards of performance to any “affected entity” it wishes, without fealty to the restriction that the regulated entity must be identical in the section 111(b) and section 111(d) regulations. EPA’s interpretation of its own regulations does not accord with the statute. Moreover, because EPA acts contrary even to its own regulations, it is entitled to no deference in interpreting them. Finally, for the same reasons, EPA’s proposal for modified and reconstructed sources under section 111(b) is also not sufficient to justify the section 111(d) proposal. See 79 Fed. Reg. at 34,852 (“The EPA intends to complete two CAA section 111(b) rulemakings regulating CO2 from new fossil fuel-fired EGUs and from modified and reconstructed fossil fuel-fired EGUs before it finalizes this rulemaking, and either of those section 111(b) rulemakings will provide the requisite predicate for this rulemaking.”). Although the CAA defines the term “new source” to include a modified source, CAA § 111(a)(2), as discussed above, the Proposed 65 Guidelines seek to regulate systems, states, and sources other than individual EGUs. The Proposed Guidelines define an “Affected Entity” as any entity with an enforceable obligation under a state plan. 79 Fed. Reg. at 34,956, Proposed 40 C.F.R. § 60.5820. This could include nearly any type of entity, “e.g., an entity that is regulated by the state, such as an electric distribution utility, or a private or public third-party entity.” Id. at 34,917. By contrast, the section 111(b) proposals for modified, and reconstructed facilities define the designated facility as a fossil fuel-fired EGU. This is a mismatch. EPA cannot regulate a different type of source under section 111(d) than the type of source that is regulated as a new (or modified or reconstructed) source under section 111(b). Because EPA has failed to meet these fundamental legal prerequisites for proposing and finalizing section 111(d) emission guidelines for CO2 emissions, the Proposed Guidelines are unlawful and must be withdrawn. VI. Definition of “Affected EGU” As discussed in Section V above, these Proposed Guidelines are unlawful because, among other reasons, the scope of entities that EPA has proposed to regulate under section 111(d) is significantly broader than the corresponding category of new sources that EPA has proposed to regulate under section 111(b). The facilities that would be subject to emission standards in state plans under the Proposed Guidelines are any “affected EGU,” along with any other “affected entity.” See, e.g., id. at 34,955-56, Proposed 40 C.F.R. § 60.5820 (defining “affected entity” and “emission standard”). The disconnect in coverage between these Proposed Guidelines and EPA’s January 2014 NSPS Proposal is further confirmed by EPA’s apparent confusion over even what sources should be considered “affected EGUs” under these different rules. Even if EPA were to adopt final emission guidelines under section 111(d) in which the “designated facility” is limited to “affected EGUs” as defined in the Proposed Guidelines, such 66 emission guidelines would still be unlawful because that definition is broader than the scope of new sources covered by the January 2014 NSPS Proposal. Specifically, the Agency has proposed to include as “affected EGUs” several types of Subpart Da units (i.e., coal- and oil/gas-fired boilers and IGCC units) that would be excluded under the applicability criteria in the January 2014 NSPS Proposal. If EPA adopts final emission guidelines, it must define “affected EGUs” to be consistent with the applicability criteria in its January 2014 NSPS Proposal. The category of existing sources subject to emission guidelines under section 111(d) cannot be broader than the corresponding category of new sources subject to standards of performance under section 111(b). In addition, as discussed in Section VI.B below, UARG notes that language in EPA’s proposals for new, modified, and reconstructed EGUs has introduced uncertainty as to which of two approaches the Agency will take to define those sources under the section 111(b) standards of performance (although EPA says the scope of EGUs that would be subject to numerical emission limits under section 111(b) would not differ under either approach). See 79 Fed. Reg. at 1461; Memorandum from EPA, Office of Air Quality Planning and Standards (“OAQPS”) to EGU NSPS Docket, “Amended Regulatory Text (Broad Applicability)” (June 2014), Docket ID No. EPA-HQ-OAR-2013-0603-0047 (“Broad Applicability Text”). This uncertainty has extended to the scope of regulation under the Proposed Guidelines. Because the same new, modified, and reconstructed EGUs would be subject to emission standards under either approach, defining the source category in the manner outlined in the Broad Applicability Text would not in fact, as EPA has suggested, expand the scope of existing EGUs that could be eligible for regulation under section 111(d). Each of the two scopes would trigger identical section 111(d) coverage for EGUs. 67 A. EPA’s Definition of “Affected EGU” in the Proposed Guidelines Must Retain the Exclusions for Subpart Da Units That Were Proposed for New Subpart Da Units. In its January 2014 proposed GHG NSPS for new EGUs, EPA proposed to define the source category of Subpart Da units for the purpose of regulating CO2 emissions to include any electric utility steam generating unit (i.e., a boiler or IGCC unit) that meets certain applicability criteria. 79 Fed. Reg. at 1459. Among these criteria, the proposed NSPS would apply to a boiler or IGCC unit only if it (1) “combusts fossil fuel for more than 10.0 percent of the heat input during any 3 consecutive calendar years,” and (2) actually “supplies more than one-third of its potential electric output and more than 219,000 MWh net-electric output to a utility power distribution system for sale on an annual basis.” 11 Id. at 1502, Proposed 40 C.F.R. § 60.46Da(a). The former applicability requirement is known as the “low fossil use criterion,” and the latter is known as the “one-third sales criterion.” 79 Fed. Reg. at 1459. In this rulemaking, EPA proposes to eliminate both of these criteria for Subpart Da units in its proposed definition of “affected EGUs,” effectively expanding its definition of those sources beyond the applicability criteria in its January 2014 NSPS Proposal. 79 Fed. Reg. at 34,954, Proposed 40 C.F.R. § 60.5795(b)(1). Although the preamble to the Proposed Guidelines purports to tie the definition of “affected EGU” to the applicability criteria in the January 2014 NSPS Proposal for new EGUs, EPA’s proposed definition would in fact eliminate the low fossil use criterion and the one-third sales criterion. See 79 Fed. Reg. at 34,854 (stating that emission guidelines would apply to existing sources that “in all other respects would meet the applicability 11 Because EPA incorporated substantially the same applicability criteria for electric utility steam generating units in both its Subpart Da and Subpart TTTT approaches in the January 2014 NSPS Proposal, UARG’s comments in this section apply equally to whichever approach the Agency adopts in that rulemaking. 68 criteria for coverage under the proposed GHG standards for new fossil fuel-fired EGUs” and that “[t]he rationale for this proposal concerning applicability is the same as that for the January 8, 2014 proposal”). EPA has not provided any legal or technical rationale for altering the scope of what Subpart Da units qualify as “affected EGUs” in this rulemaking, and for that reason alone, its proposal to eliminate the one-third sales criterion and the low fossil use criterion is invalid. See CAA § 307(d). Further, if EPA revises these Proposed Guidelines to define the “designated facility” as “affected EGUs,” it cannot eliminate the one-third sales criterion and the low fossil use criterion from the definition of “affected EGUs” because doing so would expand the scope of the Agency’s Proposed Guidelines under section 111(d) beyond the source category that EPA defined in its January 2014 NSPS Proposal for new units. Under the CAA, state plans may only impose obligations on an existing source “to which a standard of performance under this section would apply if such existing source were a new source.” CAA § 111(d)(1). Existing boilers that actually sell less than one-third of their potential electric output to the grid or that combust fossil fuel for less than 10 percent of their heat input over a 3-year period would not be subject to the standards of performance set forth in the January 2014 NSPS Proposal if they were new, and therefore cannot legally be subject to emission guidelines under section 111(d). EPA cannot define a source category differently for new, modified, and existing sources. If EPA adopts final emission guidelines, it should define “affected EGUs” to be consistent with the definitions in the January 2014 NSPS Proposal. B. The “Broad Applicability” Approach to the Proposed NSPS Would Not Support Regulating a Broader Universe of Sources Under the Proposed Guidelines. Although the Agency does not acknowledge it in the Proposed Guidelines, EPA appears to be contemplating two different approaches to defining the category of sources that it proposes 69 to regulate in the NSPS rulemakings for GHG emissions from new, modified, and reconstructed EGUs. As a result, there is uncertainty as to precisely how any final NSPS would define the source category for regulation, although EPA suggests that the same sources would be subject to numerical emission limits under either approach. See 79 Fed. Reg. at 1461. As noted above, the docket for the proposed NSPS contains two versions of the regulatory text as it would be amended by those proposals—a “Proposed Applicability” version and a “Broad Applicability”/“Alternate Applicability” version. But because both approaches would ultimately subject the same new, modified, and reconstructed EGUs to emission standards, both approaches would place the same limits on which existing EGUs are eligible for regulation under section 111(d)—limits that EPA has exceeded in these Proposed Guidelines, as discussed in Sections V and VI.A above. The “Broad Applicability” approach would not in fact expand the scope of EGUs subject to the Proposed Guidelines beyond what would be covered under the “Proposed Applicability” approach. The “Proposed Applicability” version of the NSPS contains the regulatory language that EPA has proposed to adopt governing new, modified, and reconstructed units under Subparts Da, KKKK, and TTTT. Memorandum from EPA, OAQPS to EGU NSPS Docket, “Amended Regulatory Text (Proposed Applicability)” (June 2014), Docket ID No. EPA-HQ-OAR-20130603-0044 (“Proposed Applicability Text”). That version applies to the source category of units as EPA proposed to regulate in the January 2014 NSPS Proposal and would exclude certain EGUs from the source category based on both design and operational characteristics. Under the Proposed Applicability Text, the category of boilers and IGCC units is defined as any electric utility steam generating unit that: (1) is capable of combusting more than 73 MW (250 MMBtu/hr) heat input of fossil fuel; (2) was constructed for the purpose of supplying more than 70 one-third of its potential electric output capacity and more than 219,000 MWh as net-electric sales on an annual basis; (3) combusts fossil fuel for more than 10.0 percent of the heat input during any three consecutive calendar years; and (4) actually supplies one-third or more of its potential electric output and more than 219,000 MWh net-electric output to a utility power distribution system for sale on an annual basis. 40 C.F.R. § 60.40Da(a)(1) (defining affected facility); Proposed Applicability Text, Subpart Da § 60.46Da(a)(1). The category of stationary combustion turbines is defined in largely the same manner, with the exception that criteria 2 and 4 are determined on a three-year rolling average basis and with an additional fifth criterion that the turbine must combust over 90 percent natural gas on a heat input basis on a three-year rolling average basis. Proposed Applicability Text, Subpart KKKK § 60.4305(c). 12 Any source that does not satisfy all of the relevant applicability criteria does not fall within the regulated source category and cannot be subject to obligations under section 111. By contrast, the “Broad Applicability” version would eliminate many of these applicability criteria and substantially broaden the source category. See Broad Applicability Text. 13 For boilers and IGCC units, that version would eliminate both the low fossil use criterion and—in the language of co-proposed Subpart TTTT—the one-third sales criterion, thus 12 The proposed regulatory text in the January 2014 NSPS Proposal for new EGUs and the June 2014 proposal for modified and reconstructed EGUs contained several substantial inconsistencies between the language of Subparts Da and KKKK and the new co-proposed Subpart TTTT, despite EPA’s stated desire that both approaches be “substantively the same.” 79 Fed. Reg. at 1454 n.100. UARG commented on the inconsistent language in its comments on the January 2014 NSPS Proposal. UARG New Source Comments at 198. EPA must ensure that the language of proposed Subpart TTTT is substantively identical to the language of proposed Subparts Da and KKKK. 13 The docket for EPA’s January 2014 NSPS Proposal for new EGUs also contained a broader alternative to that proposal’s applicability criteria. Memorandum from EPA, OAQPS to EGU NSPS Docket, “Alternate Applicability Approach” (Sept. 2013), Docket ID No. EPA-HQOAR-2013-0495-0062 (“Alternate Applicability Text”). Portions of the Broad Applicability Text addressing new EGUs differ from the Alternate Applicability Text. 71 expanding the source category to include non-fossil fuel boilers and low capacity factor boilers. Compare Proposed Applicability Text, Subpart Da § 60.46Da(a)(1)(ii), Subpart TTTT § 60.5509(a)(1), with Broad Applicability Text, Subpart Da § 60.46Da(a)(1)-(2), Subpart TTTT § 60.5509(a), 60.5580. Likewise, for all stationary combustion turbines, the Broad Applicability Text would eliminate all of the applicability criteria and expand the source category to include every stationary combustion turbine, including all simple cycle turbines, low capacity factor NGCC units, turbines that combust fuels other than natural gas, and turbines with design heat input less than 73 MW. Compare Proposed Applicability Text, Subpart KKKK § 60.4305(c), Subpart TTTT § 60.5509(a)(3), with Broad Applicability Text, Subpart KKKK § 60.4305(c), Subpart TTTT § 60.5509(a). Instead of defining the source categories to exclude these units that EPA has proposed to find are not appropriate for regulation, under the Broad Applicability Text, the Agency would consider many of these previously excluded sources to technically be “subject to” a standard of performance that simply specifies “no emission standard.” See, e.g., Broad Applicability Text, Subpart TTTT, Tbl. 1. Although these units would not be required to comply with a numerical emission limit, they would be subject to monitoring, reporting, and recordkeeping requirements under the applicable NSPS. 79 Fed. Reg. at 1461 (discussing Alternate Applicability Text). EPA makes no mention of the “Broad Applicability” approach in this rulemaking for existing EGUs, suggesting that EPA is no longer contemplating that approach for the proposed NSPS. In any event, EPA’s choice between these alternative approaches will not affect the scope of what existing EGUs may be eligible for regulation under section 111(d). Under section 111(d), a state plan may establish standards of performance only for “any existing source . . . to which a standard of performance under [section 111] would apply if such existing source were a 72 new source.” CAA § 111(d)(1). But the additional sources that the Broad Applicability Text would bring into the source category would be subject to “no emission standard.” “No emission standard” is not a “standard of performance.” Under the CAA, a “standard of performance” is defined as a “requirement of continuous emission reduction.” Id. § 302(l). Further, for the purposes of section 111, a “standard of performance” means a “standard for emissions of air pollutants” reflecting the degree of emission limitation achievable by applying the BSER that has been adequately demonstrated. Id. §111(a)(1). Therefore, a source that is subject to “no emission standard” cannot be one “to which a standard of performance under [section 111] would apply.” Id. § 111(d)(1). VII. EPA Has Unlawfully Ignored the Federal Power Act’s Division Between State and Federal Authority and Arbitrarily and Capriciously Ignored the Realities of Wholesale Electric Markets, Regional Transmission Grids, and Bulk Power System Reliability. EPA’s Proposed Guidelines violate the Federal Power Act (“FPA”), 16 U.S.C. §§ 791a et seq., by seeking to regulate matters that Congress either preserved as the province of exclusive state regulation or specifically assigned to FERC. The Proposed Guidelines make scant mention of the FPA or FERC, 14 but as discussed below, multiple elements of the Proposed Guidelines would effectively result in EPA making energy policy decisions in areas that the FPA expressly leaves to state control. EPA would thus be wielding regulatory authority over energy resource development and planning that the FPA denies even to FERC. Other aspects of the proposed rule, including in particular proposed “Building Block 2,” represent an unprecedented intrusion by EPA into areas that the FPA has made exclusively subject to FERC’s jurisdiction. As a result, the Proposed Guidelines also violate section 310 of the CAA, which specifies that the Act 14 See, e.g., 79 Fed. Reg. at 34,834, 34,888 n.235, and 34,914. None of these references mention FERC’s authority, or the limits on that authority, under the FPA. 73 “shall not be construed as superseding or limiting the authorities and responsibilities, under any other provision of law, of the Administrator or any other Federal officer, department, or agency.” CAA § 310(a). In addition, as discussed below, EPA’s Proposed Guidelines are founded upon numerous fundamental factual errors that violate the “reasoned decision-making” requirement under the Administrative Procedure Act (“APA”). 5 U.S.C. § 706(2)(A); see also Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). The Proposed Guidelines wrongly assume that various “technically feasible” but impracticable objectives can be achieved without serious harm to wholesale electric markets or to BPS reliability. EPA might have avoided these misunderstandings, and developed a more reasonable proposed rule, if it had seriously engaged with FERC. As discussed below, however, EPA has unreasonably excluded FERC from a substantive role in developing the Proposed Guidelines. It has irrationally deprived itself of the input of the federal agency that has real world expertise overseeing electricity markets, transmission grids, and electric reliability. The result is a proposal that has caused NERC and other independent expert entities to raise serious concerns that the Proposed Guidelines will jeopardize reliability. In short, EPA’s intrusion into regulatory spheres where it has no jurisdiction and no expertise has resulted in a proposal that will cause serious harm to wholesale electric markets and electric reliability if implemented, contrary to the requirements of reasoned decisionmaking and section 310(a) of the CAA. 74 A. The Proposed Guidelines Are Inconsistent With the FPA and Violate Section 310(a) of the CAA Because They Would Override the Division of Regulatory Authority Over the Generation, Transmission, Distribution, and Sale of Electric Energy Between the States and the Federal Government. Part II of the FPA, 15 16 U.S.C. §§ 824 et seq., was enacted in 1935 and has been amended several times in the intervening years, most recently by the Energy Policy Act of 2005, Pub. L. No. 109-58, §§ 1277-1298, 119 Stat. 594, 978-86 (2005). For nearly eight decades, the FPA has established a bright line of demarcation between the federal and state roles in overseeing the generation, transmission, distribution, and sale of electric energy. That line of demarcation is straightforward and well understood. Section 201(a) of the FPA states that “[f]ederal regulation . . . of the transmission of electric energy in interstate commerce and the sale of such energy at wholesale in interstate commerce is necessary in the public interest. . . .” 16 U.S.C. § 824(a). That provision also expresses the congressional determination that federal regulation should “extend only to those matters which are not subject to regulation by the States.” Id.; see also Fed. Power Comm’n v. S. Cal. Edison Co., 376 U.S. 205, 209 (1964). Section 201(b) establishes FERC’s exclusive jurisdiction “over all facilities for such transmission or sale of electric energy.” 16 U.S.C. § 824(b). 16 It further affirms that states, not the federal government, have regulatory authority over “facilities used for the generation of 15 Part I of the FPA, which was first enacted in 1920, addresses the licensing and oversight of hydroelectric facilities. It is not relevant to this proceeding. 16 Section 201(f) of the FPA exempts state, municipal, and federal governmental entities, as well as most rural electric cooperatives, from FERC’s jurisdiction under most provisions of the FPA. 16 U.S.C. § 824(f). Non-FERC-jurisdictional utilities that are not federal entities are instead generally subject to pervasive oversight by state, municipal, or other local bodies. Accordingly, to the extent that the proposed rule impacts the electric utility service by nonFERC-jurisdictional utilities it is invading the regulatory sphere that the FPA reserves for state and local regulation and not the sphere reserved for FERC. Thus, the Proposed Guidelines improperly invade local, state, and FERC jurisdictional spheres under the FPA. 75 electric energy” 17 as well as over electric distribution, retail sales, and intrastate transmission. Id. The Proposed Guidelines would impermissibly inject EPA into the separate spheres that the FPA has assigned to the states and to FERC. Section VII.A.1 of these comments addresses the Proposed Guidelines’ attempt to usurp states’ control over regulatory matters that the FPA expressly leaves to them in violation of the Tenth Amendment. Section VII.A.2 summarizes FERC’s relevant regulatory authority under the FPA and explains how EPA’s Proposed Guidelines conflict with and impermissibly supersede and limit that authority, in violation of section 310(a) of the CAA. For both reasons, EPA should not adopt the Proposed Guidelines. 1. The Proposed Guidelines Unlawfully Usurp Authority Over the Generation of Electricity, Resource Planning, and Related Matters that the FPA and Tenth Amendment Reserve to the States. The FPA draws clear lines between state and federal jurisdiction over electricity markets and generation facilities. S. Cal. Edison, 376 U.S. at 215 (“Congress meant to draw a bright line easily ascertained, between state and federal jurisdiction . . . .”). The Supreme Court has noted that FERC itself “has recognized that the States retain significant control over local matters even when retail transmissions are unbundled” and referred to FERC’s own acknowledgments of the limits on its jurisdiction in its landmark Order No. 888 18 to eliminate any uncertainty on this 17 The FPA also gives FERC jurisdiction over certain specific generation-related matters, in addition to wholesale sales by generators, e.g., allegations involving manipulation of FERCjurisdictional markets or compliance with FERC-approved reliability standards. As discussed below, however, “direct regulation” of generation and retail markets is reserved to the states. 18 Promoting Wholesale Competition Through Open Access Non-discriminatory Transmission Services by Public Utilities; Recovery of Stranded Costs by Public Utilities and Transmitting Utilities, Order No. 888, 61 Fed. Reg. 21,540 (May 10, 1996), FERC Stats. & Regs. ¶ 31,036 (1996), order on reh’g, Order No. 888-A, 62 Fed. Reg. 12,274 (Mar. 14, 1997), FERC Stats. & Regs. ¶ 31,048 (1997), order on reh’g, Order No. 888-B, 81 FERC ¶ 61,248 (1997), order on reh’g, Order No. 888-C, 82 FERC ¶ 61,046 (1998), aff’d in relevant part sub 76 point. New York v. FERC, 535 U.S. 1, 24 (quoting Order No. 888 at 31,782, n.543 (“‘Among other things, Congress left to the States authority to regulate generation and transmission siting’”), and n.544 (“‘This Final Rule will not affect or encroach upon state authority in such traditional areas as the authority over local service issues, including reliability of local service; administration of integrated resource planning and utility buy-side and demand-side decisions, including DSM [demand-side management]; authority over utility generation and resource portfolios; and authority to impose nonbypassable distribution or retail stranded cost charges.’”) (2002)). The Supreme Court further observed that the FPA’s “legislative history is replete with statements describing Congress’ intent to preserve state jurisdiction over local facilities.” Id. at 22. The Supreme Court has also emphasized that federal regulation by agencies other than FERC may not invade the domain that the FPA has preserved for the states. In Pacific Gas & Electric Co. v. State Energy Resources Conservation & Development Commission, 461 U.S. 190, 194 (1983), the Court concluded that federal regulation of the safety of nuclear power plants did not override state authority over “the regulation of electricity production.” The Supreme Court explained that states have “traditional authority over the need for additional generating capacity, the type of generating facilities to be licensed, land use, ratemaking, and the like.” Id. at 212. It noted that the only exception to state regulation of these activities was FERC’s authority under the FPA over interstate wholesale sales and transmission by FERC-jurisdictional utilities. Id. at 205. nom. Transmission Access Policy Study Group v. FERC, 225 F.3d 667 (D.C. Cir. 2000), aff’d sub nom. New York v. FERC, 535 U.S. 1 (2002). 77 Decisions of the U.S. Court of Appeals for the D.C. Circuit likewise confirm that federal regulation cannot directly or indirectly intrude into the sphere that the FPA reserves for state regulation. See, e.g., Elec. Power Supply Ass’n v. FERC, 753 F.3d 216 (D.C. Cir. 2014) (“EPSA”). EPSA vacated FERC Order No. 745, 19 which established revised rules for compensating demand-side resources in FERC-jurisdictional markets, because demand reductions did not constitute wholesale sales of energy and “the Federal Power Act unambiguously restricts FERC from regulating the retail market . . . .” Id. at 224; see also Duke Power Co. v. Fed. Power Comm’n, 401 F.2d 930, 935 (D.C. Cir. 1968) (explaining that the “major emphasis” of the FPA “is upon federal regulation of those aspects of the industry which—for reasons either legal or practical—are beyond the pale of effective state supervision”); Conn. Dep’t of Pub. Util. Control v. FERC, 569 F.3d 477, 481 (D.C. Cir. 2009) (reiterating established principle that FERC may not “directly regulat[e]” electric generation). Courts have also rejected attempts by FERC to use indirect means to regulate in areas that the FPA reserves to the states. See, e.g., EPSA, 753 F.3d at 221 (quoting Altamont Gas Transmission Co. v. FERC, 92 F.3d 1239, 1248 (D.C. Cir. 1996)) (noting that “FERC cannot ‘do indirectly what it could not do directly’”); Richmond Power & Light v. FERC, 574 F.2d 610, 620 (D.C. Cir. 1978). FERC itself has expressly acknowledged the rights of states to encourage the development of particular generation resources as a legitimate policy interest that is within their jurisdiction under the FPA. See, e.g., Order Accepting Proposed Tariff Revisions, Subject to 19 Demand Response Compensation in Organized Wholesale Energy Markets, Order No. 745, 76 Fed. Reg. 16,658 (Mar. 24, 2011), FERC Stats. & Regs. ¶ 31,322 (2011), order on reh’g, Order No. 745-A, 137 FERC ¶ 61,215 (2011), reh’g denied, Order No. 745-B, 138 FERC ¶ 61,148 (2012). 78 Conditions, and Addressing Related Complaint, PJM Interconnection, L.L.C., 135 FERC ¶ 61,022, at 61,106 (Apr. 12, 2011) (a state may “act within its borders to ensure resource adequacy or to favor particular types of new generation”). FERC has recognized that state law does not intrude on federal jurisdiction to the extent that a state is directing the planning and resource decisions of electric utilities under its jurisdiction. See Order on Complaint and Petition for Declaratory Order and on Petition for Enforcement, Midwest Power Sys., Inc., 78 FERC ¶ 61,067, at 61,246 (Jan. 29, 1997) (Iowa statute not preempted “to the extent that [it] require[s] [state utilities] . . . to purchase from certain types of generating facilities”); see also Order on Requests for Reconsideration, S. Cal. Edison Co., 71 FERC ¶ 61,269 (June 2, 1995) (states have broad powers to direct the planning and resource decisions of utilities under their jurisdiction). The FPA’s preservation of state jurisdiction over generation, resource planning, and other matters is consistent with basic principles of federalism. It likewise comports with the Tenth Amendment to the United States Constitution, which declares that “[t]he powers not delegated to the United States by the Constitution, nor prohibited by it to the states, are reserved to the states respectively, or to the people.” U.S. CONST. amend. X. Before Part II of the FPA was enacted in 1935, states pervasively regulated utilities within their borders based on their general police powers. The 1935 amendments to the FPA were adopted only after the Supreme Court held that states could not regulate interstate sales of electricity under the Commerce Clause. Pub. Utils. Comm’n v. Attleboro Steam & Elec. Co., 273 U.S. 83, 90 (1927). FERC was given authority over interstate transmission and interstate wholesale sales of electric energy solely to close this “Attleboro gap.” EPA’s Proposed Guidelines ignore the careful division of responsibilities embodied in the FPA—and the constitutional underpinnings for this division—and attempt to claim state 79 authority over generation, resource planning, and other matters for EPA. As stated by FERC Commissioner Clark in recent Congressional testimony, the Proposed Guidelines “ha[ve] the potential to comprehensively reorder the jurisdictional relationship between the federal government and states as it relates to the regulation of public utilities and energy development” and “fundamentally change the very fabric of how the utility industry is regulated in the country.” FERC Perspectives: Questions Concerning EPA’s Proposed Clean Power Plan and other Grid Reliability Challenges: Hearing Before the Subcomm. on Energy and Power of the H. Comm. on Energy & Commerce, 113th Cong. 4, 7 (July 29, 2014) (written testimony of Tony Clark, Comm’r, FERC), available at http://energycommerce.house.gov/hearing/fercperspectives-questions-concerning-epa%27s-proposed-clean-power-plan-and-other-grid (“Clark Testimony”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). The Proposed Guidelines would “dramatically alter the[] traditional lines of authority by creating a new paradigm of oversight of net carbon emission from a state” and potentially result in states “ceding ultimate authority of the regulation of their state’s public utilities and energy development to the EPA.” Id. at 4, 5. In enacting section 111(d) of the CAA, Congress could not possibly have intended to empower the EPA to take on such a monumental task. As the Supreme Court has recently stated, EPA cannot reasonably interpret a statute to authorize “an enormous and transformative expansion in EPA’s regulatory authority without clear congressional authorization.” UARG v. EPA, 134 S. Ct. at 2444. EPA can have no greater authority to address matters Congress has expressly reserved to state regulation than FERC has under the FPA. EPA should certainly have no more ability to “do indirectly what it could not do directly” than FERC would with respect to matters that the FPA left to the states. Similarly, the Tenth Amendment does not permit EPA to 80 displace sovereign state authority or “compel the States to enact or administer a federal regulatory program.” See, e.g., New York v. United States, 505 U.S. 144, 188 (1992). Accordingly, the Proposed Guidelines violate both the FPA and the Tenth Amendment and must be withdrawn. 2. The Proposed Guidelines Usurp Regulatory Authority that the FPA Gives Exclusively to FERC. The United States Supreme Court has repeatedly held that FERC’s exclusive jurisdiction over interstate transmission is extraordinarily broad because of the integrated nature of the interstate transmission grid and the laws of physics. For example, in New York v. FERC, 535 U.S. 1 (2002), the Court endorsed an earlier ruling, Federal Power Commission v. Florida Power & Light Co., 404 U.S. 453, 469 (1972), confirming FERC’s exclusive jurisdiction over the Florida Power & Light Company’s (“FPL”) transmission facilities. As the Court recounted: [S]ince electric energy can be delivered virtually instantaneously when needed on a system at a speed of 186,000 miles per second, such energy can be and is transmitted to FPL when needed from out-of-state generators, and in turn can be and is transmitted from FPL to help meet out-of-state demands; . . . there is a cause and effect relationship in electric energy occurring throughout every generator and point on the FPL, Corp, Georgia, and Southern systems which constitutes interstate transmission of electric energy by, to, and from FPL. New York v. FERC, 535 U.S. at 7 n.5 (internal quotation marks omitted). FERC’s jurisdiction over wholesale sales is comparably expansive. See, e.g., Schneidewind v. ANR Pipeline Co., 485 U.S. 293, 300 (1988) 20 (“The [FPA] long has been recognized as a comprehensive scheme of federal regulation of all wholesales of [energy] in 20 Schneidewind dealt with the Natural Gas Act rather than the FPA. Because “the relevant provisions of the two statutes are in all material respects substantially identical,” however, the Supreme Court has adopted an “established practice of citing interchangeably decisions interpreting the pertinent sections of the two statutes.” Ark. La. Gas Co. v. Hall, 453 U.S. 571, 577 n.7 (1981) (internal quotation marks omitted). 81 interstate commerce.”) (internal quotation marks omitted); Appalachian Power Co. v. Pub. Serv. Comm’n, 812 F.2d 898, 902 (4th Cir. 1987) (“FERC’s jurisdiction over interstate wholesale rates is exclusive . . . .”); see also Nantahala Power & Light Co. v. Thornburg, 476 U.S. 953, 966 (1986) (quoting S. Cal. Edison, 376 U.S. at 215-16) (“Congress meant to draw a bright line easily ascertained, between state and federal jurisdiction . . . . This was done in the [FPA] by making [FERC] jurisdiction plenary and extending it to all wholesale sales in interstate commerce except those which Congress has made explicitly subject to regulation by the States.”) (internal quotation marks omitted). Congress has explicitly prohibited EPA from interfering with FERC’s jurisdiction over these matters. Section 310(a) of the CAA specifies that the Act “shall not be construed as superseding or limiting the authorities and responsibilities, under any other provision of law, of the Administrator or any other Federal officer, department, or agency.” As discussed in the following sections, the Proposed Guidelines unlawfully interfere with the authority and responsibility given to FERC by Congress in the FPA. For this reason, the Proposed Guidelines are unlawful and must be withdrawn. a. The Proposed Guidelines Would Impermissibly Interfere With FERC’s Exclusive Authority Under Sections 205 and 206 of the FPA To Regulate Interstate Transmission and Wholesale Sale of Electric Energy. Sections 205 and 206 are the core FPA provisions that empower FERC to regulate both the interstate transmission and wholesale sale of electric energy. 16 U.S.C. §§ 824d, 824e. The Proposed Guidelines would intrude directly into this FERC-exclusive jurisdictional area in violation of section 310(a) of the CAA. Under section 205(c), jurisdictional public utilities must file with FERC “all rates and charges for any transmission or sale subject to the jurisdiction of the Commission, and the 82 classifications, practices, and regulations affecting such rates and charges, together with all contracts which in any manner affect or relate to such rates, charges, classifications, and services.” 16 U.S.C. § 824d(c). Courts have affirmed that this filing requirement is to be read broadly to encompass all classifications, practices, and procedures that “significantly affect” the price or non-price terms and conditions of FERC-jurisdictional services. See, e.g., City of Cleveland v. FERC, 773 F.2d 1368, 1376 (D.C. Cir. 1985) (“[T]here is an infinitude of practices affecting rates and service. The statutory directive must reasonably be read to require the recitation of only those practices that affect rates and service significantly . . . .”) (emphasis omitted). Under section 205(a), FERC may only accept filed rates, rules, and practices if it determines that they are “just and reasonable.” 16 U.S.C. § 824d(a). Section 206 requires FERC to revise any filed “rate, charge, or classification,” or “any rule, regulation, practice, or contract” affecting them, if it determines that they are “unjust, unreasonable, unduly discriminatory or preferential.” Id. § 824e(a). In recent decades, FERC has generally relied on the interplay of competitive market forces, subject to its oversight, to ensure that wholesale prices for electric energy are “just and reasonable.” EPA’s proposed Building Block 2 calls for “[r]educing emissions from affected EGUs in the amount that results from substituting generation at those EGUs with expanded low- or zerocarbon generation.” 79 Fed. Reg. at 34,836. The Proposed Guidelines assert that it would be technically feasible for states to use redispatch to achieve substantial emissions reductions. Id. at 34,862. They also posit that it would be technically feasible to increase the average utilization rates of NGCCs to seventy percent via redispatch. Id. at 34,863. EPA estimates the cost and economic impacts of achieving that NGCC utilization rate and concludes that they would be 83 reasonable. Id. (“We view these estimated costs as reasonable and therefore as supporting the use of a 70 percent utilization rate target . . . .”). Proposed Building Block 2 interferes with FERC’s authority under sections 205 and 206 in several critical ways. First and foremost, the Proposed Guidelines directly invade FERC’s core jurisdictional function by effectively determining that the price increases necessary to achieve a seventy percent NGCC utilization rate would be “just and reasonable.” Under the precedents described above, this determination is unquestionably for FERC, not EPA, to make. Building Block 2 would also require changes to the dispatch and redispatch rules applicable to EGUs. The Proposed Guidelines appear to acknowledge that dispatch and redispatch in an area encompassing nearly two-thirds of the load in the United States is conducted by ISOs and RTOs. See, e.g., 79 Fed. Reg. at 34,888 (“On the regional level, ISO/RTOs control dispatch and are responsible for reliable operation of the [BPS].”). But EPA fails to consider the legal implications of the fact that ISOs and RTOs dispatch EGUs and other resources pursuant to voluminous and detailed rules filed with FERC under section 205. See, e.g., FERC, Security Constrained Economic Dispatch: Definition, Practices, Issues, and Recommendations: A Report to Congress Regarding the Recommendations of Regional Joint Boards for the Study of Economic Dispatch Pursuant to Section 223 of the Federal Power Act As Added by Section 1298 of the Energy Policy Act of 2005 (July 31, 2006) (describing existing dispatch regimes, their emphasis on security-constrained economic dispatch, and the role of ISOs and RTOs in administering them), available at http://www.ferc.gov/industries/electric/indusact/joint-boards/final-cong-rpt.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record); Section III of the ISO New England Inc. Transmission, Markets, and Services Tariff (“Market Rule 1”) (“govern[ing] the operation of New England’s wholesale 84 electricity markets and includes detailed information on pricing, scheduling, offering, bidding, settlement, and other procedures related to the purchase and sale of electricity”), available at http://www.iso-ne.com/participate/ rules-procedures/tariff/market-rule-1. The map below depicts the territories subject to RTO dispatch. Of these, the “New England ISO” (“ISO-NE”), the “New York ISO” (“NYISO”), the “PJM ISO” (“PJM”), the “Midcontinent ISO” (“MISO”), the “Southwest Power Pool” (“SPP”), and the “California ISO” (“CAISO”) are all FERC-jurisdictional RTOs. The Electric Reliability Council of Texas (“ERCOT”) employs dispatch systems very similar to those used by other RTOs but is not subject to FERC regulation under section 205 given the physical separation of the ERCOT grid from the rest of North America. Figure 1: Map of Territories Subject to RTO Dispatch See http://www.ferc.gov/industries/electric/indus-act/rto.asp. There is no question that dispatch rules must be filed under section 205 because it is wellestablished that they significantly affect FERC-jurisdictional rates and non-rate terms and 85 conditions of service. There is also no question that ISOs and RTOs must comply with the filed dispatch rules and are potentially subject to civil penalties for deviating from them under section 316A of the FPA. 16 U.S.C. § 825o-1. FERC-approved RTO dispatch rules may be amended only by: (i) the RTOs, which must voluntarily submit amendments and persuade FERC to accept them under the standards of section 205; or (ii) by FERC, on its own initiative or in response to a third party complaint, under section 206. EPA and individual states have no authority to modify these dispatch rules. States also may not compel utilities, including ISOs and RTOs, to make involuntary section 205 filings. See, e.g., Atl. City Elec. Co. v. FERC, 295 F.3d 1 (D.C. Cir. 2002). In addition, EPA has ignored the fact that modifying redispatch rules to provide for the “environmental redispatch” contemplated by Building Block 2, see 79 Fed. Reg. at 34,863, would necessitate fundamental changes to existing ISO and RTO dispatch procedures that could conflict with FERC policies and with FERC’s FPA authority. All ISOs and RTOs use a form of regional, security-constrained, bid-based economic dispatch. These regimes are often designed to co-optimize the dispatch of resources to produce the most efficient combination of energy and “ancillary services” 21 needed to satisfy all requirements at the lowest total cost. FERC has consistently favored the use of bid-based economic dispatch mechanisms for at least the last 21 Section 1.2 of FERC’s pro forma Open Access Transmission Tariff (“OATT”) defines “ancillary services” as transmission or generation-related services “that are necessary to support the transmission of capacity and energy from resources to loads while maintaining reliable operation of the Transmission Provider’s Transmission System in accordance with Good Utility Practice.” Preventing Undue Discrimination and Preference in Transmission Service, Order No. 890, 72 Fed. Reg. 12,266 (Mar. 15, 2007), FERC Stats. & Regs. ¶ 31,241 (2007), order on reh’g, Order No. 890-A, FERC Stats. & Regs. ¶ 31,261 (2007), order on reh’g, Order No. 890-B, 123 FERC ¶ 61,299 (2008), order on reh’g, Order No. 890-C, 126 FERC ¶ 61,228 (2009), order on clarification, Order No. 890-D, 129 FERC ¶ 61,126 (2009) (“FERC pro forma OATT”). In ISO and RTO regions, several ancillary services may be procured from suppliers that offer to sell them at prices determined by ISO and RTO dispatch algorithms. 86 decade. No ISO or RTO has rules or dispatch software algorithms that would implement environmental dispatch. Nor do they have rules or software that would enable them to target a seventy percent utilization rate for NGCCs within their own regions, let alone on an interregional or even a national level. As FERC Commissioner Moeller has testified, “[s]imply put, the power plants with the lowest operating costs are called first to generate electricity—with various reliability requirements and other factors as part of the decision, depending on the structure of various markets.” FERC Perspectives: Questions Concerning EPA’s Proposed Clean Power Plan and other Grid Reliability Challenges: Hearing Before the Subcomm. on Energy and Power of the H. Comm. on Energy & Commerce, 113th Cong. 3 (July 29, 2014) (written testimony of Philip D. Moeller, Comm’r, FERC), available at http://energycommerce.house.gov/hearing/fercperspectives-questions-concerning-epa%27s-proposed-clean-power-plan-and-other-grid (“Moeller Testimony”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Switching to “environmental dispatch” would result in a system in which “units will be called to generate primarily based upon the emission profile of the unit.” Id. Commissioner Moeller has also informed Congress that FERC-jurisdictional RTO markets “would need to be fundamentally altered and redesigned to implement EPA’s proposal to accommodate environmental dispatch.” FERC Perspectives: Questions Concerning EPA’s Proposed Clean Power Plan and other Grid Reliability Challenges: Hearing Before the Subcomm. on Energy and Power of the H. Comm. on Energy & Commerce, 113th Cong. 3 (July 29, 2014) (Answers to Preliminary Questions for the FERC, Philip D. Moeller, Comm’r, FERC), available at http://energycommerce.house.gov/hearing/ferc-perspectives-questions-concerning- 87 epa%27s-proposed-clean-power-plan-and-other-grid (“Moeller Responses”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). It is, at best, disingenuous for the Proposed Guidelines to emphasize that ISOs and RTOs’ existing economic dispatch rules have accommodated existing environmental rules on a plantspecific basis, e.g., by permitting sellers to incorporate environmental costs into their offers. 79 Fed. Reg. at 34,862. Revising dispatch rules to focus on NGCC utilization rates instead of economic dispatch would in fact be a radical transformation. Id.at 34,863. FERC Commissioner Clark has explained that: [T]he Commission has allowed RTOs to acknowledge the operating limits of certain plants. Also the Commission allows generators to recognize various governmentally imposed costs like taxes and cap-and-trade schemes, but this is simply a matter of allowing generators to bid-in costs that they have legally incurred. To go beyond that by changing the fundamental market dispatch algorithms in the ways some have suggested would be a major change, to say the least. FERC Perspectives: Questions Concerning EPA’s Proposed Clean Power Plan and other Grid Reliability Challenges: Hearing Before the Subcomm. on Energy and Power of the H. Comm. on Energy & Commerce, 113th Cong. 3 (July 29, 2014) (Answers to Preliminary Questions for the FERC, Tony Clark, Comm’r, FERC), available at http://energycommerce.house.gov/hearing/ferc-perspectives-questions-concerning-epa%27sproposed-clean-power-plan-and-other-grid (“Clark Responses”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Even if EPA’s proposed overhaul of RTO dispatch mechanisms were technically feasible, which has not been established and which UARG does not concede, it could be approved only by FERC—not mandated by EPA or by states acting at EPA’s behest. FERC, not EPA, would have to conclude that environmental dispatch is just, reasonable, and not unduly discriminatory. The 88 Proposed Guidelines’ usurpation of these decisions from FERC interferes with the authority given to FERC by Congress in the FPA. As such, the Proposed Guidelines violate section 310(a) of the CAA, which expressly prohibits EPA from such interference with other federal agencies or federal laws. The same legal considerations apply to FERC-jurisdictional public utilities that have not transferred control of their transmission facilities to an RTO, i.e., principally utilities in the southeastern and western states. These utilities provide transmission service under tariffs that are modeled closely on FERC’s pro forma open-access transmission tariff (“OATT”). See FERC pro forma OATT (establishing the current version of the pro forma OATT used by most FERCjurisdictional utilities that are not participating in RTOs). Non-RTO tariffs also include provisions governing redispatch of EGUs, and the curtailment of transmission customers, in certain situations. See, e.g., id. §§ 13.3, 13.6, 15.4, 27, 30.5, and 33. Like RTO tariffs, non-RTO tariffs are filed with, and approved by FERC, and can be amended only by the filing utilities or by FERC—not EPA—under sections 205 or 206. The Proposed Guidelines violate section 310(a) of the CAA because they interfere with FERC’s exclusive authority under sections 205 and 206 of the FPA to regulate interstate transmission and wholesale sales of electric energy. As a result, the Proposed Guidelines are contrary to law, and EPA must withdraw them. b. The Proposed Guidelines Conflict With Section 202(a) of the FPA and Impermissibly Intrude on FERC’s Authority To Coordinate Regional Transmission. Section 202(a) of the FPA authorizes FERC to “divide the country into regional districts for the voluntary interconnection and coordination of facilities for the generation, transmission, and sale of electric energy” in order to assure “an abundant supply of electric energy” with “the greatest possible economy” and with regard to the “proper utilization and conservation of natural 89 resources.” 16 U.S.C. § 824a. But it is “definitively” clear that such coordination must be voluntary and that FERC has no power to compel any particular “‘technique of coordination.’” Atl. City Elec. Co., 295 F.3d at 12 (quoting Duke Power Co. v. Fed. Power Comm’n, 401 F.2d 930, 943 (D.C. Cir. 1968)); see also Cent. Iowa Power Coop. v. FERC, 606 F.2d 1156, 1161 (D.C. Cir. 1979) (holding that the voluntariness requirement of section 202(a) prevented FERC from revising a proposed power-pooling agreement that “provide[d] a mechanism for coordinated daily operation of generation facilities” to create an even more “fully integrated electric system with central dispatch of generating units”). Thus, for example, FERC may not force utilities to form or join an RTO under section 202(a). Recently, in South Carolina Public Service Authority v. FERC, 762 F.3d 41, 59 (D.C. Cir. 2014) (per curiam), the court held that FERC’s mandatory transmission planning reforms under Order No. 1000 were lawful because section 202(a) only governed the “coordinated operation of existing transmission facilities, not to the planning of future facilities” and was “textually limited to coordination for purposes of generation, transmission and sale, all activities that require operating facilities.” The decision thus affirmed that section 202(a) does not permit FERC to impose greater involuntary coordination involving existing facilities, such as operational EGUs. The Proposed Guidelines intrude on FERC’s section 202(a) jurisdiction by promoting the development of multi-state plans for the coordinated operation and dispatch of EGUs and by specifying rules to govern the development and evaluation of such plans. 79 Fed. Reg. at 34,851, 34,897. FERC has encouraged the voluntary formation of RTOs that dispatch EGUs within their regional boundaries. But the Proposed Guidelines call for greater regional coordination both through existing RTO structures and alternative multi-state arrangements. Id. 90 at 34,865 n.142, 34,867 (describing EPA’s use of six regions for analytical purposes that were “informed by” but do not correspond to RTO boundaries), 34,888 (proposing that regional coordination be undertaken through non-RTO entities). Although EPA professes that any additional regional coordination under the Proposed Guidelines would be voluntary, it appears that the rule would, as a practical matter, effectively require further coordination to address the integrated nature of the interstate grid and to reduce compliance costs. See, e.g., MISO, GHG Regulation Impact Analysis – Initial Study Results at 3, 11, 14 (Sept. 17, 2014) (finding that compliance costs in the MISO footprint could be reduced by approximately $3 billion annually by taking a MISO-wide approach to reducing CO2 emissions compared to making similar emission reductions on a “sub-regional” basis) (“Initial MISO Study Results”), available at https://www.misoenergy.org/WhatWeDo/EPARegulations/Pages/ 111(d).aspx (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). To the extent that the Proposed Guidelines would mandate new regional coordination, e.g., by effectively compelling the creation of multi-state dispatching arrangements in non-RTO regions, expanding the scope of existing RTO dispatch, or requiring RTOs to coordinate conflicting state plans, this would conflict with the voluntariness requirement of section 202(a) of the FPA, intrude into an area of FERC authority in violation of section 310(a) of the CAA, and constitute an unlawful EPA attempt to exercise even greater power over the coordination of transmission and sale of electricity than the FPA gives to FERC. 91 c. The Proposed Guidelines Unlawfully Intrude on FERC’s Authority Under Section 215 of the FPA To Ensure Reliability of the BPS. Section 215 of the FPA gives FERC the ultimate responsibility for the reliability of the BPS 22 by authorizing it to approve and enforce reliability standards developed by NERC and the various regional reliability entities. 16 U.S.C. § 824o. The Proposed Guidelines reference NERC, suggest that NERC and EPA have had some discussions about the Proposed Guidelines’ potential reliability implications, 79 Fed. Reg. at 34,836, and “mention[] the concept of reliability more than a hundred times,” Moeller Testimony at 6. But the Proposed Guidelines give no indication that EPA is prepared to defer to, or to even seriously account for, the extensive expertise that: (i) FERC has developed as the federal agency principally responsible for the reliability of the BPS; or (ii) NERC possesses as the independent body with the statutory responsibility under the FPA to “evaluate and improve the reliability of the BPS,” 23 and to develop and enforce reliability standards. Commissioner Moeller has warned that “it’s the details of calculating proper reserve margins and specific load pockets that matter from a reliability perspective.” Id. He has urged EPA to include FERC in evaluating the reliability questions raised by the proposed rule, and in potentially, developing answers: “Just as the Commission does not have expertise in regulating air emissions, I would not expect the EPA to have expertise on the intricacies of electric markets and the reliability implications of transforming the electric generation sector.” Id. at 7. NERC 22 The statute defines the “bulk-power system” as “(A) facilities and control systems necessary for operating an interconnected electric energy transmission network (or any portion thereof); and (B) electric energy from generation facilities needed to maintain transmission system reliability.” 16 U.S.C. § 824o(a)(1). 23 See NERC, Potential Reliability Impacts of EPA’s Proposed Clean Power Plan: Initial Reliability Review at ii (Nov. 2014) (“Initial NERC Report”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). 92 has also identified numerous concerns with EPA’s Proposed Guidelines, which are discussed in detail in Section VII.B.4 below. EPA’s lack of expertise and knowledge in this area shows in the Proposed Guidelines. The Proposed Guidelines assume that electrons are fully fungible, when in reality the regional nature of transmission grids and the prevalence of transmission constraints seriously restrict the extent to which one EGU can be substituted for another. EPA has ignored technical requirements, e.g., the need for voltage support and other ancillary services, that complicate the integration of intermittent renewable resources, and practical limitations on the development of new transmission and pipeline infrastructure. It has also ignored the requirements of existing FERC-approved reliability standards and failed to consider how the obligation of RTOs, transmission-owning utilities, and EGUs to comply with them might conflict with the Proposed Guidelines’ emission reduction goals. Unless EPA reverses course, invites FERC and others with reliability expertise to the table, and modifies the Proposed Guidelines as necessary to reflect input from FERC, NERC, and other industry experts, the rule will result in conflicts between FERC-approved enforceable reliability standards and EPA requirements. By disregarding the reliability standards and by making its own reliability-related determinations, EPA is improperly supplanting FERC as the ultimate authority over the reliability of the BPS in contravention of section 215 of the FPA and section 310(a) of the CAA. In sum, a federal agency may not assert authority over issues that Congress has exclusively reserved for another. See, e.g., Hunter v. FERC, 711 F.3d 155, 156 (D.C. Cir. 2013) (rejecting FERC’s attempts to regulate matters that the court believed were exclusively subject to the jurisdiction of the Commodity Futures Trading Commission). It is also clear that 93 “jurisdiction cannot arise from the absence of objection, or even from affirmative agreement” by a regulatory body, see Columbia Gas Transmission Corp. v. FERC, 404 F.3d 459, 463 (D.C. Cir. 2005), any more than it can be created by the consent of a private party. See, e.g., Weinberger v. Bentex Pharms., Inc., 412 U.S. 645, 652 (1973) (“Parties . . . cannot confer jurisdiction; only Congress can.”); Cal. Indep. Sys. Operator Corp. v. FERC, 372 F.3d 395, 398 (D.C. Cir. 2004); Michigan v. EPA, 268 F.3d 1075, 1081 (D.C. Cir. 2001) (“EPA is a federal agency—a creature of statute. It has no constitutional or common law existence or authority, but only those authorities conferred upon it by Congress. ‘It is axiomatic that an administrative agency’s power to promulgate legislative regulations is limited to the authority delegated by Congress.’ Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 208 . . . (1988). Thus, if there is no statute conferring authority, a federal agency has none.”). These principles are embodied in the CAA. Section 310(a) makes clear that EPA cannot use the CAA as a vehicle to usurp the authority of another federal agency. This is precisely what EPA proposes to do, however, in the Proposed Guidelines. Because the Proposed Guidelines intrude on the exclusive authority given to FERC in sections 205, 206, 202(a), and 215 of the FPA, they are arbitrary and capricious and contrary to law. As a result, EPA must withdraw the Proposed Guidelines. B. EPA Has Arbitrarily and Capriciously Ignored the Realities of the FERCRegulated Wholesale Electric Markets and Bulk Power System. EPA action must be overturned if it is “arbitrary, capricious, an abuse of discretion or otherwise not in accordance with law,” or if it is “in excess of statutory jurisdiction, authority, or limitations . . . .” CAA § 307(d)(9)(A), (C). An agency action will be upheld if it “articulate[s] a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.’” Motor Vehicle Mfrs. Ass’n, 463 U.S. at 43 (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168 (1962)). Agency factual findings will be upheld if 94 supported by substantial evidence. 16 U.S.C. § 825l(b). For the reasons set forth below, EPA’s Proposed Guidelines do not represent the kind of “reasoned decision-making” that the CAA, the APA, and the courts require. The Proposed Guidelines therefore must not be adopted in their current form and should instead, be withdrawn, or, at a minimum, substantially revised to correct their many factual errors and faulty assumptions. 1. EPA Has Unreasonably Failed To Consult With FERC Regarding the Proposed Guidelines’ Technical Feasibility and Impacts on Reliability and Markets. EPA claims that the Proposed Guidelines were “substantially informed by the extensive input from states and a wide range of stakeholders . . . since the summer of 2013.” 79 Fed. Reg. at 34,835; see also id. at 34,845-51. It is clear, however, from statements made by FERC’s commissioners, including their sworn testimony before the House of Representatives’ Committee on Energy and Commerce, Subcommittee on Energy and Power on July 29, 2014, that consultations with FERC have been extremely limited. It appears that only the office of FERC’s Chairman had any contact with EPA and that those contacts were both limited and “private and paperless.” Moeller Testimony at 7. No other Commissioners appear to have been involved in, or made aware of the details of, the limited FERC staff communications with EPA. The Commissioners’ testimony and responses to written questions also make it clear that FERC: (i) was not involved in preparing the Technical Support Document (“TSD”) entitled “Resource Adequacy and Reliability Analysis,” EPA-HQ-OAR-2013-0602-0368, that accompanied the Proposed Guidelines; (ii) has not conducted an independent study of the Proposed Guidelines’ reliability impacts; and (iii) did not verify the feasibility of EPA’s optimistic assumptions concerning the major new electric transmission and gas pipeline infrastructure development that must take place for the emission reductions contemplated by the Proposed Guidelines to be achieved (which are discussed separately below). To date, EPA has 95 also not accepted requests that FERC be given a formal role, along with NERC and other entities with reliability expertise, in discussing the reliability implications of the Proposed Guidelines. Moeller Testimony at 6 (“Going forward and at a minimum, I will reiterate my request for a formal role for the Commission with the EPA as it relates especially to the reliability implications of the proposal. Convening the appropriate reliability experts (including the Commission, and possibly the North American Electric Reliability Corporation, electric wholesale market operators, power generators, electricity consumers, along with input from the states) to examine the reliability implications is necessary to avoid additional unintended consequences.”). Commissioner Clark recently reiterated that EPA has not seriously engaged FERC on reliability issues. See John Sicliano, FERC’s Clark Urges States To Delay Climate Plans Until EPA Issues FIP Guide, INSIDEEPA/CLIMATE (Nov. 5, 2014), available at http://insideepa climate.com/climate-daily-news/fercs-clark-urges-states-delay-climate-plans-until-epa-issuesfip-guide (subscription required) (“‘I think there needs to be a much more transparent process in relation to how we’re modeling reliability and how reliability is being taken into consideration. I recently read in the press clippings that an EPA official said “we are working closely with FERC on these reliability matters.” And I read it, and I thought that’s news to me.’”). Most recently, the Chairmen of the House Energy and Commerce Committee and the Subcommittee on Energy & Power and the Ranking Member of the Senate Energy & Natural Resources Committee wrote a letter to FERC Chairman Cheryl A. LaFleur that noted the evident lack of substantive outreach by EPA. The letter asked FERC to take action to “begin to mitigate EPA’s failure to engage FERC and other relevant agencies . . . .” Letter from Fred Upton, The Hon. Chairman, House Energy & Commerce Comm., et al., to The Hon. Cheryl A. LaFleur, 96 Chairman, FERC at 4 (Nov. 24, 2014) (“November 24 Congressional Letter”), available at http://1.usa.gov/1xO4co5 (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). EPA’s refusal to consider FERC’s expertise regarding wholesale power markets, transmission operations, the real world impact of transmission constraints, and the BPS is inexplicable, arbitrary, and capricious. Article III courts frequently recognize and defer to FERC’s expertise in these matters, see, e.g., S.C. Pub. Serv. Auth., 762 F.3d at 54-55 (“[I]n raterelated matters, the court’s review of the Commission’s determinations is particularly deferential because such matters are either fairly technical or ‘involve policy judgments that lie at the core of the regulatory mission.’”) (quoting Alcoa Inc. v. FERC, 564 F.3d 1342, 1347 (D.C. Cir. 2009)). EPA’s failure to show similar deference has resulted in a proposed rule that fundamentally misunderstands the technical realities of the modern utility sector and is therefore not the product of reasoned decisionmaking. This conclusion has been reinforced by NERC, as described in further detail below. 2. EPA’s Erroneous Assumption that States Control the Dispatch of EGUs Is Likely to Result in Serious Distortions and Unintended Consequences. As discussed above, proposed Building Block 2 wrongly assumes that states have “command-and-control” style authority over the redispatch of EGUs when a significant majority of EGU dispatch is in reality managed by FERC-jurisdictional RTOs. These comments have already explained how this error results in Building Block 2 violating the FPA. But the same oversight also makes the Proposed Guidelines arbitrary and capricious. EPA has not considered the practical difficulties—and potential for state-federal and inter-agency disputes—that pressing states to pursue environmental dispatch is likely to entail. As Commissioner Clark told Congress: 97 [E]ven if all states in a region band together under the regional grid operator, any changes to the wholesale markets must necessarily be vetted and approved by FERC. The Commission would be charged with the awkward task of evaluating fundamental wholesale market design changes driven by environmental priorities approved by the EPA. Yet FERC is an economic and reliability regulator. Any decisions made by FERC must be rooted not in the Clean Air Act, but in our ‘just and reasonable’ and ‘not unduly discriminatory or preferential’ rate standard in the Federal Power Act. FERC’s ability to alter or reject an RTO-proposed compliance mechanism would present a conflict with EPA’s evaluation of the compliance plans. Absent Congress stepping in and clearly defining FERC authority and EPA authority, it is hard not to envision a future jurisdictional train wreck. Clark Testimony at 7. Similarly, if states do not coordinate their policies then “regional grid operators will be faced with an increasingly complex task of implementing multiple compliance mechanisms into what was once an efficiently-dispatched regional electric grid.” Id. at 6. Commissioner Moeller has also explained that because electricity markets are actually “interstate in nature” that the Proposed Guidelines’ “state-by-state approach results in an enforcement regime that would be awkward at best, and potentially very inefficient and expensive.” Moeller Testimony at 2. EPA’s failure to consider these complications only serves to increase the likelihood that they will pose serious problems. EPA’s approach does not represent the kind of “reasoned decision-making” that the APA requires. 3. EPA Underestimates the Challenges of Energy Infrastructure Development and the Impact of Transmission Constraints. Although the Proposed Guidelines are based on the erroneous presumption that states can dictate how EGUs are dispatched, at times they recognize the interconnected, integrated, and multi-state nature of the actual transmission grid. For example, EPA states that “[t]he utility power sector is unique in that, unlike other sectors where the sources operate independently and on a local scale, power sources operate in a complex, interconnected grid system that typically is regional in scale.” 79 Fed. Reg. at 34,844. EPA also suggests that “system operators typically 98 have flexibility to choose among multiple EGUs when selecting where to obtain the next MWh of generation needed” and that this results in electricity being highly “fungible.” Id. at 34,880. Similarly, the Proposed Guidelines assert that EPA has examined the “capability of the natural gas supply and delivery system to provide increased quantities of natural gas and the capability of the electricity transmission system” to accommodate a seventy percent utilization rate for NGCCs. Id. at 34,863. The Agency concludes that these systems would be capable of supporting the degree of increased NGCC utilization needed for states to achieve EPA’s proposed goals because: (i) the pipeline system is already supporting national NGCC utilization rates of sixty percent or higher during peak hours; (ii) its proposed state goals are supposedly sufficiently flexible so that “even if isolated natural gas or electricity system constraints were to limit NGCC unit utilization rates in certain locations in certain hours, this would not prevent an increase in NGCC generation overall across a state or broader region and across all hours;” and (iii) pipeline and transmission planners “have repeatedly demonstrated the ability to methodically relieve bottlenecks and expand capacity.” Id. at 34,864. Congressional testimony and other statements by multiple FERC commissioners (from both political parties) demonstrate that these assumptions are unrealistically simplistic, unreasonable, and incompatible with reasoned decisionmaking. Then-acting FERC Chairman LaFleur told Congress that “FERC staff emphasized that in light of EPA’s proposal to rely on increased capacity factors for natural gas fired generation resources, gas pipeline adequacy should be considered from a regional perspective, not just a national perspective, due to existing constraints on the system.” FERC Perspectives: Questions Concerning EPA’s Proposed Clean Power Plan and other Grid Reliability Challenges: Hearing Before the Subcomm. on Energy and Power of the H. Comm. on Energy & Commerce, 113th Cong. 4 (July 29, 2014) (written 99 testimony of Cheryl A. LaFleur, Acting Chairman, FERC), available at http://energycommerce.house.gov/hearing/ferc-perspectives-questions-concerning-epa%27sproposed-clean-power-plan-and-other-grid (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Commissioner Moeller stated that he was “skeptical of EPA’s contention that the modeled [pipeline] capacity increases are feasible by 2020 . . . partly due to the fundamental manner in which the proposed rule would change the way that electricity is dispatched.” Moeller Responses at 5. Moreover, expanding gas pipeline infrastructure has proven extremely difficult in recent years in certain regions that are widely understood to need more capacity, e.g., New England, which is particularly dependent on natural gas for heat and electricity generation during the winter. Commissioner Moeller explained that the fundamental problem has been finding ways to finance new construction when generators competing in RTO-administered electricity markets are reluctant to sign the long-term contracts that traditionally supported the financing of new pipelines before the markets were established. Id.; Moeller Testimony at 3-4. These financing difficulties are likely to be exacerbated by the Proposed Guidelines. See Moeller Testimony at 3-4. With respect to electric transmission, the Proposed Guidelines pay scant attention to the regional nature of transmission grids, the significant limitations imposed by transmission constraints, and the reality that such constraints are hardly “isolated” as the Proposed Guidelines suggest. 79 Fed. Reg. at 34,864. As Commissioner Moeller informed Congress: As we have seen with the implementation of EPA’s mercury rule (MATS), load pockets matter because the laws of physics trump written words. Although a specific generating plant may not contribute significant power to the grid, its other outputs such as voltage support or “inertia” qualities may contribute significantly to grid stability. Moreover, the details of how reserve margins are calculated can have a significant impact on the ability of excess capacity in one load pocket to 100 transfer power to another load pocket that is short. These challenges can be addressed, but it takes engineering expertise, especially when designing optimal infrastructure improvements. Moeller Testimony at 6. In connection with Building Block 3, i.e., promoting greater deployment of intermittent renewable resources, EPA failed to consider transmission grid “integration issues” (e.g., voltage control, natural gas backup power, etc.) that would have to be addressed to accommodate the substantial influx of renewable resources that the Proposed Guidelines contemplate. See id. at 7. Commissioner Moeller noted that FERC had not been asked to, and had not, analyzed these significant implications. See Moeller Responses at 5. He also stated that he did not believe that FERC had “studied whether additional transmission lines would need to be built under the EPA Proposal to integrate more renewables, where the lines may be built, and how long it may take to site, permit, and build these lines” or estimated the cost of transmission necessary to supply increased renewable resources. Id. at 6. 4. EPA Relies on Assumptions that Independent Expert Analyses Demonstrate Are Faulty. EPA’s assumption that the Proposed Guidelines’ emission reductions could be achieved on the contemplated timetable without significant negative impacts on electric reliability and reliability markets is contradicted by independent expert analyses. Most notably, NERC has issued an initial reliability review of EPA’s Proposed Guidelines that has identified numerous serious concerns. See Initial NERC Report. As the Initial NERC Report explains, NERC is independent and impartial. The Initial NERC Report is a detailed review by the entity with the greatest collective expertise and experience regarding the reliability of the BPS in the United States. It is not intended to advocate a policy position on behalf of any entity or to promote any specific compliance approach. Id. at 1. But it does establish that “detailed and thorough analysis 101 will be required to demonstrate that the proposed rule and assumptions are feasible and can be resolved consistent with the requirements of BPS reliability.” Id. NERC is planning to conduct three more special assessments between now and December 2016 to evaluate more comprehensively the impact of EPA’s Proposed Guidelines and of the state plans that will be submitted if the Proposed Guidelines are not withdrawn. Id. at 4. The mere fact that NERC is already questioning the feasibility of the Proposed Guidelines ought to be enough to cause EPA to pause and reconsider them at a fundamental level. The specific flaws the Initial NERC Report identifies with each of the four proposed Building Blocks would make it irrational, arbitrary, and capricious for EPA to fail to do so. With respect to Building Block 1, the Initial NERC Report acknowledges that improving the existing coal generation fleet’s efficiency by six percent “may be difficult to achieve.” Id. at 8. It explains that EPA calculated unit-specific heat rates using gross generation data from the CEMS, thereby ignoring “generation-reducing effects from post-combustion environmental controls . . . .” Id. EPA also used net-generation data, without considering the fact that these retrofits would reduce the net output and net heat rate efficiency of affected units. “Not considering these reductions creates an inconsistent approach, especially considering that most coal-fired EGUs will require control retrofits to comply with environmental regulations, such as [MATS] and Section 316(b) of the Clean Water Act.” Id. EPA also assumed that turbine overhauls would result in major heat rate improvements but “[r]egular turbine overhauls are generally not practical or economical” because they require units to be out of service for protracted periods. Id. The notion that there are significant efficiencies to be obtained through overhauls also ignores the reality that generation owners have strong incentives to operate at peak efficiency, including profit incentives in FERC-regulated wholesale power markets. 102 Moreover, the Initial NERC Report notes that EPA fails to account for factors that have “profound effects on the process efficiency of a coal-fired EGU,” including: (1) subcritical versus supercritical boiler designs; (2) fluidized bed combustion, integrated gasification combined-cycle (IGCC), and pulverized coal; (3) unit size and age; and (4) coal quality variations in moisture and ash (i.e., every 5 percent change in coal moisture results in a 1 percent change in boiler heat rate efficiency). Id. (footnotes omitted). The Initial NERC Report also explains how EPA failed to consider the interaction between Building Blocks 1 and 2. Assuming that Building Block 2 will cause coal units to cycle more often, resulting in lower capacity factors and higher heat rates, the heat rate improvements assumed by Building Block 1 are unlikely to be achievable. Id. Although EPA recognized capacity effects in its regression analysis, it did not evaluate the effects of lower capacity factors from dispatching natural gas generation before coal generation. Id. With respect to Building Block 2 itself, the Initial NERC Report identifies “a number of reliability concerns regarding increased reliance on natural-gas-fired generation that should be evaluated.” Id. at 9. First, NGCC units have traditionally been used for load following purposes, and while some are capable of operating at a high capacity factor, they are generally less wellsuited to serve as baseload resources than coal generation. Id. Second, the Proposed Guidelines would accelerate an ongoing utility industry shift from coal- to gas-fired generation, threatening the fuel diversification that utilities traditionally rely on to help to preserve reliability. Id. Third, as the events of the winter 2014 “polar vortex” demonstrated, natural gas generation can face serious fuel availability issues during extended periods of cold weather. The Initial NERC Report expresses concern that over-reliance on natural gas generation will degrade utilities’ ability to manage cold weather conditions and threaten reliability. Id. Fourth, Building Block 103 2’s assumption that “the power industry in aggregate can support higher gas consumption without the need for any major investments in pipeline infrastructure” is unrealistic. Id. at 10. There are “critical areas,” including Arizona, Nevada, and New England that already face pipeline constraints, and those areas will need still more pipeline capacity to the extent that coal generation is replaced by natural gas generation. Id.; see also Section VII.B.3 (noting that FERC Commissioners have highlighted these same infrastructure issues). With respect to Building Block 3, the Initial NERC Report identifies various flaws in the Proposed Guidelines’ assumptions regarding the preservation of nuclear generation resources and the expansion of renewable energy. Initial NERC Report at 11-13. Among other issues, EPA’s assumed “rapid growth in non-hydro renewable generation exceeds its own forecast” in the RIA of 356 TWh/year by 2030. Id. at 11. EPA also overlooked real world differences among state renewable portfolio standards that are contrary to EPA’s assumptions. Id. at 12-13. Furthermore, EPA completely failed to address the well-known grid reliability challenges, which have also been noted by FERC’s Commissioners, see Section VII.B.3, associated with increased dependence on “variable” resources. Initial NERC Report at 13. Specifically: Conventional generation (e.g., steam and hydro), with large rotating mass, has inherent operating characteristics, or [Essential Reliability Services], needed to reliably operate the BPS. These services include providing frequency and voltage support, operating reserves, ramping capability, and disturbance performance. Conventional generators are able to respond automatically to frequency changes and historically have provided most of the power system’s essential support services. As variable resources increase, system planners must ensure the future generation and transmission system can maintain essential services that are needed for reliability. A large penetration of [variable energy resources] will also require maintaining a sufficient amount of reactive support and ramping capability. More frequent ramping needed to provide this capability could increase cycling on conventional generation. This could contribute to increased maintenance hours or higher 104 forced outage rates, potentially increasing operating reserve requirements. While storage technologies may help support ramping needs, successful large-scale storage solutions have not yet been commercialized. Id. (footnotes omitted). Thus, NERC makes it clear that increased dependence on renewables may seriously strain the utility industry’s ability to provide or procure essential reliability services. Id. at 2. With respect to Building Block 4, the Initial NERC Report notes that EPA has projected that energy efficiency will grow faster than electricity demand, an assumption with “complex” implications. Among other things, if EPA’s aggressive rate of efficiency growth cannot be met, it will be necessary to adopt additional carbon reduction measures to meet the state targets, presumably through further reductions in fossil fuel generation that may not be achievable. Id. at 2, 14,16. More generally, the Initial NERC Report indicates that, even if the Proposed Guidelines could be implemented, they will likely take longer to implement than EPA envisions. State and regional plans must be approved by the EPA, which is anticipated to require up to one year, leaving as little as six months to two years to implement the approved plan. Areas that experience a large shift in their resource mix are expected to require transmission enhancements to maintain reliability. Constructing the resource additions, as well as the expected transmission enhancements, may represent a significant reliability challenge given the constrained time period for implementation. While the EPA provides flexibility for meeting compliance requirements within the proposed time frame, there appears to be less flexibility in providing reliability assurance beyond the compliance period. Id. at 3-4. NERC’s 2014 Long-Term Reliability Assessment highlights the issues raised in the Initial NERC Report and emphasizes that environmental regulations, including specifically the Proposed Guidelines, are creating uncertainty and require assessment. NERC, 2014 Long-Term Reliability Assessment at 8-17 (Key Reliability Finding #2), available at http://www.nerc.com/ 105 pa/RAPA/ra/Reliability%20Assessments%20DL/2014LTRA_ERATTA.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). The November 24 Congressional Letter asked FERC to hold a technical conference with participation by the Department of Energy, the National Association of Regulatory Utility Commissioners, and stakeholders to “examine the significant concerns, as identified by NERC’s report, that EPA’s Clean Power Plan presents for grid reliability.” November 24 Congressional Letter at 3. That letter emphasized that “EPA lacks the mission and the expertise to determine what is necessary to maintain the reliability of the nation’s electric grid. Indeed, Congress specifically established the Electric Reliability Organization for this very purpose.” Id. at 2 (emphasis in original) (NERC is the FERC-approved “Electric Reliability Organization” under section 215 of the FPA). Beyond NERC’s concerns, the RTOs, which actually operate and plan for the future of the BPS and thus have enormous reliability expertise, have also raised concerns with the Proposed Guidelines. For example, on September 16, 2014, SPP informed Missouri regulators that it “anticipates that there will be significant reliability impacts as a result of compliance with [the Proposed Guidelines].” Responsive Comments of the Southwest Power Pool, Inc., In the Matter of an Investigation of the Cost to Missouri’s Electric Utilities Resulting from Compliance with Federal Environmental Regulations at 3 (Sept. 16, 2014), Missouri Public Service Commission, File No. EW-2012-0065, Item No. 97, available at https://www.efis.psc.mo.gov/mpsc/commoncomponents/view_ itemno_details.asp?caseno=EW2012-0065&attach_id=2015006331 (“SPP Comments”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). SPP noted that addressing these reliability impacts would likely require transmission upgrades but that because transmission upgrades 106 normally require up to eight and a half years to be placed into service such solutions would likely not be feasible given the Proposed Guidelines’ 2020 compliance deadline. Id. An October 2014 assessment of reliability impacts by SPP concluded that it was: clear that new generation and transmission expansion will be necessary to maintain reliability during summer peak conditions if EPA’s projected generator retirements occur. Even the scenario that assumes optimal resource expansion using new natural gas-fired resources could be problematic during extreme winter load conditions with gas supply and delivery challenges. SPP, Reliability Impact Assessment of the EPA’s Proposed Clean Power Plan, at 6 (Oct. 8, 2014), available at http://www.spp.org/publications/CPP%20Reliability%20Analysis% 20Results%20Final%20Version.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). In addition, “[m]ore comprehensive planning efforts with stakeholders and new tools/metrics will be required. Unprecedented coordination and cooperation beyond regional planning efforts will be necessary but may not be timely given significant challenges with interregional planning and necessary system expansion.” Id. Similarly, the MISO indicated that as much as an additional 14 gigawatts (“GW”) of coal-fired generating capacity in the MISO footprint could be at-risk for retirement because of the Proposed Guidelines. Initial MISO Study Results at 13-14. This amount is beyond the 12,600 MW within MISO’s area that is slated to retire by the end of 2016 to comply with MATS. ERCOT has also evaluated the potential implications of the Proposed Guidelines for grid reliability and the non-FERC jurisdictional wholesale electricity market in Texas. ERCOT’s analyses indicated that: [I]mplementation of the proposed Clean Power Plan will have a significant impact on the planning and operation of the ERCOT grid. ERCOT estimates that the proposed CO2 emissions limitations will result in the retirement of between 3,300 MW and 8,700 MW of coal generation capacity, could result in transmission reliability issues due to the loss of generation resources in and around major urban centers, and will strain ERCOT’s ability to integrate new intermittent renewable 107 generation resources. The Clean Power Plan will also result in increased energy costs for consumers in the ERCOT region by up to 20% in 2020, without accounting for the costs of transmission upgrades, procurement of additional ancillary services, energy efficiency investments, capital costs of new capacity, and other costs associated with the retirement or decreased operation of coal-fired capacity in ERCOT. ERCOT, ERCOT Analysis of the Impacts of the Clean Power Plan (Nov. 17, 2014), available at http://www.ercot.com/content/news/presentations/2014/ERCOTAnalysis-ImpactsCleanPower Plan.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). NYISO also noted that the Proposed Guidelines “present[] potentially serious reliability implications . . . .” NYISO, Comments of the New York Independent System Operator, Inc. on the Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units (Dec. 1, 2014). MISO has also expressed reliability concerns regarding the Proposed Guidelines. Specifically, MISO says that the Proposed Guidelines’ interim goals “will force decisions that pit environmental compliance against electric reliability.” Letter from John R. Bear, President and CEO, MISO, to The Hon. Gina McCarthy, Adm’r, EPA, at 4 (Nov. 25, 2014) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). To meet the interim goals, reductions in CO2 emissions will need to begin in 2020, and MISO says “there will not be nearly enough time to plan for the replacement capacity, transmission upgrades, and natural gas delivery infrastructure that will be needed to maintain reliability and resource adequacy.” Id. at 2. More generally, RTOs have called on EPA to establish some form of “reliability safety valve” similar to the one-year compliance extension that is available to avoid retirement-related reliability impacts from the MATS compliance deadline. The Initial NERC Report states that 108 even if EPA were to do so, a “reliability safety valve will be of limited utility if the EPA’s proposal is implemented as currently designed . . . .” Initial NERC Report at 22. 5. EPA Has Not Accounted for the EPSA Decision’s Potential Impact on Demand-Side Resources. Building Block 4 assumes that states will be able to reduce “emissions from affected EGUs in the amount that results from the use of demand-side energy efficiency that reduces the amount of generation required.” 79 Fed. Reg. at 34,836. The Proposed Guidelines focus on the role played by state-administered programs but appear to ignore the fact that demand-side resources have been active participants in RTO-administered markets, including especially RTOadministered capacity markets, for many years. The Proposed Guidelines likewise do not address the D.C. Circuit’s decision in EPSA, which both vacated FERC Order No. 745 and arguably prohibits some demand-side resources from directly participating in FERCjurisdictional markets, although it does not affect state-administered demand-response programs. 24 It would not be consistent with reasoned decisionmaking for EPA to continue to simply assume that demand-side resources will play an expanded role in supporting emissions reductions given EPSA’s potential implications. It might be that the change from FERC to state regulation of demand response will be beneficial in the long term, but EPA must recognize that demand response is likely entering a potential period of transition and uncertainty that makes the Building Block 4 assumptions unrealistic. See Mot. of FERC To Stay Issuance of Mandate at 11, Elec. Power Supply Ass’n v. FERC, 753 F.3d 216 (D.C. Cir. 2014) (No. 11-1486) (“Whether the Opinion is read narrowly or broadly, its implementation will involve both setting new 24 EPSA was decided less than a month before the Proposed Guidelines were published in the Federal Register. 109 prospective rules and unwinding demand response participation in certain wholesale markets—a task with significant market and financial implications that go far beyond the typical burdens of litigation . . . .”). 6. EPA Inaccurately Describes RTO Capacity Markets and Overstates Their Potential To Advance the Proposed Guidelines’ Goals. Finally, the Proposed Guidelines suggest that: In most states where generation is no longer subject to price regulation, [RTOs] . . . ensure the adequacy of future generation supplies by administering auctions for forward capacity. In these auctions, owners of existing EGUs (with consideration of building blocks 1 and 2), developers of new EGUs including renewable generating capacity (building block 3), and developers of demand-side resources (building block 4) all compete to provide potential resources for meeting the projected demand for electricity services. 79 Fed. Reg. at 34,881 (footnote omitted). The Proposed Guidelines suggest that RTO capacity markets can serve as the functional equivalent of traditional state integrated resource planning mechanisms and that the BSER for CO2 emissions should account for this. Id. The Proposed Guidelines’ description of RTO capacity markets is inaccurate because only two FERC-jurisdictional RTOs, ISO-NE and PJM, have what are generally considered to be “forward capacity markets.” The NYISO is currently evaluating a possible move to a forward capacity market design but has concluded on two prior occasions that such a change was not needed in New York. See, e.g., Written Statement of Emilie Nelson, Vice President – Market Operations, on Behalf of New York Independent System Operator, Inc., FERC Docket No. AD14-18-000 at 26-27 (Nov. 5, 2014) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Three FERC-jurisdictional RTOs, namely CAISO, MISO, and SPP, do not have any form of mandatory capacity markets. See, e.g., Commission Staff Report, Centralized Capacity Market Design Elements, FERC Docket No. AD13-7-000 at 11-15 (Aug. 23, 2013) (“FERC Staff Report”) (describing various RTO capacity market 110 designs), available at http://www.ferc.gov/CalendarFiles/20130826142258- Staff%20Paper.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). As noted above, the ability of demand-side resources to continue to participate in FERCjurisdictional capacity markets is also in doubt in light of the EPSA decision. Moreover, the Proposed Guidelines appear to overstate the potential of the capacity markets that actually exist in ISO-NE, the NYISO, and PJM to function as de facto approximations of traditional integrated resource planning systems. RTO capacity markets were not intended or designed to play such a role. Their principal purpose is to “ensure resource adequacy at just and reasonable rates through a market-based mechanism that is not unduly discriminatory or preferential as to the procurement of resources.” FERC Staff Report at 2. FERC is considering a number of complex and controversial issues related to this goal, such as the types of payments to be made to different kinds of resources and the extent to which new resources should be subject to market power mitigation measures if they are deemed to be “uneconomic.” See, e.g., PJM Staff Proposal, PJM Capacity Performance Proposal (Aug. 20, 2014) (preliminary proposal to establish new categories of capacity products to address defects in existing PJM capacity market design), available at http://www.pjm.com/~/media/documents/ reports/20140820-pjm-capacity-performance-proposal.ashx (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record); FERC, Supplemental Notice of Technical Conference, Centralized Capacity Markets in Regional Transmission Organizations and Independent System Operators, FERC Docket No. AD13-7-000 (Aug. 23, 2013) (identifying and inviting comment on various capacity market design questions), available at http://www. ferc.gov/CalendarFiles/20130823115125-AD13-7-000TC1.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Consequently, EPA has unreasonably 111 overstated the potential for RTO-administered capacity markets to support the Proposed Guidelines’ objectives. VIII. If Lawful, The Proposed Guidelines Would Violate the Fifth Amendment Because Uncompensated Takings of Property Will Result. As these comments explain, EPA’s Proposed Guidelines are unlawful for numerous reasons. But assuming arguendo that the Proposed Guidelines were otherwise consistent with the CAA and other federal statutes (which they are not), they would nevertheless constitute an unconstitutional taking of property without just compensation. The Proposed Guidelines violate the CAA, the FPA, and are inconsistent with key court rulings. Nevertheless, for the sake of argument, UARG assumes EPA’s Proposed Guidelines are authorized by statute and otherwise lawful for the purposes of this evaluation of whether the Proposed Guidelines, under those circumstances, would constitute a taking under the Fifth Amendment. As EPA acknowledges, to meet the goals set forth in the Proposed Guidelines, numerous coal-fired EGUs throughout the country will need to be shut down. 79 Fed. Reg. at 34,933 (estimating retirement of 46-49 GW of coal-fired EGUs and 16 GW of oil/gas steam EGUs under EPA’s Building Block Option 1, and retirement of 24-32 GW of coal-fired EGUs and 3-5 GW of oil/gas steam EGUs under its Option 2 approach); see also EPA, Goal Computation Technical Support Document (June 2014), Docket ID No. EPA-HQ-OAR-2013-0602-0460 (“Goal Computation TSD”); id. at App. 1, Docket ID No. EPA-HQ-OAR-2013-0602-0255. And for those coal-fired EGUs that do remain open, many will have their operations so severely curtailed that it will result in a partial taking. The Fifth Amendment of the Constitution prohibits the taking of “private property . . . without just compensation.” U.S. CONST. amend. V. The Supreme Court long ago extended the takings concept to regulation that goes too far: “The general rule . . . is that while property may 112 be regulated to a certain extent, if regulation goes too far it will be recognized as a taking.” Pa. Coal Co. v. Mahon, 260 U.S. 393, 415 (1922). A total, or per se, taking occurs where regulations “completely deprive an owner of ‘all economically beneficial use’ of her property.” Lingle v. Chevron U.S.A. Inc., 544 U.S. 528, 538 (2005) (emphasis in original) (quoting Lucas v. S.C. Coastal Council, 505 U.S. 1003, 1019 (1992)). Partial takings, like total takings, must be compensated under the Fifth Amendment. Courts evaluate partial takings claims by making “essentially ad hoc, factual inquiries” into three primary factors: (1) the character of the government action; (2) the economic impact of the regulation on the claimant; and (3) the extent to which the regulation interferes with distinct investment-backed expectations. Penn Cent., 438 U.S. at 124. The constitutional concept of “property” is broad enough to encompass the EGUs affected by the Proposed Guidelines, the generating capacity of those EGUs, and their remaining useful life (and corresponding future generating capacity). See United States v. Gen. Motors Corp., 323 U.S. 373, 378 (1945) (“[Property] deals with what lawyers term the individual’s ‘interest’ in the thing in question . . . . [T]he constitutional provision is addressed to every sort of interest the citizen may possess.”); see also Kaiser Aetna v. United States, 444 U.S. 164, 179 (1979) (stating that the government must give just compensation for taking “sufficiently important” “expectancies embodied in the concept of ‘property’”) (emphasis added). Takings claims may be brought for the taking of real property, see Lucas, 505 U.S. at 1019, tangible property, see Andrus v. Allard, 444 U.S. 51, 65 (1979), and intangible property, see Ruckelshaus v. Monsanto Co., 467 U.S. 986, 1003-04 (1984). The EGUs at issue here are, themselves, obviously tangible property, their generating capacity is intangible property, and the remaining useful life (and associated future generating capacity) of the EGUs represent investment-based 113 expectations in the property, all of which are encompassed by the protections of the Fifth Amendment. The economic value of an EGU lies in the generation of electricity. Thus, eliminating or curtailing the generating capacity or remaining useful life of an EGU, as the Proposed Guidelines would do, directly interferes with an owner’s primary “‘interest’ in the thing in question.” Gen. Motors Corp., 323 U.S. at 378. Accordingly, if the Proposed Guidelines are otherwise lawful—which they are not—their promulgation would require just compensation for all owners of EGUs that are forced to shut down or severely curtail operations because of EPA’s regulatory action. A. The Proposed Guidelines, if Lawful, Would Result in a Total Taking of Any EGU Forced To Shut Down To Ensure Compliance with the Guidelines. EPA admits that the Proposed Guidelines will result in the forced retirement of coal- and oil-fired EGUs based on EPA-established state emission goals that can be met only by shifting generating capacity away from these EGUs to other sources of generation. 79 Fed. Reg. at 34,933; Goal Computation TSD, App. 1. If EPA were authorized to issue the Proposed Guidelines, the premature retirement of these EGUs would constitute a total taking because it would completely deprive the owners of these EGUs of all economically beneficial use of their property. See Lingle, 544 U.S. at 537. Because EGUs will be required to shut down to meet the state goals set forth in the Proposed Guidelines, the owner of the EGU would lose its capital investment in the EGU, the generating capacity of the EGU, and the remaining useful life of the EGU (and its associated future generating capacity). Although the Supreme Court has permitted the government, without compensation, to restrict lawful property from being put to unlawful uses, see Mugler v. Kansas, 123 U.S. 623 (1887) (breweries banned under Prohibition Era law), or to restrict the use or sale of a previously legal product to the extent that it is nearly valueless, see Andrus v. Allard, 444 U.S. 51 (1979), 114 the Court has not allowed government appropriation of property without compensation where there is no showing that the property is a nuisance or creates an impending harm to others, see Mugler, 123 U.S. at 669 (“In the one case, a nuisance only is abated; in the other, unoffending property is taken away from an innocent owner.”); cf. Lucas, 505 U.S. at 1030. There is no nuisance or impending harm here. Indeed, the EGUs that would be shut down under the Proposed Guidelines are heavily regulated facilities that operate subject to numerous permits, including Title V permits under the CAA. Moreover, which EGUs would shut down has no relationship to the features of or environmental controls in place at the EGUs themselves. In fact, older, less environmentally-controlled EGUs may continue to operate while newer EGUs with state-of-the-art environmental controls may shut down, simply by virtue of which state the EGU is located in and how much “excess” natural-gas fired, renewable, or other low carbonemitting capacity the state possesses. In other words, factors that have nothing to do with the environmental characteristics of the EGUs themselves would determine which EGUs must shut down to comply with the rule. The shutting down of these EGUs under the Proposed Guidelines would result in a total loss of all economically beneficial use to their owners and thus constitute a total regulatory taking in the event that EPA’s Proposed Guidelines were found otherwise to be lawful. B. The Proposed Guidelines, if Lawful, Would Result in a Partial Regulatory Taking of Any EGU That Must Significantly Curtail its Operation or Remaining Useful Life To Ensure Compliance with the Proposed Guidelines’ Emission Goals. Under Building Blocks 2, 3, and 4 of the Proposed Guidelines, as reflected in the EPA state emission goals, many coal- and oil-fired EGUs that are not shut down entirely would experience serious curtailment of their operations as the Proposed Guidelines result in natural gas-fired generation being dispatched before coal- or oil-fired generation (Building Block 2), 115 increased use of renewable energy that will also be dispatched before coal- or oil-fired generation (Building Block 3), and the implementation of demand-side energy efficiency programs that are expected to result in less demand for electricity (Building Block 4). Although these actions would not rise to the level of a total taking under Lucas, partial takings claims may be brought under Penn Central, which established a set of factors that courts consider to determine whether a regulation results in a compensable partial taking: (1) the character of the government action; (2) the economic impact of the regulation on the claimant; and (3) the extent to which the regulation interferes with distinct investment-backed expectations. Penn Cent., 438 U.S. at 124. The Penn Central factors are “designed to allow ‘careful examination and weighing of all the relevant circumstances.’” Tahoe-Sierra Pres. Council, Inc. v. Tahoe Reg'l Planning Agency, 535 U.S. 302, 322 (2002) (quoting Palazzolo v. Rhode Island, 533 U.S. 606, 636 (2001) (O'Connor, J., concurring)). The factors focus “directly upon the severity of the burden that government imposes upon private property rights.” Lingle, 544 U.S. at 539. As discussed below, analysis of these three factors demonstrates that the Proposed Guidelines, if otherwise lawful, would result in a partial taking for coal- and oil-fired EGUs that would be forced to significantly curtail their operations as a result of the Proposed Guidelines. 1. The Character of the Proposed Guidelines Weighs in Favor of a Takings Claim. Under the Takings Clause of the Fifth Amendment, courts look not to the effectiveness of the regulation or legitimacy of the public purpose, but instead to the magnitude or character of the burden a particular regulation imposes upon private property rights and how that regulatory burden is distributed among property owners. Lingle, 544 U.S. at 541. A regulation causes a taking where it forces “‘some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.’” Tahoe-Sierra, 535 U.S. at 336 (quoting Penn 116 Cent., 438 U.S. at 123-24). Where a regulation has a severe or an “unduly harsh impact upon the owner’s use of the property,” it weighs in favor of a takings claim. Penn Cent., 438 U.S. at 127; see also Lingle, 544 U.S. at 539. Likewise, where a government regulation is of an “extraordinary” character and destroys “‘one of the most essential sticks in the bundle of rights that are commonly characterized as property,’” it weighs heavily toward a takings claim. See, e.g., Hodel v. Irving, 481 U.S. 704, 716 (1987) (quoting Kaiser Aetna, 444 U.S. at 176). For the owner of an EGU, one of the “most essential sticks” in the owner’s bundle of property rights is the right to use the EGU for its lawful, economically beneficial, and only intended use: the generation of electricity. The Proposed Guidelines would require significant reductions in the operations of many coal- and oil-fired EGUs. In particular, reflecting EPA’s Building Block 2, the EPA-required state emission goals would shift generation from coal-fired EGUs, which currently provide baseload generation, to NGCC units. Requiring severe curtailments in the operation of an EGU would prevent the owner of that property from realizing its sole purpose—the generation and sale of electricity. Penn Central is instructive here. In Penn Central, which involved regulations affecting Grand Central Station in New York City, the Supreme Court found that regulations restricting the ability to build above the train terminal in historic Grand Central Station in New York City did not interfere with owner “Penn Central’s primary expectation” because it did “not interfere in any way with the present uses of the Terminal . . . . [A]ppellants may continue to use the property precisely as it has been used for the past 65 years.” 438 U.S. at 136. The Court further determined that the law “not only permit[ted] reasonable beneficial use of the landmark site but also afforded appellants opportunities further to enhance not only the Terminal site proper but also other properties” by permitting Penn Central to: (1) “profit from the Terminal but also to 117 obtain a ‘reasonable return’ on its investment,” (2) expand into the air space as long as the new construction “would harmonize . . . with [the Terminal]”; and (3) to transfer its “pre-existing air rights . . . to at least eight parcels in the vicinity of the Terminal.” Id. at 136-38. By contrast, the Proposed Guidelines would prevent the use of EGUs in accordance with their intended purpose. The severe curtailment of operation of an EGU would prevent it from generating electricity, which is its sole purpose. The Proposed Guidelines would destroy the value of these sources by confiscating their generating capacity and preventing the owners of these EGUs from obtaining a reasonable return on their investment. Indeed, because many of the owners of these EGUs have obligations to provide electricity to their customers, the Proposed Guidelines would not only prevent the reasonable return on their investment but could require certain owners to build capacity to replace the severely curtailed operation of their existing coalfired units or to purchase electricity from other electric generators on the open market to ensure that their customers’ power remains on—even though the owner could provide such electricity to its customers from the affected EGUs if their operation was not restricted as a result of EPA’s rule. In sum, for those EGUs whose operation is severely restricted as a result of the Proposed Guidelines, EPA will have confiscated the “most essential sticks” in their bundle of property rights, assuming the Proposed Guidelines were otherwise lawful. The fact that these measures would be taken for the “public good” of reducing GHGs is of no relevance to the takings question. Owners of certain coal- and oil-fired EGUs should not be forced to bear this societal burden alone. Thus, the extraordinary character and severe impact of the Proposed Guidelines on those EGUs weigh heavily in favor of finding that a taking would occur if EPA were able to lawfully finalize its proposal. 118 2. The Proposed Guidelines Would Cause Severe Economic Impacts. The second factor in the Penn Central analysis focuses on the economic impact of the regulation. 438 U.S. at 124. The inquiry is whether the regulation “impair[s] the value or use of [the] property” according to the owner’s general use of its property. PruneYard Shopping Ctr. v. Robins, 447 U.S. 74, 83 (1980); see also Penn Central, 438 U.S. at 136-38. EPA admits that there would be significant economic impacts associated with the Proposed Guidelines. 79 Fed. Reg. at 34,935-36; EPA, Regulatory Impact Analysis for the Proposed Carbon Pollution Guidelines for Existing Power Plants and Emission Standards for Modified and Reconstructed Power Plants at ES-7 to ES-9 (June 2014), Docket ID No. EPA-HQ-OAR-2013-0602-0391 (“RIA”). For those coal- and oil-fired EGUs that would be forced to curtail their operations substantially, economic impacts would be severe. Those EGUs would no longer be able to run at the capacity at which they were constructed to run. This would result not only in the generation of less electricity—and thus fewer profits—but also higher operating costs and inability to recover investments that were made in anticipation of continued operations. Coal- and oil-fired EGUs are not designed to operate as peaking units, and the costs associated with doing so are significant. Failing to operate these units fully results in their needing to burn more fuel per MWh generated (which perversely results in greater CO2 emissions). This “is not a case in which the Government is exercising its regulatory power in a manner that will cause an insubstantial devaluation of petitioners’ private property,” Kaiser Aetna, 444 U.S. at 180, but is instead a case where the government regulation would be “functionally equivalent” to a direct appropriation, Lingle, 544 U.S. at 539. Both the value and the use of these EGUs as electric generating facilities would be significantly impaired by the Proposed Guidelines, and their owners would lose significant control over the manner of operation of their sources. These 119 major economic impacts weigh heavily in favor of finding a taking assuming the Proposed Guidelines are otherwise lawful. 3. The Proposed Guidelines Would Significantly Interfere With Reasonable Investment-Backed Expectations. The Proposed Guidelines would interfere significantly with the reasonable investmentbacked expectations of the owners of those coal- and oil-fired EGUs whose operation will be significantly restricted. See Penn Cent., 438 U.S. at 124 (establishing as a primary factor in a regulatory takings analysis the extent of interference with “distinct investment-backed expectations”); see also Palazzolo, 533 U.S. at 618 (adopting reasonableness requirement); Ark. Game & Fish Comm’n v. United States, 133 S. Ct. 511, 522 (2012) (adopting reasonableness requirement). Although the Supreme Court has indicated that the extent to which an industry is regulated plays a role in evaluating the reasonableness of investment-backed expectations, see Monsanto, 467 U.S. at 1008-09, this does not mean that prior regulation of an industry per se forecloses the reasonableness of investment-backed expectations, cf. Palazzolo, 533 U.S. at 633 (“Today’s holding does not mean that the timing of the regulation’s enactment relative to the acquisition of title is immaterial to the Penn Central analysis. Indeed, it would be just as much error to expunge this consideration from the takings inquiry as it would be to accord it exclusive significance.”) (O’Connor, J., concurring); see also Tahoe-Sierra, 535 U.S. at 336 (“Accordingly, we have eschewed any set formula . . . . The outcome instead depends largely upon the particular circumstances in that case.”) (internal quotation marks and citations omitted). As the Supreme Court has noted, whether a regulation has gone “‘too far’” is instead “‘a question of degree—and therefore cannot be disposed of by general propositions.’” MacDonald v. Yolo Cnty., 477 U.S. 340, 348 (1986) (quoting Pa. Coal, 260 U.S. at 416). Thus, the appropriate question for a regulated industry like the electric generation industry is the degree to 120 which, when the affected sources were developed and constructed, the EGUs had a reasonable investment-backed expectation that they would be able to operate the source as the type of source it was designed to be and at a certain level of generating capacity in light of the existing regulatory climate, among other factors. Although this determination will require an “essentially ad hoc, factual inquir[y],” Penn Cent., 438 U.S. at 124, regarding how exactly the Proposed Guidelines would affect the EGU in question, the date that the investment decision was made for the EGU, and the laws and regulations existing at that time, the Proposed Guidelines, if valid, would lead to a number of strong takings claims by the owners of the coal- and oil-fired EGUs that are having their operations severely curtailed. The Proposed Guidelines would require drastic reductions in generating capacity for many sources. Despite the fact that EGUs are heavily regulated, their owners had no reason to expect that their operations would be curtailed as a result of section 111(d) of the CAA. First, most, if not all, of the investment decisions for these EGUs were likely made before the Supreme Court decided Massachusetts v. EPA, 549 U.S. 497 (2007), and held that GHGs could be regulated under the CAA. Prior to that time, EPA had not regulated GHGs under the Act and said it had no authority to do so. 68 Fed. Reg. 52,922, 52,925-29 (Sept. 8, 2003). Furthermore, all of the investment decisions were made well before EPA proposed to adopt its novel “building blocks” approach of determining BSER, which it set forth for the first time in the Proposed Guidelines. Nothing in EPA’s prior rulemakings under section 111(d)—or indeed under section 111(b)—would give the owners of EGUs reason to anticipate that section 111(d) would be used to require reduced utilization of a source. In sum, the owners of those coal- and oil-fired EGUs that would have significant restrictions placed on the operation of those EGUs as a result of the Proposed Guidelines had no reason to expect at the time that the 121 investment decisions were made that this could occur. The owners instead constructed these EGUs with the reasonable investment-backed expectation that, although the sources would be subject to some regulation under the CAA and other statutes and regulations, the EGUs would operate for their remaining useful lives and at a certain level of generation and profitability. In sum, the Proposed Guidelines would result in a total taking of those coal- and oil-fired EGUs that are required to shut down in order for the states in which they are located to meet the goals set forth in the Proposed Guidelines. And for those EGUs whose operation would be significantly curtailed as a result of the Proposed Guidelines, after examining the three Penn Central factors, it is clear that a partial taking would result. The impact of the Proposed Guidelines on these EGUs would be severe and that burden would be unfairly distributed by focusing exclusively on coal- and oil-fired EGUs. The Proposed Guidelines would also impose severe economic impacts on the owners of those EGUs and would significantly interfere with those owners’ investment backed expectations. If EPA decides to proceed with the Proposed Guidelines and if those Guidelines could be lawfully promulgated, which for the many reasons presented in these comments they cannot, then EPA would have to justly compensate the owners of those EGUs that are forced to shut down or to restrict their operations significantly. Otherwise, the owners of these EGUs will be forced “‘alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.’” Tahoe-Sierra, 535 U.S. at 336 (quoting Penn Cent., 438 U.S. at 123-24). IX. EPA Usurps the Role of States and Treads Into Areas that Are Quintessentially Within the Purview of State Control in Violation of Section 111(d). A. State Primacy Under Section 111(d) and EPA’s Regulations Implementing Section 111(d) EPA’s authority under section 111(d) is narrowly tailored; Congress intended for section 111(d) regulation to employ the same cooperative federalism framework that exists in section 122 110 of the CAA, giving primacy to the states. See CAA § 111(d)(1) (EPA to establish “procedure similar to that provided by section [110]”). In contrast to the exclusive authority Congress gave to EPA to set new source performance standards, section 111(d) does not give EPA direct regulatory authority over existing sources. Instead, section 111(d) directs EPA to establish a “procedure” for states to submit plans establishing performance standards for existing sources. Section 111(d) gives the states broad discretion to develop such plans and to implement and enforce them based on specific state concerns and needs. EPA recognized this in 1975 when it promulgated the Subpart B regulations to interpret and implement section 111(d): “States will have primary responsibility for developing and enforcing control plans under section 111(d).” 40 Fed. Reg. at 53,343 (emphasis added). EPA’s proposal to establish hard CO2 emission intensity caps contravenes the language of the statute and that intent. 79 Fed. Reg. at 34,953, Proposed 40 C.F.R. § 60.5765; id. at 34,957-58, Proposed Subpart UUUU, Tbl. 1. EPA may require states to submit plans that contain performance standards for emissions of certain pollutants from designated facilities, but it does not have the authority to dictate the form and content of those performance standards. In section 111(d), Congress directs EPA merely to “establish a procedure . . . under which each State shall submit to [EPA] a plan which . . . establishes standards of performance” for existing sources within the state. CAA § 111(d)(1) (emphasis added). EPA established this procedure in 1975 when it promulgated the Subpart B regulations, codified at 40 C.F.R. Part 60, Subpart B, §§ 60.20-60.29. Whereas section 111(d)(1) limits EPA’s authority to setting up procedures, the states are given the authority and responsibility to establish the actual, substantive standards of performance. In contrast, section 111(d)(2)(A) authorizes EPA to set substantive standards of performance, but only in situations where a state has failed to submit an 123 acceptable plan. Congress carefully limited EPA’s authority to establish substantive standards of performance for existing sources to those limited situations where a state fails to act. Thus, unlike the very different language in section 111(b) governing the standards of performance for new sources, section 111(d) gives EPA no direct regulatory authority over existing sources, and instead gives states broad discretion to develop such plans subject to a general requirement that the state’s exercise of discretion be “satisfactory.” CAA § 111(d)(2)(A). The Act provides criteria for establishing whether a state plan is “satisfactory,” including the definition of “standard of performance,” and a direction that states shall be allowed to consider remaining useful life, “among other factors.” Id. §§ 111(a)(1), 111(d)(1), (2). The Act leaves to the states substantial freedom as to the factors to be considered in formulating a state plan and how those factors are to be weighed, and it does not dictate any particular outcome for the state. States thus have significant discretion to adopt state plans that vary from EPA’s emission guidelines, as long as they have considered relevant factors. Congress envisioned the section 111(d) system to be like the state implementation plan system under section 110, under which as long as a state considers relevant factors, EPA may not second-guess the state’s judgment. This conclusion is well-established in case law governing section 110 and arguably applies in this context as well, given the explicit connection Congress made between section 110 and section 111. Courts have noted that states, under section 110, are “given wide discretion in formulating” their plans. Union Elec. Co. v. EPA, 427 U.S. 246, 250 (1976). EPA’s role is “confine[d] . . . to the ministerial function of reviewing [state plans] for consistency with the Act’s requirements” and little more. Luminant Generation Co. v. EPA, 675 F.3d 917, 921 (5th Cir. 2012). The Fifth Circuit has also noted that: The great flexibility accorded the states under the Clean Air Act is further illustrated by the sharply contrasting, narrow role to be played by EPA. Quite 124 simply, the Act provides that the Administrator shall approve the proposed plan if it has been adopted after public notice and hearing and if it meets [the] specified criteria. Fla. Power & Light Co. v. Costle, 650 F.2d 579, 587 (5th Cir. 1981) (internal quotation marks omitted). Similarly, EPA has no authority to question the wisdom of a State’s choices of emission limitations if they are part of a plan which satisfies the standards of § 110(a)(2) . . . . Thus, so long as the ultimate effect of a State’s choice of emission limitations is compliance with the national standards for ambient air, the State is at liberty to adopt whatever mix of emission limitations it deems best suited to its particular situation. Virginia v. EPA, 108 F.3d 1397, 1407-08 (D.C. Cir. 1997) (citing Train v. NRDC, 421 U.S. 60, 79 (1975)). EPA cannot disagree with a state plan merely because it would exercise its judgment differently. The Act is supposed to “leave[] it to the individual States to determine, in the first instance, the particular restrictions that will be imposed on particular emitters within their borders.” EME Homer City Generation, L.P. v. EPA, 696 F.3d 7, 12 (D.C. Cir. 2012). Anything in EPA’s 1975 Subpart B rules that requires EPA “approval” must be read consistently with the CAA and case law indicating that as long as the state considers relevant factors, its plan must be deemed “satisfactory.” In the Proposed Guidelines, however, EPA interprets the statutory term “satisfactory” to mean that it has the authority to impose “binding” emission rates on each state. See, e.g., 79 Fed. Reg. at 34,844, 34,892. This interpretation ignores the structure and context of section 111(d), which gives substantive authority to the states to establish the standards of performance in the first instance, and to consider whatever factors it wishes in applying those standards of performance. EPA instead has proposed to establish hard numeric standards of performance 125 (though it calls them “emission guidelines” 25 or “emission performance level[s]” rather than “standards of performance,” EPA Legal Memorandum at 16-18, 29), and would apply them to states rather than to the sources. 26 Nothing in either the CAA or the Subpart B regulations empowers EPA to establish binding emissions performance obligations for states. Standards of performance are to be set by states and must apply to sources. EPA’s statewide performance goals are emission limitations not authorized by the Act or the Subpart B regulations. EPA may not lawfully enlarge its authority in this way. Under the CAA and the Subpart B regulations, EPA’s responsibility under section 111(d) is to specify the BSER that limits the emissions of the pollutant in question from the source at hand. As discussed at length in Section III of these comments, EPA’s establishment of BSER here is unlawful because it requires measures that occur beyond the source itself and over which the source has no control. But even EPA’s application of the one Building Block that arguably falls within the scope of section 111 (Building Block 1) is flawed. Building Block 1 calls for heat rate improvements at coal-fired EGUs. Under the statute, EPA’s determination of BSER for those units does not involve giving each state a rate that each unit must meet (either individually or collectively). Rather, EPA’s determination of BSER in this instance might involve EPA identifying for consideration by the states a list of measures that could be made at a coal-fired 25 EPA’s Subpart B regulations define “emission guidelines” as emission limitations that apply to “designated facilities,” not to the states in which those designated facilities are located. 40 C.F.R. § 60.21(e). 26 As described in Section V of these comments, EPA may establish emission guidelines only for the same type of sources the Agency has regulated under section 111(b). CAA § 111(d)(1)(A); Am. Elec. Power Co., 131 S. Ct. at 2537; see also 40 C.F.R. § 60.22(b)(5) (“The Administrator will specify different emission guidelines or compliance times or both for different sizes, types, and classes of designated facilities . . . .”) (emphasis added). The Court also confirmed that it is the states that “issue performance standards for stationary sources within their jurisdiction,” not EPA that does so. Am. Elec. Power, 131 S. Ct. at 2537-38. 126 EGU to improve heat rate that EPA has determined to be BSER. States would then apply that BSER to each unit within the state to set the actual standard of performance for the EGU. And as the Act and the Subpart B regulations make clear, states could deviate from EPA’s BSER determination based on things such as remaining useful life of the EGU or other factors such as the inability to do a certain improvement at a unit because, for instance, the cost would be excessive, the measure would be impossible to implement at that unit, or the measure had been done previously. This is how the partnership between EPA and the states should work under section 111(d). The Proposed Guidelines utterly fail in this regard. If EPA finalizes the Proposed Guidelines, it will violate the Act, which requires section 111(d) performance standards to be state-led and promulgated, and to apply to existing fossil fuel-fired EGUs, the same sources to which EPA’s section 111(b) standards apply. EPA has not given states the flexibility to set state performance standards—authority to which states are entitled under the statute. Moreover, EPA would force states to violate the terms of the Act by regulating sources that may not be regulated under a section 111(d) rule for existing fossil fuelfired EGUs. See, e.g., 79 Fed. Reg. at 34,853 (“EPA is proposing that states be authorized to submit state plans that do not impose legal responsibility on the affected EGUs for the entirety of the emission performance level, but instead, by adopting what this preamble refers to as a ‘portfolio approach,’ impose requirements on other affected entities (e.g., renewable energy and demand-side energy efficiency measures) that would reduce CO2 emissions from the affected EGUs.”). Although EPA asserts that it is giving states flexibility to adopt in their state plans whatever “building blocks” they see fit, id. at 34,859, in reality, because the standard of performance is an emission intensity cap applied to the state, rather than to existing facilities, 127 states in most cases must adopt emission limitations on facilities apart from fossil fuel-fired EGUs in order to meet the EPA-mandated, binding caps. EPA’s Subpart B rules generally require state emission standards to “be no less stringent than” the emission guidelines set forth by EPA. 40 C.F.R. § 60.24(c). When it first adopted its section 111(d) regulations, however, EPA took pains to emphasize that, in contrast to the “emission standards” that states adopt under their section 111(d) plans, emission guidelines are not “legally enforceable.” 40 Fed. Reg. at 53,341 (“to emphasize that a legally enforceable standard is not intended, the term ‘emission limitation’ has been replaced with the term ‘emission guideline’”); see also 40 C.F.R. § 60.21(f) (state-imposed “[e]mission standard means a legally enforceable regulation . . . .”). Thus, EPA cannot in the first instance prescribe to states a standard of performance, which it would do in the Proposed Guidelines. EPA stated in 1975 that emission guidelines “will not be requirements enforceable against any source. Like the national ambient air quality standards prescribed under section 109 and the items set forth in section 110(a)(2)(A)-(H), they will only be criteria for judging the adequacy of State plans.” 40 Fed. Reg. at 53,343. Thus, the emission guidelines and BSER determinations that EPA publishes are supposed to be largely procedural and contain only nonbinding factors and descriptions of demonstrated systems for states to consider as the states set their own standards. They are not, as EPA has proposed, supposed to be “binding” emission rate limitations on states. 79 Fed. Reg. at 34,897 (“the state-specific goals will be binding emission guidelines”). What the Administrator determines to be BSER is merely one of the many factors states must consider in determining the level and form of any existing source performance standard as applied to a 128 specific EGU. 27 EPA’s Section 111(d) guideline document is supposed to contain an “emission guideline” for designated facilities. 40 C.F.R. § 60.22(b)(5). EPA’s Proposed Guidelines go beyond that mandate to propose binding emission limitations on entities that neither the Act nor the Subpart B regulations authorize EPA to regulate directly. Because states are not and cannot be regulated under section 111(b), emission limitations cannot be imposed on them under section 111(d). Similarly, under the Subpart B rules, to which EPA claims to be adhering, 79 Fed. Reg. at 34,951, Proposed 40 C.F.R. § 60.5700 (“These emission guidelines are developed in accordance with section[] 111(d) of the Clean Air Act and subpart B of this part.”), states have considerable flexibility to deviate from EPA’s emission guidelines in adopting plans and emission standards. 28 27 As discussed above in Section III, the BSER is supposed to be reflected in stateestablished standards of performance that apply to “sources.” In 1975, when it promulgated the Subpart B regulations, EPA interpreted section 111(d) being source-specific and technologybased. EPA stated that BSER is something that has been “adequately demonstrated for designated facilities.” 40 Fed. Reg. at 53,340 (emphasis added). And it stated that it believed Congress “intended a technology-based approach” under section 111(d). Id. at 53,343. It emphasized that it viewed this interpretation as “legally correct in view of the language, statutory context, and legislative history of the provision.” Id. But EPA has turned this long-established construct on its head by proposing to reflect the BSER in binding goals that apply to states. See, e.g., RIA at ES-2 (“[T]his proposed rule contains state-specific goals that reflect the EPA’s calculation of the emission reductions that a state can achieve through the application of BSER.”); 79 Fed. Reg. at 34,892 (“[T]he interim and final [state] goals will be binding emission guidelines for state plans.”). 28 To the extent EPA is acting contrary to its own regulations, it is not entitled to deference in doing so. Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512 (1994) (deference to agency’s interpretation of its own regulation does not confer when it is “plainly erroneous or inconsistent with the regulation” or “alternative reading is compelled by the regulation’s plain language or by other indications of the [Administrator’s] intent at the time of the regulation’s promulgation”) (internal quotations omitted); see also INS v. Cardoza-Fonseca, 480 U.S. 421, 446 n.30 (1987) (“An agency interpretation of a relevant provision which conflicts with the agency’s earlier interpretation is ‘entitled to considerably less deference’ than a consistently held agency view.”) (quoting Watt v. Alaska, 451 U.S. 259, 273 (1981)); Good Samaritan Hosp. v. Shalala, 508 U.S. 402, 417 (1993) (“[T]he consistency of an agency’s position is a factor in assessing the weight that position is due.”). To the extent it is interpreting its Subpart B 129 For example, states may apply “less stringent emission standards or longer compliance schedules” to particular facilities or classes of facilities if the costs of adopting the standards suggested by the emission guidelines would be unreasonably costly, physically impossible, or for other reasons. 40 C.F.R. § 60.24(f)(1)-(3). States have significant discretion in designing their own plans and determining how individual sources or classes of sources may demonstrate compliance, including taking into account the “remaining useful life of the existing source.” CAA § 111(d)(1). Applying these criteria, states may grant individual sources or types of sources longer periods of time to comply, or may apply less stringent standards than set forth in EPA’s emission guidelines. EPA has also stated that “[s]tates will be free to vary from the levels of control represented by the emission guidelines . . . . In most if not all cases, the result is likely to be substantial variation in the degree of control required for particular sources, rather than identical standards for all sources.” 40 Fed. Reg. at 53,343. Thus, states have flexibility to implement EPA guidelines as they deem “reasonable” where costs are determined to be “unreasonable,” where they would be “physically impossible,” or otherwise “unreasonable.” EPA also made clear that “[s]tates will be free to set more lenient standards [than those set out in EPA’s emission guidelines], subject to EPA review . . . in cases of economic hardship.” Id. Moreover, in accordance with its Subpart B rules, EPA must address in its guideline document sub-categories of “different sizes, types and classes” of existing sources where factors like “costs of control, physical limitations [or] geographical location” warrant the application of different guidelines. 40 C.F.R. § 60.22(b)(5). EPA has failed to do so here. The Agency regulations in a way that conflicts with section 111(d), that interpretation is unlawful. A regulation—or interpretation thereof—that is “manifestly contrary to the statute” is not only unlawful, but is entitled to no deference. United States v. Mead Corp., 533 U.S. 218, 227 (2001); see also Chevron, 467 U.S. at 844; United States v. Morton, 467 U.S. 822, 834 (1984); 5 U.S.C. § 706(2)(A), (D). 130 underscored as early as 1975 that “emission guidelines will reflect subcategorization within source categories where appropriate,” and the guidelines “will in effect be tailored to what is reasonably achievable by particular classes of existing sources . . . .” 40 Fed. Reg. at 53,343. Because a standard of performance must be “adequately demonstrated” for each source, EPA has an obligation to establish highly subcategorized emission guidelines within a broad source category like existing fossil fuel-fired EGUs. B. The “Flexibility” EPA Claims Exists in the Proposed Guidelines Is Illusory. EPA has turned the flexibility embedded in section 111(d) on its head. In the Proposed Guidelines, EPA would set binding state emission reduction targets that cannot be adjusted by states once promulgated. This is the antithesis of state flexibility. EPA’s aggressive assumptions in the Building Blocks further erode state flexibility and disregard the primacy that Congress explicitly gave to the states over EPA under section 111(d). The Proposed Guidelines, while frequently vague and incomplete, provide information on a broad range of issues related to state implementation. Generally, the Proposed Guidelines provide EPA’s perspective on a particular detail of implementation and request comment on various, often more complex issues that do not appear to have been fully developed by the Agency. UARG’s comments address many of these issues below. As a preliminary matter, however, the Proposed Guidelines’ discussion of state implementation highlights two key flaws in the proposal: (1) the Proposed Guidelines’ false claims of flexibility; and (2) their intrusion into areas of exclusive state jurisdiction. EPA’s Proposed Guidelines tout the “flexibility” the guidelines would provide to states. In reality, however, the proposed emission guidelines would dictate all key policy choices, including those that the CAA and EPA’s Subpart B regulations expressly reserve to the states. The states, on the other hand, would be tasked with addressing issues over which EPA itself has 131 no authority and answering the unprecedented questions the Proposed Guidelines raise where EPA has apparently been unable to devise solutions. The Proposed Guidelines lack flexibility because the state goals that are its centerpiece are set assuming significant emission reductions resulting from use of all four Building Blocks. EPA claims that states are free to disregard the policies represented by the Building Blocks when devising the elements of their plans. EPA states: We also note that a state is not required to achieve the same level of emission reductions with respect to any one building block as assumed in the EPA’s BSER analysis. If a state prefers not to attempt to achieve the level of performance estimated by the EPA for a particular building block, it can compensate through over-achievement in another one, or employ other compliance approaches not factored into the state-specific goal at all. The EPA has estimated reasonable rather than maximum possible implementation levels for each building block in order to establish overall state goals that are achievable/while allowing states to take advantage of the flexibility to pursue some building blocks more aggressively, and others less aggressively, than is reflected in the goal computations, according to each state’s needs and preferences. 79 Fed. Reg. at 34,926. Under EPA’s reasoning, states could in fact implement none of the Building Block policies and instead create an entirely different program to control CO2 emissions. This is a fantasy. EPA claims that its “reasonable rather than maximum” approach to implementing the Building Blocks means that there is room for states to trade emission reductions among Building Block policies (and policies that were not included in the Building Blocks) to achieve each state’s goal. But EPA’s assumptions in modeling the Building Blocks were anything but reasonable. As explained in Section XIV, EPA has in fact made entirely unrealistic assumptions. The results of those analyses are inconsistent with the conclusions reached by the Arizona Department of Environmental Quality (“ADEQ”), NERC, SPP, the Kansas Corporation Commission (“KCC”), and other independent reviewers. For example, in initial comments filed with EPA, ADEQ presented an analysis supporting its finding that “Arizona has no flexibility to 132 shift from one Building Block to another to meet its rate-based goal under the program as proposed.” Email from Steve Burr, Legal Support Section, Air Quality Div., ADEQ, to A-and R-Docket et al. at 1 (Aug. 22, 2014), Docket ID No. EPA-HQ-OAR-2013-0602-14064. Other states will face the same situation, as confirmed by UARG’s assessment. Given these facts, there is no way for a state to meet its goal without adopting all four Building Blocks and implementing them largely in line with EPA’s assumptions. Indeed, even more may actually be required for states to meet EPA’s goals. EPA’s approach to assessing each Building Block assumed that nothing would go wrong in the implementation of the Building Block policies. Accordingly, EPA’s state goals leave no room to accommodate unexpected events that could disrupt generation from sources needed to ensure compliance with state goals. Yet disruption in generation from renewable sources and NGCC units will be unavoidable. Because the state goals ignore these realities, EPA cannot claim that its state goals reflect reasonable rather than maximum implementation of the Building Blocks. As a result, there is no true flexibility in the Proposed Guidelines. Another reason the Proposed Guidelines would limit rather than enhance flexibility for states is that it would encroach on areas that fall under the exclusive jurisdiction of the states. Most—if not all—of the programs in Building Blocks 2, 3, and 4 are the quintessential purview of state electricity regulators—not state environmental agencies. These programs have been developed pursuant to well-established state sovereign powers over matters relating to electricity regulation, including determining the appropriate mix of generating resources within a state. EPA is barred by U.S. Supreme Court precedent from infringing upon a traditional state sovereign function unless Congress has adopted clear statutory language expressly authorizing the Agency to do so. Rice v. Santa Fe Elevator Corp., 331 U.S. 218, 230 (1947) (courts begin 133 preemption analysis “with the assumption that the historic police powers of the States were not to be superseded by [a federal law] unless that was the clear and manifest purpose of Congress”). Nothing in the CAA expressly authorizes EPA to regulate the generation of electricity or other such energy regulatory matters traditionally reserved to states. Indeed, as discussed in Section VII, the language of section 111(d) suggests the precise opposite. Further, as described in Section IV, given EPA’s regulation of EGUs under MATS, EPA lacks authority to regulate EGUs under section 111(d) at all. The Proposed Guidelines acknowledge that interference with the sphere of exclusive state jurisdiction is a considerable concern: [I]ncluding [renewable energy (“RE”)] and demand-side [energy efficiency (“EE”)] measures in state plans would render those measures federally enforceable and thereby extend federal presence into areas that, to date, largely have been the exclusive preserve of the state and, in particular, state public utility commissions and the electric utility companies they regulate.” 79 Fed. Reg. at 34,902. EPA seeks comment on a proposal to remedy this situation by essentially creating a technical loophole. EPA suggests that it can avoid entanglement in areas over which the states are sovereign simply by not including things such as RE and EE in state plans, preventing such requirements from becoming federally enforceable, while states impose the requirements directly to allow compliance with what would be otherwise unachievable emission limits contained in state plans. Id. EPA does not address how elements necessary to ensure that state plans are implemented to achieve EPA’s goals can be left out of those plans. EPA’s proposed approach, moreover, would not prevent federal interference in these areas of traditional state primacy. It would require that states take action outside section 111(d) plans themselves but would nevertheless improperly compel state action. 134 EPA also proposes what it calls the “‘state commitment approach.’” Id. The problems with that approach are even more glaring. Again, this approach would not make requirements for non-EGUs part of a federally enforceable state plan: “Instead, the state plan would include an enforceable commitment by the state itself to implement state-enforceable (but not federally enforceable) measures that would achieve a specified portion of the required emission performance level on behalf of affected EGUs.” Id. This approach plainly does not avoid EPA interference in areas that are traditionally regulated only by the states. It merely seeks to avoid the appearance of such interference. Further, it fails to address how measures that are necessary to ensure achievement of a requirement under section 111(d) can be left out of a federally enforceable plan. C. The Proposed Guidelines Conflict With State Laws and Fail To Take Account of Other CAA Requirements. The Proposed Guidelines actually conflict with key state laws, including state constitutions. To override these laws, the CAA would have to provide clear authority, grounded in the Commerce Clause, for EPA to take such action. The Act provides no such authority. State laws that would interfere with the implementation of EPA’s Proposed Guidelines are widespread. For instance, a September 9, 2014 letter from several governors identifies existing laws in Kansas, Kentucky, Louisiana, Missouri, and West Virginia that prohibit those states from regulating power plant emissions “by shifting pollution-control costs to other parts of the economy.” Letter from Governors of Fifteen States to President Barack Obama at 1 n.3 (Sept. 9, 2014), available at http://governor.alabama.gov/ assets/2014/09/RGA-Letter-to-POTUS.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). In those states “[e]missions reductions must occur at the power plant source.” Id. 135 Similarly, legislation is under consideration or has been adopted that conflicts with EPA’s Proposed Guidelines. As reported in a June 9, 2014 article, at least nineteen states had introduced resolutions targeting the Proposed Guidelines, eleven states had introduced legislation, and six states had enacted bills as of that time. Nathanael Massey, Some states push back against proposed EPA carbon rule, CLIMATEWIRE, June 9, 2009, available at http://www.eenews.net/climatewire/stories/1060000921/search?keyword=Some+states+push+ back+against+proposed+EPA+carbon+rule (subscription required). An August 2014 report identifies conflicting resolutions and legislation in 18 states. Raymond L. Gifford, et al., EPA’s CO2 Rule and 18 States’ Resolutions and Legislation: EPA’s Proposed CO2 Rule Collides with Flexibility Asserted By States at 3 (Aug. 2014), available at http://www.wbklaw.com/News/ Articles__Publications/EPAs_CO2_Rule_and_18_States_Resolutions_and_Legislation__August _15_2014. Although the content of these laws and legislative proposals differ, many require state agencies to employ the full measure of state flexibility contemplated by the CAA, in contrast to the lack of discretion to implement section 111(d) requirements that EPA would dictate under its Proposed Guidelines. A law enacted by the Wyoming legislature, for instance, provides that EPA’s emission guidelines must recognize state primacy in establishing carbon regulations. J.R. 6, 61st Leg., Gen. Sess., 2011 Wyo. Sess. Laws 559. Kentucky has also adopted legislation at odds with EPA’s Proposed Guidelines. Signed by the governor on April 2, 2014, H.B. 388 limits BSER to “measures undertaken at each coal-fired electric generating unit” and precludes such measures from including fuel-switching, co-firing with other fuels, or limiting utilization. H.B. 388, 2014 Leg., Reg. Sess., 2014 Ky. Acts 18. Other states have longstanding laws, not enacted in reaction to the Proposed Guidelines, that will prevent the program EPA has proposed from being implemented. Arizona, for instance, enacted H.B. 2442 136 in 2010. ARIZ. REV. STAT. § 49-191. That statute prohibits any state agency from adopting or enforcing a state or regional program to regulate GHG emissions for the purposes of addressing changes in atmospheric temperature without express legislative authorization. EPA has no authority to override these state laws due to the Supremacy Clause of the U.S. Constitution. U.S. CONST. art. VI, cl. 2. Wherever possible courts avoid determining that federal preemption of state law has occurred. N.Y. State Conference of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., 514 U.S. 645, 654-55 (1995); Maryland v. Louisiana, 451 U.S. 725, 746 (1981); Napier v. Atl. Coast Line R. Co., 272 U.S. 605, 611 (1926). Furthermore, when a court examines whether preemption occurs, it looks first to the intent of Congress in passing the statute. See, e.g., Gade v. Nat’l Solid Wastes Mgmt. Ass’n, 505 U.S. 88, 96 (1992). Here, EPA is acting pursuant to its authority under the Clean Air Act, which is a statute that regulates air emissions. But EPA’s proposed action here would not directly regulate air emissions from municipal utilities and instead would regulate the amount and source of energy generation from utilities. Similarly, courts typically presume that “the historic police powers of the States were not to be superseded by [federal law] unless that was the clear and manifest purpose of Congress.” Rice, 331 U.S. at 230; see also Wisconsin Public Intervenor v. Mortier, 501 U.S. 597, 605 (1991). As discussed in Section VII above, Congress intended for states to retain authority over energy generation and retail sales and for FERC to retain authority over interstate transmission of electricity. EPA cannot reasonably assert that Congress intended to grant EPA preemption authority over state constitutions and laws through the CAA to regulate energy production and transmission of municipal utilities. In addition, little thought was given to how the Proposed Guidelines interact with other CAA rules. For example, Arizona will have to close all coal units to achieve its interim and final 137 targets under the Proposed Guidelines, but utilities in the state recently invested $500 million to comply with the Regional Haze Program. Furthermore, the Proposed Guidelines will require the ramping up of units located in nonattainment areas, which would violate NAAQS requirements. Section 111(d) explicitly requires that consideration of the remaining useful life of sources be considered when designing performance standards for existing sources. Contrary to that requirement, the Proposed Guidelines will lead to the stranding of assets if finalized, especially given its interaction with other CAA rules. D. Selecting a Rate-Based or Mass-Based Goal According to EPA, states decide the “emission performance level” to include in their plans, 79 Fed. Reg. at 34,900, but under the Proposed Guidelines, states are empowered only to select a rate-based or mass-based emission goal. EPA presents this choice as providing considerable “flexibility” for states. See id. at 34,897 (“[W]e are also proposing to provide states with the flexibility to translate the rate-based goals to mass-based goals.”); see also id. at 34,90809. It does not. EPA is merely affording states a choice between accepting an EPA-prescribed and inflexible rate-based state goal or translating that rate-based goal into an equivalent massbased goal. Id. at 34,892 (translated mass-based goal must “achieve[] the same degree of emission limitation” as EPA’s rate-based goal). Under the Proposed Guidelines, states are invited to attempt to calculate a mass-based rate. Id. at 34,897. The preamble to the Proposed Guidelines states that: “[i]f the plan adopts a mass-based goal, the plan must include a description of the analytic process, tools, methods, and assumptions used to translate from the rate-based goal to the mass-based goal.” Id. at 34,911. Similarly, section III of EPA’s “Projecting EGU CO2 Emission Performance in State Plans” TSD (June 2014), Docket ID No. EPA-HQ-OAR-2013-0602-0462 (“Projections TSD”), which was released with the Proposed Guidelines, purports to provide guidance to states on how to make 138 such a goal translation. That guidance, however, does little more than direct states to project the CO2 emissions that would result from application of EPA’s rate-based limit. These statements suggest that states have broad leeway to calculate mass-based goals. Proposed 40 C.F.R. § 60.5770 includes an additional requirement, however: The conversion must represent the tons of CO2 emissions that are projected to be emitted by affected EGUs, in the absence of emission standards contained in the plan, if the affected EGUs were to perform at an average lb CO2/MWh rate equal to the rate-based goal for the state identified in Table 1 of [the Proposed Guidelines]. 79 Fed. Reg. at 34,953, Proposed 40 C.F.R. § 60.5770. Although this roughly sketched methodology might produce rate-to-mass goal conversions that are equivalent, other approaches might prove better fits for some states, and EPA should provide states with flexibility to demonstrate that differing methods are appropriate, as the Agency suggests in the preamble language. That other state approaches might be valid is seemingly confirmed by EPA’s November 2014 TSD, entitled “Translation of the Clean Power Plan Emission Rate-Based CO2 Goals to Mass-Based Equivalents, Docket ID No. EPA-HQ-OAR-2013-0602-22187 (“Translation TSD”). Unlike the Proposed Guidelines, which require a protection of emissions “in the absence of emission standards contained in the plan,” 79 Fed. Reg. at 34,953, Proposed 40 C.F.R. § 60.5770, i.e., in the absence of application of the Building Block, the Translation TSD directs states to adjust their 2012 emissions to “reflect[] the deployment of incremental building block three and four resources.” Translation TSD at 5. The approaches contained in the proposed regulatory language of the Proposed Guidelines and the Translation TSD cannot be reconciled. Yet, EPA appears to believe that either approach could lead to accurate and satisfactory massbased goals. EPA should therefore adopt the approach to mass-based goal conversion suggested 139 by the language in the preamble to the Proposed Guidelines: EPA should recognize state discretion to adopt different approaches to goal conversion provided states provide a reasonable explanation for the choices they have made. UARG also notes that the Translation TSD includes an appendix presenting mass-based goals for each state calculated in accordance with the Translation TSD methodology. Id. at 9-16. UARG takes no position on the accuracy of the mass-based goals EPA has calculated. If accurate, however, EPA’s mass-based goals confirm the exceptional stringency of the Proposed Guidelines. The results of EPA’s conversions to mass-based standards and the impact they will have should demonstrate to the Agency how extreme its Proposed Guidelines are. Moreover, EPA should state expressly that the mass-based goals it has identified are not binding on states and do not establish a baseline or standard of reasonableness against which state calculations will be judged. E. Types of State Plans The Proposed Guidelines contemplate that states can develop “different types of state plans” to meet EPA’s CO2 emission goals. 79 Fed. Reg. at 34,901. As discussed in the Proposed Guidelines and EPA’s “State Plan Considerations” TSD (June 2014), Docket ID No. EPA-HQOAR-2013-0602-0463 (“Plan TSD”), EPA has identified four plan types: • Rate-based CO2 emission limits applied to affected EGUs; • Mass-based CO2 emission limits applied to affected EGUs; • A state-driven portfolio approach; and • A utility-driven portfolio approach. Plan TSD at 5; see also 79 Fed. Reg. at 34,901-02. The primary distinction between the EGUdriven and portfolio approach-driven plans is that the portfolio approach includes measures, such 140 as for RE and EE, that are legally enforceable against entities other than EGUs, including state agencies and nonprofit entities. As an initial matter, and as EPA acknowledges, “section 111(d) gives states the primary responsibility for designing their own state plans for submission to the EPA.” 79 Fed. Reg. at 34,901. States, accordingly, have broad discretion as to how they structure their plans. EPA, therefore, cannot disapprove a state plan that does not conform to the plan “types” it has identified. In addition, although the Proposed Guidelines and the Plan TSD note some salient features of each plan type, neither provides a thorough assessment of those plan types or the questions and potential problems posed by their new and untested characteristics. For instance, EPA states that for plans that place all legal responsibility for compliance on EGUs, states can nevertheless rely on RE and EE to ensure that EGUs can meet their obligations. Id. EPA does not, however, explain how the availability of adequate RE and EE is to be ensured under such an approach, leaving states unsure of what it takes to obtain plan approval and leaving EGUs vulnerable to noncompliance. EPA suggests that states could force EGUs to invest in EE and RE, id. at 34,902, but does not grapple with whether or how states could impose such requirements or whether adequate RE and EE would be feasible to construct. EPA’s discussion of portfolio approaches is similarly ill-conceived. In the Plan TSD, EPA suggests that state agencies, public or private nonprofits, or electric distribution utilities could be assigned section 111(d)-related responsibilities. Plan TSD at 10. EPA does not, however, explain how any of this might be accomplished legally or practically, except to suggest that state legislatures might have to enact new laws. Id. at 14-18. EPA also suggests that the integrated resource planning (“IRP”) process might be a vehicle for imposing portfolio approach- 141 type obligations, but does not explain what states that do not, or cannot, use the IRP process would be expected to do. Id. at 10-12. The inadequacy of EPA’s Proposed Guidelines is illustrated by the fundamental unanswered questions it raises with respect to current state programs, like the Regional Greenhouse Gas Initiative (“RGGI”), that EPA suggests could be a means by which states could comply with the Proposed Guidelines. See 79 Fed. Reg. at 34,901. RGGI is a mass-based capand-trade program that applies to CO2 emissions from fossil fuel-fired EGUs and is carried out among nine northeastern states. The member states voluntarily undertook the RGGI program and designed its primary features through interstate working groups. Each state implements RGGI pursuant to state-specific legislation and regulations. Despite the fact that RGGI has been in operation since 2008 and that its features are well known, EPA makes only vague assertions that RGGI might be a mechanism for compliance with section 111(d) requirements. This leaves many key questions unanswered and is not an approach that will encourage states to undertake the difficult process of negotiating multi-state plans and devising cap-and-trade programs. For instance, the state-by-state emission budgets for the RGGI states are known, as is the regional cap from the program. EPA has not explained, however, if the regional goal for the RGGI program is consistent with the Proposed Guidelines, how it might need to be altered if it is not consistent, or how states can ensure that other regional programs or additions to the RGGI program can achieve or maintain consistency with EPA’s proposed goals. The Proposed Guidelines also do not address a key component of the RGGI program: its investment of the proceeds from allowance auctions in energy efficiency and other measures to reduce demand as a mechanism for reducing CO2 emissions. Would a program that simply includes investment requirements satisfy EPA’s requirements for EE and other demand 142 reduction-related emissions reductions? Or would something more be required to make such emission reductions binding? Further, in a number of instances, RGGI member states have diverted funds intended for investment in demand reduction measures to fill budget shortfalls. If a state legislature were to take similar action in the future, would the state be deemed to have violated its plan? Would EPA find a sole state in violation, or would that violation be attributed to the region? The Proposed Guidelines note that a number of these questions exist, but until EPA attempts to provide an answer to them, it cannot expect states to implement these policies, and it cannot claim to have provided adequate notice through its proposed rulemaking. EPA also asks for comment on whether state plans can legally include measures such as RE and EE under a broad interpretation of “‘standards of performance for [affected sources]’” or whether they could be included as “‘implement[ing]’ measures.” Id. at 34,903. Although states may have the authority to enact laws or trading programs to impose RE and EE investment requirements on EGUs or to impose portfolio approach type measures pursuant to their own authorities, EPA fails to engage the question of whether it can require (or effectively require) states to include such measures in section 111(d) plans as a result of establishing emission goals that are so stringent as to be unachievable in the absence of such requirements. As discussed in Section XIII, not only are the goals EPA has proposed unlawful, but it would be unlawful for EPA to require the states to develop plans that are beyond the scope of section 111(d) and that command states to adopt new statutory and regulatory programs. F. Timing for Plan Implementation and Achievement of State Emission Performance Goals The Proposed Guidelines would create final state-specific goals to be met by 2030 and interim goals to be achieved on average over the 10-year period from 2020 to 2029. Id. at 143 34,904. EPA claims that this approach provides states with flexibility to design their programs and the trajectory of emission reductions. Again, this flexibility is an illusion. EPA itself acknowledges that states will have between 1.5 and 3.5 years from the expected deadlines for plan submittal before compliance actions must commence. Id. at 34,905. EPA insists this is not unreasonable because affected sources “will have knowledge of state requirements as they are adopted” and will therefore have more time to act. Id. The additional time EPA imagines is itself quite short, and it is unrealistic to expect sources to take costly actions before states finalize their plans and before EPA has reviewed and approved them, especially considering the uncharted territory the Proposed Guidelines force states to enter. Moreover, the emission goals EPA has set are too stringent and require such draconian actions to ensure compliance that many states will have no choice but to take significant action by 2020. For example, the State of Arkansas has concluded that the infrastructure that will be required to implement the Proposed Guidelines makes plan implementation “as early as 2017 to meet an interim goal in 2020 . . . far too soon and too abrupt to allow Arkansas utilities to meet the challenges of the proposed rule in a manner that is cost-effective for their ratepayers.” Letter from Dustin McDaniel, Attorney General of the State of Arkansas to Avi S. Garbow, General Counsel, EPA at 2 (Aug. 4, 2014), available at https://static.ark.org/eeuploads/ag/EPA_letter.pdf (“Arkansas Letter”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). For that reason, Arkansas has requested that EPA eliminate the interim goal to allow for a less burdensome glidepath to compliance. Id. Indeed, it is nearly inconceivable that states could develop renewable energy, build out infrastructure, or even develop more traditional sources of generation to replace the sources that the Proposed Guidelines will force into retirement in time to meet the interim goals. The Congressional Research Service has even found that EPA’s targets are “front 144 loaded,” “with a disproportionate percentage of emission rate reductions required in the early years of the program (2020-2024).” Jonathan L. Ramseur, Cong. Research Serv., “EPA’s Clean Power Plan Proposal: Are the Emission Rate Targets Front-Loaded?” at 1, CRS Insights IN10172 (Nov. 3, 2014), available at http://fas.org/sgp/crs/misc/IN10172.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). CRS finds, for example, that “90% of Arizona’s total emission rate reductions occur between its 2012 baseline and its 2020 emission rate.” Id. at 2. The Proposed Guidelines improperly tie states’ hands and will wreak havoc on electricity markets in the near-term—EPA’s assertions to the contrary notwithstanding. See Initial NERC Report at 17; SPP Comments at Attachment 1, slides 13-14. In the event that EPA decides to proceed with the Proposed Guidelines despite the myriad legal, technical, practical, and political issues that they present, EPA should, at the very least, eliminate the interim goals. EPA seems to suggest that it is at least considering this option in the NODA issued on October 30, 2014. See Section XV. The Proposed Guidelines discuss a number of other issues related to timing of implementation measures. First, EPA proposes to require states to include achievement “demonstrations” for the interim and final goals in plan submittals. 79 Fed. Reg. at 34,904. The Plan TSD indicates that EPA believes such demonstrations will entail considerable efforts, including the use of utility-scale capacity expansion and dispatch planning models. Plan TSD at 28-31. The level of effort EPA apparently envisions requiring states to undertake is more extensive than can be justified under section 111(d)’s permissive “satisfactory” standard for state plans and likely to require far more work than can be completed in the time EPA proposes to allow for states to develop their plans. Indeed, states have three years to develop state implementation plans under section 110 of the CAA to implement the NAAQS, which is the 145 longest period of time EPA proposes to allow from the development of section 111(d) state plans. Yet states have no obligation to conduct this sort of modeling during the development of a state implementation plan. It is unreasonable for EPA to expect states to develop novel plans to implement a program that will transform the electric industry and the economy all while performing burdensome analysis that is not typically required to comply with environmental regulation. In addition, EPA proposes that state plans must be either “self-correcting” or contain milestones, periodic checks, and corrective measures. 79 Fed. Reg. at 34,906-07. The Proposed Guidelines do not provide in-depth discussion of what EPA is considering, apart from the suggestion that corrective measures that are included in a state plan should be triggered by a ten percent deviation from projected performance, and corrective measures that would be included in a plan amendment to be submitted after such a deviation should be triggered by an eight percent deviation from projected performance. EPA is free to make these policy suggestions to the states, but nothing in section 111(d) authorizes the Agency to micromanage these policy choices. EPA also requests comment on whether the consequences for failing to meet a state goal should include a requirement “to achieve additional emission reductions to offset any emission performance deficiency that occurred during a performance period for the interim or final goal.” Id. at 34,908. Attempting to impose emission reduction requirements more stringent than the application of BSER is inconsistent with the CAA. And again, the consequences of a failure to meet a state goal are policy decisions left to the states. EPA also asks for comment on whether it should promulgate “a mechanism under CAA section 111(d) similar to the SIP call mechanism in CAA section 110” in order to redress the failure of any state plan to achieve its emission goal. Id. Section 110 expressly authorizes EPA 146 to initiate a SIP call when the Administrator finds that a SIP is “substantially inadequate.” CAA § 110(k)(5). There is no similar authority for section 111(d) plans. Finally, EPA requests comment on its alternative state goals and compliance timing requirements. 79 Fed. Reg. at 34,909. Each of the flaws identified with respect to the primary proposal apply to the alternative. G. State Plan Approvability Criteria and Components of Approvable Plans The Proposed Guidelines include “four general plan approvability criteria, [and] twelve required components for a state plan to be approvable.” Id. at 34,900. EPA notes that the approvability criteria and required components of a section 111(d) plan can differ from those required for a SIP under section 110. Id. at 34,909. EPA also states that compliance with the four criteria and the twelve components is required for a plan to meet the 111(d) “satisfactory” standard. Id. (“The EPA proposes to use the combination of these twelve plan components and four general criteria to determine whether a state’s plan is ‘satisfactory’ under CAA section 111(d)(2)(A).”). EPA fails to recognize, however, that the “satisfactory” criterion under section 111(d) is far more permissive than the substantive requirements of section 110. The requirements of section 110, moreover, are themselves minimal. See Train, 421 U.S. at 79. Nevertheless, the four criteria and twelve components frequently appear to require more of states than submission of a “satisfactory” plan and arguably more than could be required pursuant to section 110. With the limitations on EPA’s authority over section 111(d) plans in mind, the criteria and components are discussed below. EPA’s four proposed approvability criteria are: (1) Enforceable Measures; (2) Emission Performance; (3) Quantifiable and Verifiable Emission Performance; and (4) Reporting and Corrective Actions. 79 Fed. Reg. at 34,909. As to enforceability, EPA acknowledges “there may be challenges to practically enforcing against some such entities in the event of 147 noncompliance” resulting from the untested nature of its Proposed Guidelines. Id. And as explained in Section III.B, there are serious legal problems with the enforceability of EPA’s Proposed Guidelines. Indeed, the Plan TSD also acknowledges the enforceability challenges presented by the Proposed Guidelines, and improperly suggests that it falls to states to resolve these questions. Plan TSD at 13-19. EPA cannot make enforceability a prerequisite to plan approval without first resolving these issues. The second criterion for approvability is that the projected CO2 emission performance by affected EGUs must be equivalent to, or better than, the required CO2 emission performance level in the state plan. 79 Fed. Reg. at 34,910. As described above, however, this leaves states none of the authority section 111(d) provides to establish different performance standards for individual sources based on their individual characteristics and needs. Moreover, EPA appears to be contemplating highly involved demonstration requirements that are not consistent with what the Agency may impose under section 111(d) and that cannot realistically be performed in the time that would be permitted under the Proposed Guidelines. The third criterion is that a state plan specify how the effects of each state plan measure will be quantified and verified. Id. Issues related to monitoring, reporting, and recordkeeping by EGUs are addressed in Section XVI of these comments. EPA also mentions quantification, monitoring, and verification of EE and RE measures. The Plan TSD in particular discusses a requirement that states include in their plans a separate “evaluation, measurement, and verification plan (EM&V plan).” Plan TSD at 36. To the extent EPA includes such a requirement in the final rule and to the extent it prepares guidance on EM&V plans, EPA must first provide adequate notice and opportunity to comment on such issues. It should, moreover, recognize that states have discretion over how to address these issues. 148 The final approvability criterion is that a plan contain provisions addressing annual reporting on plan performance and measures providing for corrective actions. As stated above, states have discretion to determine the manner in which they will address any shortfalls in their plans. In addition to the four criteria, the Proposed Guidelines describe twelve components that EPA would require for a plan to be approvable. These components are: 1. Identification of affected entities; 2. Description of plan approach and geographic scope; 3. Identification of state emission performance level; 4. Demonstration that the plan is projected to achieve the emission performance level; 5. Identification of emissions standards; 6. Demonstration that each emissions standard is quantifiable, nonduplicative, permanent, verifiable, and enforceable; 7. Identification of monitoring, reporting, and recordkeeping requirements; 8. Description of state reporting; 9. Identification of milestones; 10. Identification of backstop measures; 11. Certification of hearing on state plan; and 12. Supporting material. Id. at 34,852, 34,911-14. Many of the components EPA identifies are not, on their face, objectionable policies, such as identifying affected EGUs and other entities subject to the plan. Nevertheless, EPA should recognize that states have primary authority over the contents of their section 111(d) plans and that plans that do not contain these components or that adopt different 149 approaches may also satisfy the CAA’s requirements. EPA cannot disapprove such plans merely because they might fail to include one or more of the components the Agency has created. Furthermore, not all of the proposed components are unobjectionable. These components incorporate requirements already addressed in these comments: demonstration of projected performance, milestones, and corrective measures. The problems identified with these elements of the Proposed Guidelines are not remedied in any way by repackaging them as required plan components. The timeframe for developing and submitting state plans is inadequate to allow for the types of performance demonstrations EPA envisions, and EPA’s milestone and corrective measure requirements would usurp state decisionmaking authority over the most minute details of a state plan. EPA can express its policy preferences for these measures, but it cannot, consistent with its limited authority under section 111(d), impose them as requirements for states. H. Process and Timing for Submittal of State Plans Unlike the specific contents and components of state plans, the CAA does empower EPA to establish “a procedure similar to that provided by section [110] . . . under which each State shall submit to the Administrator” a section 111(d) plan. CAA § 111(d)(1). EPA’s proposed procedures, however, pose practicability concerns and, even in this area, EPA’s Proposed Guidelines would overstep the Agency’s legal authority. The Proposed Guidelines’ discussion of plan submittal and approval processes begins with an acknowledgement of the complexities posed by the Proposed Guidelines and state concerns over the time it will take to develop a plan. 79 Fed. Reg. at 34,915. EPA therefore proposes to require plan submission by June 30, 2016, longer than the nine months allowed for under the existing Subpart B rules. Id. Further, EPA proposes that states may seek an additional year if they are entering into single-state plans and may seek an additional two years if they seek 150 to join multi-state plans. Id. For the many reasons explained throughout these comments, EPA’s Proposed Guidelines are unlawful. But even assuming for the sake of argument that they were lawful, states cannot develop plans fundamentally reconfiguring the electric generation industry and power supply market in the short time EPA proposes to provide. Indeed, the State of Arkansas has confirmed that, if the rule is finalized in June 2015, it “would then have to develop a state plan, seek any required legislation from the General Assembly and only then . . . submit the state plan to EPA for approval.” Arkansas Letter at 2. The state concluded that the timeframe EPA has provided is not sufficient for the state to take those actions, noting that “the General Assembly will not even meet in regular session again to consider any necessary legislation until January of 2017, when the state is scheduled to meet some of the renewable goals.” Id. Further, EPA proposes to require states, in initial submissions, to justify any one-year or two-year deadline extension, and the Agency specifically asks for comments on whether there are justifications for extensions that EPA should reject. 79 Fed. Reg. at 34,915. EPA should take at face value states’ good faith efforts to attempt to comply with EPA’s exceedingly complicated and amorphous Proposed Guidelines and accept any state assertion that more time is needed to develop a plan unless there is clear evidence to the contrary. In addition, EPA proposes a number of required elements for initial plan submission, including a detailed roadmap to plan completion, before any deadline extension will be considered. Id. at 34,915-16. Placing additional burdens on states already struggling to meet extended deadlines is a waste of resources. Moreover, EPA should acknowledge that states have primary authority over plan contents and should defer to state judgments about what should properly be included in an initial 151 plan submittal, consistent with section 111(d)’s requirement that states need only submit plans that are “satisfactory.” With respect to the Agency’s role in reviewing and approving state plans, EPA requests comment on whether it should authorize itself to use the hybrid approval mechanisms it has developed under section 110—the “partial approval/partial disapproval” and the “conditional approval.” Id. at 34,916. Recognizing that disapprovals should be rare given EPA’s limited review authority under section 111(d), EPA should make use of its section 110 hybrid approval mechanisms where use of such mechanisms will assist the states in implementing the section 111(d) program. The Proposed Guidelines also briefly address modifications of state plans. Id. at 34,917. EPA states that such modifications may become necessary as implementation moves forward and that they will be permissible provided the revision “does not result in reducing the required emission performance for affected EGUs specified in the original approved plan.” Id. EPA cites no authority and provides no justification for imposing this limitation. Provided a modified plan reflects consideration of relevant section 111(d) factors, it is “satisfactory,” and the modification should be allowed. Finally, EPA asks for comment on whether it should develop template plans. Id. If EPA is clear that states are not bound to use any template and that states that choose not to follow the template may nevertheless have approvable submissions, development of a template would be appropriate. I. Key Considerations for States The Proposed Guidelines’ section on state implementation includes a subsection addressing what EPA calls “key considerations,” which are issues that cover a wide variety of topics. Each of these key considerations is addressed below. 152 1. Affected Entities Other Than Affected EGUs EPA notes that it wants to give states “broad discretion” to implement the section 111(d) program, including through the portfolio approach EPA has developed to reflect the Building Blocks it used to identify BSER. Id. at 34,917. EPA goes on to admit, however, that the program it has envisioned may prove “challenging to implement.” Id. EPA suggests that it is a key state consideration, therefore, to figure if the program as imagined by the Agency could be workable. EPA has it backwards. In developing the Proposed Guidelines, EPA should have done the work necessary to determine if the program it has devised is practical. Under section 111(d), EPA cannot seek to impose stringent emission control requirements on states without knowing if the plans needed to implement those requirements can be developed. That is the antithesis of section 111’s requirement that “performance standards” reflect levels of emission control that are “achievable” and “demonstrated.” 2. Treatment of Existing State Programs EPA states that it intends not to disadvantage states that have already adopted programs that reduce CO2 emissions from EGUs. Id. at 34,918. With respect to early emission reductions (e.g., reductions from 2005 to 2012), EPA states that those actions are reflected in the 2012 baseline. EPA provides an example: For example, in such instances a significant shift to NGCC generation prior to 2012 may result in a lower potential for further re-dispatch to these units, as witnessed in the 2012 base period data. This would influence the calculated ratebased emission goal for the state, reducing the percentage improvement required relative to the base period CO2 emission rate. Id. at 34,918 n.291. Similarly, EPA says that states with established EE programs “would be closer to, or in some cases already achieving, the level of demand-side [EE] reflected in the state goals.” Id. at 34,918. This approach does not give adequate credit for early action. Therefore, 153 as explained below, EPA should expand its proposed treatment of emission reductions from existing EE programs. EPA proposes that existing programs that will result in future emission reductions can be used for purposes of complying with a plan. Id. EPA proposes that for any existing state requirements, a state may apply toward its required emission performance level the emission reductions that existing state programs achieve during a plan performance period as a result of actions taken after the date of this proposal. EPA further explains that “[t]his option would ensure that actions taken after proposal of the emission guidelines and prior to 2020 as a result of requirements in a state plan, could be recognized as contributing toward meeting a state’s required emission performance level for affected EGUs.” Id. EPA is also requesting comment on two alternatives. Under the first alternative, EPA would change the cutoff for the date by which actions to put EE into place may be taken. Specifically, EPA seeks comment on changing the action date from after the date of proposal to: (1) the start date of the initial plan performance period; (2) the date of promulgation of the emission guidelines; (3) the end date of the base period for the EPA’s BSER-based goals analysis (e.g., the beginning of 2013 for Building Blocks 1, 2, and 3, and the beginning of 2017 for Building Block 4, end-use energy efficiency); (4) the end of 2005; or (5) any other date. Id. Second, EPA requests comment on counting emission reductions that occur not only during a plan performance period but also during a specified date prior to the initial performance period, including (1) the date of promulgation of the final emission guidelines; (2) the date of proposal of the emission guidelines; (3) the end date of the base period for the EPA’s BSER-based goals analysis (e.g., the beginning of 2013 for Building Blocks 1, 2, and 3 and the beginning of 2017 for Building Block 4, end-use energy efficiency); (4) the end of 2005; or (5) any other date. Id. at 34,919. EPA 154 should expand the scope of the existing programs provisions of its Proposed Guidelines to give the broadest favorable treatment possible to EE actions taken to reduce CO2 emissions regardless of whether those actions were taken prior to the proposal of the Proposed Guidelines. Many, if not most, of those actions taken to date have been pursued with the expectation that appropriate credit for early action would be reflected in any new legislation or regulation to address CO2 emissions. Failure to provide adequate recognition for these actions will upset those expectations and will undoubtedly discourage future forward looking actions from being taken as a result. EPA does not give adequate recognition or credit to states that have already undertaken significant measures to increase efficiency, retire coal-fired units, add new renewable or gasfired capacity, or to otherwise reduce GHG emissions. For instance, Mississippi Power invested $600 million to upgrade pollution controls at Plant Daniel, and Entergy recently uprated what is now the country’s largest single-unit power plant, Grand Gulf Nuclear Station. See Kristi E. Swartz, EPA plan gives no credit for Southern states’ strides, utility regulators say in Atlanta, ENERGYWIRE (July 30, 2014), available at http://www.eenews.net/energywire/stories/ 1060003769/print (subscription required). Furthermore, the Tennessee Valley Authority (“TVA”) has reduced its CO2 emissions more than 17 percent below 2005 levels and aims to reduce them to 40 percent below 2005 levels by 2020, exceeding even EPA’s targets. Kristi E. Swartz, Regional carbon approach a ‘difficult concept’ for sprawling TVA, ENERGYWIRE (Aug. 6, 2014), available at http://www.eenews.net/energywire/2014/08/06/stories/1060004112 (subscription required). TVA notes that a new nuclear plant it is building will not count toward its reduction targets in the Proposed Guidelines. Id. Similarly Duke Energy reduced CO2 emissions 20 percent through 2013. Kristi E. Swartz, Duke CEO emphasizes customer benefits 155 in scrutiny of EPA power plant proposal, ENERGYWIRE (Aug. 8, 2014), available at http://www.eenews.net/energywire/stories/1060004240/print (subscription required). In another example, Georgia Power has been penalized for undertaking a nuclear power project whose operation is assumed in the state’s baseline, even though the plant has not yet commenced operation. Jean Chemnick, EPA rule not such a boon for nuclear after all – utilities, GREENWIRE (Aug. 8, 2014), available at http://www.eenews.net/greenwire/stories/1060004265/print (subscription required). Although EPA claims that “states with currently existing programs and policies, and states that put in place new programs and policies early, will be better positioned to achieve the goals,” 79 Fed. Reg. at 34,839, this is not the case as such efforts have been put into the baseline, typically resulting in more stringent state targets. It is unfair for EPA to disadvantage companies and states that have acted to reduce emissions and not credit these measures toward compliance with state goals. 3. Incorporating RE and Demand-Side EE Measures Under a RateBased Approach EPA asks for comment on what approaches should be permissible for crediting or adjusting CO2 emission rates to take RE and EE into account. Id. at 34,919. EPA has no reason—or authority—to circumscribe the mechanisms that are available to states. These issues fall squarely within the purview of state authority under section 111(d) of the CAA. 4. Quantification, Monitoring, and Verification of RE and Demand-Side EE Measures As described above, EPA is developing guidance on EM&V plan requirements for state plans and seeks comment on what that guidance should contain. Id. at 34,920. The Proposed Guidelines note that states and utilities have led the way in developing EM&V standards. Id. As such, this is again an area that is within the purview of the states and one in which EPA should 156 not interfere. EPA’s guidance should provide states with assistance in how they might develop EM&V plans, but it should not attempt to impose EM&V requirements on the states or to burden and thereby discourage the use of EE and RE measures by imposing onerous monitoring and verification requirements. 5. Reporting and Recordkeeping for Affected Entities Implementing RE and Demand-Side EE Measures EPA’s Proposed Guidelines state that plans that incorporate RE and EE measures will require different reporting and recordkeeping requirements than those that would be applicable solely to affected EGUs. Id. at 34,921. EPA, therefore, seeks comment on what would constitute appropriate reporting and recordkeeping requirements for such entities. Id. EPA’s Plan TSD does little more than summarize current efforts by states to address reporting and recordkeeping requirements used to implement existing state RE and EE programs. Plan TSD at 75-81. That TSD goes on to suggest, however, that the Agency might require more than the states’ current programs contemplate. Id. at 82-83. There is no basis for that assertion. EPA’s state goals are based on a Building Block analysis of what states are currently achieving through their own RE and EE programs. It is unreasonable for EPA to rely on these programs for the purpose of developing state goals and, at the same time, to find those programs insufficiently reliable due to the robustness of their reporting and recordkeeping requirements. Moreover, EPA has not established that any additional reporting or recordkeeping requirements are, in fact, needed. EPA merely suggests that “supplemental reporting information . . . may be necessary” or that additional information “may also be valuable.” Id. at 82. These vague assertions do not establish that such information is actually needed or valuable. EPA should instead defer to the states, as is appropriate under section 111(d), and refrain from attempting to circumscribe state discretion. 157 6. Treatment of Interstate Effects The Proposed Guidelines recognize that the electricity market is interstate in nature and that EGUs in one state can provide electricity to customers in other states. 79 Fed. Reg. at 34,921. EPA further notes that state RE and EE programs frequently allow actions in other states, such as renewable generation across state lines, to satisfy intrastate program requirements. Id. The nature of this system creates enormous problems for EPA’s proposal. EPA’s proposed response to these problems is to adopt one set of rules for EE measures, a separate set of rules for RE measures, and yet another set of rules for multistate plans. Id. at 34,922. These simplistic rules, however, will not prevent serious, real-world problems from emerging and impacting states’ abilities to comply with the Proposed Guidelines. Under the Proposed Guidelines, for instance, states would be able to take into account all RE measures implemented for purposes of compliance with a renewable portfolio standard (“RPS”) regardless of whether the renewable energy is generated in-state or out-of-state. Although such an approach will benefit some states with RPS requirements, it could also significantly disadvantage states that have constructed renewable generation that happens to be consumed across state lines. With respect to EE measures, EPA would take the opposite approach, limiting credit for EE-related emission reductions to in-state activities only and ignoring the possibility that some states rely on out-ofstate EE to satisfy their existing requirements. Id. There is no simple solution that can remedy the inequities presented by these conflicting approaches to crediting RE- and EE-related emission reductions, and the simple fact that these problems are so inescapable demonstrates that EPA has strayed far beyond the bounds of its CAA authority. EPA’s Proposed Guidelines will engender conflict among the states. EPA should recognize the seriousness of this issue and withdraw the Proposed Guidelines. 7. Projecting Emission Performance 158 As described above, EPA’s Proposed Guidelines suggest the Agency is seeking to impose unnecessary burdens on states that the Agency is not authorized under section 111(d) to require. States should determine the manner in which they will evaluate the performance of their plans. At the very least, EPA must provide more time to states if it wants to encourage states to undertake the sort of detailed modeling and related work EPA appears to prefer. 8. Potential Emission Reduction Measures Not Used To Set Proposed Goals According to EPA, “States may include measures in their plans beyond those that the EPA included in its determination of the BSER.” Id. at 34,923. States should be free to include any measures they deem appropriate, consistent with the CAA, as part of their section 111(d) plans. EPA requests comment on several additional issues in this section of the Proposed Guidelines. First, it recognizes the possibility of using new NGCC generation to replace higher emitting coal-fired EGUs. EPA asks how that new NGCC generation should be treated for purposes of demonstrating compliance with section 111(d) requirements: The agency requests comment on how emissions changes under a rate-based plan resulting from substitution of generation by new NGCC for generation by affected EGUs should be calculated toward a required emission performance level for affected EGUs. Specifically, considering the legal structure of CAA section 111(d), should the calculation consider only the emission reductions at affected EGUs, or should the calculation also consider the new emissions added by the new NGCC unit, which is not an affected unit under section 111(d)? Should the emissions from a new NGCC included as an enforceable measure in a mass-based state plan (e.g., in a plan using a portfolio approach) also be considered? Id. at 34,924. As an initial matter, as discussed in Section X, new NGCC units are outside the purview of section 111(d) as those units will be subject to the requirements of section 111(b). To the extent new NGCC generation can help offset requirements under section 111(d) plans, however, states should be given wide latitude to decide how to address new NGCC capacity for 159 purposes of calculating compliance. For example, to the extent newly constructed NGCC units displace higher-emitting generation, states that employ the rate-based approach under section 111(d) should be allowed to include the megawatts generated from these newly constructed units in the denominator for the state’s rate. This approach would be similar to how the Proposed Guidelines treat renewable energy and new nuclear units. For states that use the mass-based approach under section 111(d), the emissions from newly constructed units should not be included in the program at all. EPA also requests comment on allowing incremental emission reductions beyond what is required of new units under section 111(b) to count toward compliance with section 111(d). Id. For example, if a new NGCC plant were to install CCS, thereby exceeding the emission reductions required under section 111(b), those additional reductions could be credited toward the section 111(d) program. UARG supports this concept, and EPA should make clear that states are free to find other, similar incremental emission reductions that could also be applied toward compliance with the section 111(d) program. Finally, EPA states that it is developing a framework for counting biomass-related emissions and notes that biomass can be carbon-neutral. Id. at 34,924-25. EPA should work to encourage the use of biomass and avoid unnecessarily burdening its use by seeking to impose unworkable and onerous verification (and similar) requirements. 9. Consideration of a Facility’s “Remaining Useful Life” in Applying Standards of Performance EPA’s Proposed Guidelines include a discussion attempting to justify EPA’s proposed decision to significantly curtail state discretion to consider remaining useful life and other factors in developing a state plan. Id. at 34,925-26. EPA first asserts that deviating from the Proposed Guidelines based on remaining useful life or other factors is simply unnecessary given the 160 flexibility EPA has designed into the Proposed Guidelines. Id. at 34,925. EPA argues instead that states can adjust requirements applicable to specific EGUs to accommodate their needs and make up for any adjustments to apply less stringent requirements to one EGU by making requirements for other EGUs (or other entities) more stringent. This argument assumes that there is enough room within each state’s goal to make such accommodations. As discussed in these comments, the flexibility EPA’s argument relies on does not exist. Indeed, it also assumes that EPA has itself adequately accounted for the remaining useful life of the facilities it seeks to regulate in setting the state goals. It has not. Over the last several years, EPA has approved or directly imposed requirements on numerous facilities, forcing them to install expensive air pollution controls, such as scrubbers and SCR, to control visibility-impairing pollutants pursuant to the Regional Haze Program and to control hazardous air pollutants under the Agency’s MATS requirements. In doing so, the Agency has frequently relied on 20 or 30-year remaining useful life estimates to support the imposition of these costly requirements. See, e.g., 76 Fed. Reg. 52,388, 52,402 (Aug. 22, 2011). Yet, in many instances, compliance with the Proposed Guidelines will require the shutdown of coal-fired units in much shorter timeframes—in many cases by 2020. This is the case, for example, with respect to the Coronado Generating Station (“CGS”) in Arizona. There, EPA imposed a requirement that CGS install SCR to control nitrogen oxides (“NOx”) emissions by December 5, 2017. EPA based that requirement on a 20-year remaining useful life in determining best available retrofit technology (“BART”). See 77 Fed. Reg. 42,834, 42,864 (July 20, 2012). Now, however, EPA’s own analysis supporting the Proposed Guidelines shows that the Agency expects all coal-fired generation in Arizona, including CGS, to shut down by 2020, 161 leaving CGS with hardly more than two years of remaining useful life. Goal Computation TSD, App. 1. EPA similarly failed to take into account major pollution control projects that are mandated by MATS. Thus, as a practical matter, EPA’s argument that deviation from the Proposed Guidelines based on remaining useful life or other factors is unnecessary is incorrect, and EPA’s failure to account for its existing requirements would lead to massive and needless waste. EPA’s argument also depends on its assertion that it has authority under the CAA to place limits on state consideration of remaining useful life. It does not. EPA claims that its existing Subpart B regulations allow for state consideration of remaining useful life and other factors and deviation from EPA emission guidelines based on that consideration, but that EPA is free to take that right away from the states. 79 Fed. Reg. at 34,925 (“EPA has discretion to alter the extent to which states may authorize relaxations to standards of performance that would otherwise apply to a particular source or source category . . . .”). The source of state authority to consider remaining useful life in determining whether to adjust EPA’s emission guideline is the CAA, not EPA regulations. 29 The Act states: “Regulations of the Administrator under this paragraph shall permit the State in applying a standard of performance to any particular source under a plan submitted under this paragraph to take into consideration, among other factors, the remaining useful life of the existing source to which such standard applies.” CAA § 111(d)(1) (emphasis added). This language provides EPA with no discretion to limit or prohibit states 29 Moreover, even if the source of state authority to consider remaining useful life and other factors solely came from EPA’s Subpart B regulations (which is not the case), EPA still must follow those regulations. If EPA wishes to change the Subpart B regulations, it must do so through notice and comment rulemaking. EPA has made clear that it is not proposing to revise its Subpart B regulations in this rulemaking. 79 Fed. Reg. at 34,844, 34,925 n.302. 162 from taking remaining useful life and other factors into account in applying a standard of performance to individual sources. It does not matter if EPA thinks this authority is unnecessary or if its rule might allow states an alternative that could approximate consideration of remaining useful life in some circumstances. The law grants states this right. Finally, EPA argues that it can alter the state’s authority to take remaining useful life into account and to deviate from or adjust EPA’s emission guideline because the emission guideline the Agency has proposed differs so dramatically from the usual “presumptive standard of performance that must be fully and directly implemented by each individual existing source within a specified source category.” 79 Fed. Reg. at 34,925. The fact that the Proposed Guidelines deviate so significantly from other section 111 standards speaks to the validity of the Proposed Guidelines themselves; it does not give EPA carte blanche to rewrite other provisions of the Act to better suit EPA’s distorted rule. J. Multi-State Plan Considerations EPA’s Proposed Guidelines address issues that will affect states that pursue multi-state plans for implementing section 111(d) requirements. Some of these issues are practical. For instance, the Proposed Guidelines assert that states participating in multi-state plans should submit a single plan signed by authorized officials from each participating state and that all required plan components must be included and address these requirements for the group as a whole. Id. at 34,911. EPA also requests comment on whether states should submit a single joint multi-state plan along with individual state plans addressing separately joint and state-specific plan elements. Id. EPA also asks for comment on whether each state participating in a multistate plan should submit its own plan without submitting any joint plan. Id. EPA should allow states to proceed how they wish. Different processes will work more effectively for different parties, and EPA gains nothing by arbitrarily limiting state discretion. 163 The Proposed Guidelines also address joint demonstration of emission performance and propose that multi-state plans include a joint multi-state performance goal calculated based on each state’s individual goal. Id. EPA seeks comment on two approaches for calculating the joint goal, described by EPA as follows: [1.] Under the first option, the weighted average emission rate goal for a group of participating states is computed using each state’s emission rate goal from the emission guidelines and the quantity of electricity generation by affected EGUs in each of those states during the 2012 base year that the EPA used in calculating the state-specific goals. Different levels would be computed for the interim and final goals . . . . [2.] Under the second option, the weighted average emission rate goal for a group of participating states is computed using each state-specific emission rate goal and the quantity of projected electricity generation by affected EGUs in each state. The calculation would be performed for the 2020 through 2029 period to produce a multi-state interim goal, and for 2030 to produce a multistate final goal. Id. at 34,911-12. The results of either approach will differ depending of the mix of generation in each multi-state group. States should be free to decide among themselves which approach works best for their region and adopt either calculation method or any other calculation method they develop or deem appropriate. EPA does not have a basis in section 111(d) for limiting state discretion over this matter. EPA also indicates that it would treat EE differently under multi-state plans than it would under single-state plans. For single-state plans, EPA proposes that “a state could take into account in its plan only those CO2 emission reductions occurring (or projected to occur) in the state that result from demand-side EE measures implemented in the state.” Id. at 34,922. For multi-state plans, however, EPA proposes to allow participating states “to distribute the CO2 emission reductions among states in the multistate area, as long as the total CO2 emission reductions claimed are equal to the total of each state’s in-state emissions reductions that result from demand-side EE measures implemented in those states.” Id. Given the nature of multi164 state plans, attempting to limit EE in such plans based on origin would unnecessarily complicate implementation. Accordingly, EPA’s proposed approach to allow states to determine how to distribute such EE-related emission reductions is appropriate. UARG generally supports allowing states to enter into multi-state plans if they wish. As noted above, however, developing and finalizing multi-state plans in the timeframe allowed by EPA’s Proposed Guidelines would be extremely challenging if not impossible. EPA should provide more time for the development of these plans and provide clear guidelines for multi-state plans. This should include a mechanism for trading between rate-based and mass-based programs. X. EPA Cannot Require States To Regulate EGUs Simultaneously Under Sections 111(b) and 111(d) of the CAA. EPA states that any EGU that becomes subject to a state plan under section 111(d) that subsequently undergoes a modification or a reconstruction must remain subject to the regulatory requirements of the state plan in addition to having to comply with EPA’s requirements for modified and reconstructed EGUs under section 111(b). 79 Fed. Reg. at 34,962-63, 34,965, 34,969, 34,974-75, 34,987. UARG strongly disagrees with EPA’s statement. The CAA does not allow EPA to regulate the same source as both an existing and new source. Although a state may have authority under section 116 of the CAA to impose additional regulatory requirements on a modified or reconstructed source as a matter of state law, EPA cannot compel states to regulate a given unit as both an existing and new source. As discussed below, under the plain language of the CAA, a source that is regulated under section 111(b) because it is a “new” source cannot simultaneously be subject to regulation under section 111(d) as an “existing” source under federal law, and vice versa. 165 In this rulemaking, EPA states that “a modified or reconstructed source would be subject to both (1) the CAA section 111(d) requirements that it had previously been subject to and (2) the modified source or reconstructed source standard being promulgated under CAA section 111(b).” 79 Fed. Reg. at 34,903. Likewise, in its proposed standards for modified and reconstructed EGUs under section 111(b), EPA asserts that “existing sources that are subject to requirements under an approved CAA section 111(d) plan would remain subject to those requirements after undertaking a modification or reconstruction.” 79 Fed. Reg. at 34,974. EPA claims that it can adopt this position because section 111 “is silent on whether requirements imposed under a CAA section 111(d) plan continue for a source that ceases to be an existing source because it modifies or reconstructs.” 79 Fed. Reg. at 34,904. This is simply incorrect. Section 111 speaks clearly and directly to the issue. The CAA states that “[t]he term ‘new source’ means any stationary source, the construction or modification of which is commenced after the publication of regulations (or, if earlier, proposed regulations) prescribing a standard of performance under this section which will be applicable to such source.” CAA § 111(a)(2) (emphasis added). Section 111(a)(6) states that “[t]he term ‘existing source’ means any stationary source other than a new source.” These definitions, by their own terms, are mutually exclusive. Thus, under section 111, an existing source is one that has commenced construction before the publication of regulations (or proposed regulations) applicable to new sources of the same type. Such existing sources are subject to standards promulgated pursuant to section 111(d) of the Act. A new source can come into existence either through greenfield construction that commences after publication of proposed rules applicable to such new sources or through 166 commencing at such time a modification or reconstruction 30 of what would otherwise be an existing source. The critical fact here is that once a modification or reconstruction commences, the “existing” source ceases to exist and a “new” source takes its place. New sources are subject to regulation pursuant to section 111(b). Under the CAA’s explicit terms, section 111(d) cannot apply to sources subject to section 111(b). EPA has long codified this strict division between existing and new sources in binding legislative rules governing CAA programs. For instance, modified and reconstructed sources are regulated as new sources under the CAA’s new source “maximum achievable control technology” program pursuant to section 112(g)(2). In finalizing NSPS for Subpart Da units in 1998, the Agency emphasized that the inclusion of modified sources in the definition of “new source” in section 111(a)(2) is a “clear statutory mandate.” 63 Fed. Reg. at 49,448. EPA’s newfound discovery of supposed statutory ambiguity cannot be reconciled with the Agency’s longstanding implementation of the CAA. Ignoring this clear statutory division, the Agency offers two policy reasons to justify some newly found statutory ambiguity and its proposal to subject new sources to both existing and new source regulatory requirements. First, EPA claims that modified or reconstructed units should remain subject to the existing source program because “[u]ncertainty about whether units would remain in the program could be very disruptive to the operation of the program.” 79 Fed. Reg. at 34,904. Second, EPA argues that it should continue to apply the existing source program to modified and reconstructed sources because “potential discrepancies in the stringency of the 30 A physical or operational change that increases the capacity of an existing source to emit may trigger a “modification” analysis. See, e.g., 40 C.F.R. § 60.14. Replacement of components that involves investments exceeding 50 percent of the fixed capital costs of a comparable, entirely new facility triggers the reconstruction analysis. See id. § 60.15(d). 167 two programs” might “creat[e] incentives for sources to seek to avoid their obligations under a CAA section 111(d) plan by undertaking modifications.” Id. Both of EPA’s concerns would be resolved by withdrawing the proposed standards for modified and reconstructed EGUs. As discussed in Section II.A of the UARG Modified/Reconstructed Comments at 13-14, EPA has discretion not to impose NSPS for modified or reconstructed units and by choosing not to do so in this instance, it would avoid the concerns it has raised. Moreover, even assuming that EPA’s concerns are valid regarding disruption and uncertainty to the section 111(d) program over whether units would leave the program, this does not justify ignoring the plain language of the statute. State plans devised with the correct statutory interpretation in mind can include provisions to accommodate such changes. Indeed, EPA envisions many sources shutting down as a result of the Proposed Guidelines. RIA at 3-34, Tbl. 3-12. If state plans can accommodate units leaving the program through cessation of operations, there is no reason why they cannot accommodate the transition of modified or reconstructed units from the existing source program to the section 111(b) program for new sources. In addition, to the extent EPA has designed a section 111(d) program that cannot function while adhering to the requirements of section 111(a)(2) and section 111(a)(6), it is EPA’s proposal, not the CAA, that must be revised. See, e.g., ASARCO, 578 F.2d at 329 (where EPA is concerned about practical application of proposed rule, solution is to alter standard). Finally, although states can opt to impose additional requirements on modified and reconstructed EGUs as a matter of state law under section 116 of the Act, EPA cannot require the states to regulate a unit as both a new and existing source as this would violate the CAA. It should also be noted that newly constructed EGUs that commence construction after January 8, 2014 (the date of publication of the proposed NSPS rule for new EGUs under section 168 111(b)) are outside of the section 111(d) program. To the extent these newly constructed EGUs displace higher-emitting generation units (such as an NGCC unit that replaces a coal-fired EGU), however, it would be permissible for a state that employs the rate-based approach under section 111(d) to allow the megawatt hours generated by these newly constructed units to be included in the denominator for a state’s rate. This approach would be similar to how renewable energy and new nuclear units are treated in the Proposed Guidelines. For states that use the mass-based approach, the emissions from newly constructed units would not be included in the program at all. With regard to EPA’s concern that there could be an incentive for sources to modify or reconstruct “to avoid their obligations under a CAA section 111(d) plan,” 79 Fed. Reg. at 34,904, for the reasons explained above, a source that undertakes modifications or undergoes reconstruction after the date of the proposal of the applicable NSPS has no obligations under a section 111(d) plan by operation of the CAA; it could only have such obligations as a matter of state law at the state’s discretion. Further, to the extent EPA has proposed discrepancies in the new source and existing source programs that might create incentives the Agency dislikes, the solution is not rewriting the terms of the CAA. See UARG v. EPA, 134 S. Ct. at 2445 (“An agency has no power to ‘tailor’ legislation to bureaucratic policy goals by rewriting unambiguous statutory terms.”). Of course, the reason for this discrepancy is EPA’s decision to propose a section 111(d) program that bears no relationship to the requirements of section 111 or to any section 111 standard EPA has previously written. See, e.g., 61 Fed. Reg. 9905 (Mar. 12, 1996) (municipal solid waste landfills); 42 Fed. Reg. 55,796 (Oct. 18, 1977) (sulfuric acid production units). If 169 EPA were to write a section 111(d) standard that complied with the CAA, this discrepancy would disappear. EPA’s proposed interpretation of the CAA is inconsistent with the statutory text of sections 111(a)(2) and 111(a)(6) and is unsupported by rational policy considerations. EPA must, therefore, revise the Proposed Guidelines to conform to the CAA requirement that new sources subject to section 111(b) requirements (including modified and reconstructed sources) cannot also be required to comply with section 111(d) state plans under the CAA. XI. New Source Review Issues Building Block 1 of EPA’s Proposed Guidelines consists of measures aimed to reduce CO2 emissions from coal-fired EGUs by improving heat rate, which reduces the amount of fuel needed to produce the same amount of electricity. Heat rate improvements increase efficiency and “yield important benefits to affected sources by reducing their fuel costs.” 79 Fed. Reg. at 34,859. In the Proposed Guidelines, EPA noted that “[s]everal studies have examined the opportunities to employ heat rate improvements” at coal-fired EGUs, and specifically cited a 2009 report by Sargent & Lundy as identifying “equipment upgrades . . . at a facility [that] could provide total heat rate improvements in a range of approximately 4 to 12 percent.” Id. Likewise, EPA cites EPA’s 2014 Technical Support Document (TSD) for GHG Abatement Measures, EPA, Technical Support Document for Carbon Pollution Guidelines for Existing Power Plants: Emission Guidelines for Greenhouse Gas Emissions from Existing Stationary Sources, GHG Abatement Measures (Sept. 16, 2014) (clarified version), Docket ID No. EPA-HQ-OAR-20130602-17180 (“GHG Abatement Measures TSD”), which also lists upgrades that may be employed to reduce heat rate. Id. at 34,859 n.114. The projects identified in the 2009 Sargent & Lundy report and the GHG Abatement Measures TSD include upgrades to the following components: soot blowers, boiler feed pumps, 170 economizers, turbines, boilers, air heaters, feedwater heaters, condensers, forced draft and induced draft fans, pulverizers, condensate pumps, flue gas conditioning systems, SCR systems, ash handling systems, neural network optimization systems, electrostatic precipitators, and system controls. See Sargent & Lundy LLC, Coal-Fired Power Plant Heat Rate Reductions at 21 to 5-4 (Jan. 22, 2009) (identifying projects), available at http://www.epa.gov/airmarkets/ resource/docs/coalfired.pdf; GHG Abatement Measures TSD at 2-1 to 2-16 (same). EPA further explained that “EGUs achieve heat rate improvements by . . . installing and using equipment upgrades . . . such as, extensive overhaul or upgrade of major equipment (turbine or boiler) or replacing existing components with improved versions.” GHG Abatement Measures TSD at 216. As EPA knows, however, every one of these upgrades has been targeted by EPA and citizen plaintiffs as triggering the NSR provisions of the CAA. Indeed, shortly after EPA launched its NSR enforcement initiative in 1999, EPA issued a formal applicability determination for a turbine upgrade at Detroit Edison’s Monroe Plant called the Dense Pack project. In that determination, EPA cited that efficiency improvement as a major factor weighing against a finding that the project was routine maintenance, repair or replacement (“RMRR”) and thus excluded from NSR permitting requirements: The purpose of the Dense Pack project, to significantly enhance the present efficiency of the high pressure section of the steam turbine, signifies that the project is not routine …. It would result in greater efficiency above the level that can be reached by simply replacing deteriorated blades with ones of the same design and, in addition, will substantially increase efficiency over the original design. Specifically, the Dense Pack upgrade would not only restore the 7 percent of the efficiency rating lost over the years at each unit but would improve the unit’s efficiency by an additional 5 percent over its original design capacity. Letter from Francis X. Lyons, Reg’l Adm’r, EPA, to Henry Nickel, Hunton & Williams at 2, 3 (May 23, 2000) (“Detroit Edison Determination”), available at www.epa.gov/ttn/nsr/gen/ 171 letterf3.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). EPA went on to assert: In general, a physical change in the nature of the Dense Pack project, which provides for the more economical production of electricity, would be expected to result in the increased utilization of the affected units, and thus, increased emissions. Notwithstanding the fact the Monroe units may be high on the dispatch order, the Dense Pack project would allow Detroit Edison to produce electricity more cheaply per unit of output, thereby creating an incentive to run Units 1 and 4 above current levels. Id. at 4-5. UARG has analyzed EPA notices of violation, citizen notices of intent to sue, and complaints available in the public record since EPA began its NSR enforcement initiative in 1999. Based on that analysis, UARG has identified over 400 efficiency improvement projects targeted by EPA or citizens since 1999 as allegedly violating NSR. These projects are listed in Attachment A and are the same types of projects identified in the 2009 Sargent & Lundy report and GHG Abatement Measures TSD. Moreover, during that same period, EPA and citizens have targeted almost 600 other projects as violating NSR. These projects are listed in Attachment B and—like the projects cited by Sargent & Lundy and EPA—are also important to maintaining and improving the efficiency of coal-fired generating units. Indeed, these projects consist of “like-kind replacement of worn existing components,” which EPA also identifies as essential to maintaining and improving the efficiency of coal-fired generating units. See GHG Abatement Measures TSD at 2-16. EPA recognizes the potential NSR consequences of implementing Building Block 1, but states that it expects there will be “few” instances where “an NSR permit would be required.” 79 Fed. Reg. at 34,928. UARG agrees that very few efficiency projects, if any, should trigger NSR. Efficiency improvements of the types identified in Attachments A and B constitute routine repair 172 and replacement of deteriorated components and do not trigger NSR. See, e.g., Nat’l Parks Conservation Ass’n v. TVA, No. 3:01-CV-71, 2010 WL 1291335 (E.D. Tenn. Mar. 31, 2010) (“NPCA”) (finding economizer and superheater replacements RMRR); Pa. Dep’t of Envtl. Prot. v. Allegheny Energy, Inc., No. 05-885, 2014 WL 494574 (W.D. Pa. Feb. 6, 2014) (finding superheater, lower slope panel, and reheater replacements RMRR). But see United States v. Louisiana Generating, LLC, No. 09-100, 2012 WL 4107129 (M.D. La. Sept. 19, 2012) (finding reheater replacements non-RMRR). Moreover, efficiency improvements of 1 to 6 percent are too small to significantly and identifiably change the dispatch of a unit in a large system, and in any event, such changes are minute with respect to the fluctuations in all other factors that can affect dispatch of a unit at any particular year. But clearly these views are not shared by EPA’s enforcement arm, as is evident from the Detroit Edison Determination and the hundreds of projects targeted in the enforcement initiative since then. Based on the history of EPA’s NSR enforcement initiative, EPA’s statement that there will be “few” instances in which Building Block 1 projects would trigger NSR rings hollow and is hardly reassuring to the utility industry. Over 20 years ago, EPA made similar statements regarding life extension projects amid concerns that EPA’s determination that led to the Seventh Circuit’s decision in Wisconsin Electric Power Co. v. Reilly, 893 F.2d 901 (7th Cir. 1990) (“WEPCo”), would apply to such projects. At that time, EPA told industry that WEPCo would not “significantly affect power plant life extension projects.” Letter from William G. Rosenberg, Assistant Adm’r for Air and Radiation, EPA, to The Hon. John D. Dingell, Chairman, Subcomm. on Oversight & Investigations, Comm. on Energy & Commerce, at 5-6 (June 19, 1991) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Moreover, in the preamble to the 1992 NSR reform rule, known as the WEPCo rule, EPA confirmed that 173 whether repair and replacement projects are judged routine would be determined with respect to industry practice. 57 Fed. Reg. 32,314, 32,326 (July 21, 1992). As EPA knows, however, many utilities relying on these and similar statements regarding repair and replacement projects were then targeted about a decade later for allegedly violating NSR for performing those very projects. See, e.g., United States v. Ala. Power Co., 681 F. Supp. 2d 1292, 1310 (N.D. Ala. 2008) (EPA “could not tell Congress it envisioned very few future WEPCO-type enforcement actions on the one hand, and then argue in subsequent enforcement actions that the utility industry was unreasonable in relying on those, or similar, EPA statements.”). Moreover, even if EPA believes there will be “few” instances where an NSR permit would be required, there is no suggestion that all states or citizens share that belief. Citizen plaintiffs have been just as active as EPA in litigating NSR suits over the past 15 years. Even when those citizen suits have lacked merit, they often delay the implementation of efficiency improvement projects, take several years to litigate, are very expensive, and drain scarce resources of the parties and courts. 31 The NSR citizen suits in Allegheny Energy, Inc., 2014 WL 494574, and NPCA, 2010 WL 1291335, for example, each took approximately nine years to litigate at the district court level. 32 Both resulted in a complete dismissal of plaintiffs’ claims. 31 The same is often true for NSR enforcement actions. For example, in United States v. Cinergy Corp., after more than a decade of litigation, a jury returned a verdict for EPA on only 4 of 14 projects, but even that limited success was reversed by the Seventh Circuit. 618 F. Supp. 2d 942 (S.D. Ind. 2009), rev’d 623 F.3d 455 (7th Cir. 2010). Likewise, EPA’s enforcement actions against Duke Energy and Alabama Power are still pending after nearly 14 years of litigation. United States v. Duke Energy Corp., No. 00-1262 (M.D.N.C. filed Dec. 22, 2000); United States v. Ala. Power Co., No. 01-152 (N.D. Ala. filed Jan. 12, 2011). Trial has not been scheduled in either case. 32 Although New York, Maryland, and Connecticut appealed the district court’s Allegheny Energy decision dismissing their citizen suit to the U.S. Court of Appeals for the Third Circuit, Pennsylvania and New Jersey chose not do so. The appeal is still pending. 174 For this reason as well, EPA’s assurances provide no comfort to UARG’s members, who are routinely targeted by citizens for alleged NSR violations relating to equipment upgrades. This threat of litigation is compounded by the fact EPA maintains it can bring a suit alleging an NSR violation years after a project is undertaken and regardless of whether emissions in fact increase after the project. EPA’s theory of liability under NSR has three fundamental tenets: (1) NSR enforcement can proceed on the basis of any repair or replacement project other than the most minor, repetitive maintenance; (2) for such repair and replacement projects, NSR enforcement can proceed on the basis of a projection alone (regardless of whether post-project actual emissions increase, much less increase as a result of the project), see, e.g., United States v. Ohio Edison Co., 276 F. Supp. 2d 829, 865 (S.D. Ohio 2003) (NSR enforcement can proceed on the basis of projection, even when post-project data are available (applying 1980 and 1992 rules)); United States v. S. Ind. Gas & Elec. Co., 245 F. Supp. 2d 994, 1022 & n.20 (S.D. Ind. 2003) (same (applying the 1980 rules)); and (3) even if a company performed a pre-project projection, EPA can always challenge that projection, see, e.g., United States v. DTE Energy Co., 711 F.3d 643, 650-52 (6th Cir. 2013) (rejecting EPA’s theory). As a result, utilities implementing Building Block 1 that do not project an increase in emissions as a result of an equipment upgrade will face the ongoing threat of NSR litigation, years after their projects. In an enforcement action filed just last year, for example, EPA sued Oklahoma Gas & Electric Company (“OG&E”) for allegedly violating NSR even though emissions since the projects have decreased. United States v. OG&E, No. 13-690 (W.D. Okla. filed July 8, 2013). And EPA brought these claims following the very same types of upgrades it is now recommending to implement Building Block 1—the replacement of economizers and turbine blades, and the addition of heat transfer surface in boilers. See id., Compl. ¶ 42. 175 EPA’s own recognition that “the NSR program has impeded or resulted in the cancellation of projects which would … improve … efficiency” confirms the need for regulatory certainty for utilities implementing Building Block 1. EPA, “New Source Review: Report to the President,” at 1 (June 2002), available at http://www.epa.gov/NSR/documents/nsr_report_to_ president.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Otherwise, EPA’s Proposed Guidelines are destined to create the same uncertainty and encounter the same problems that have plagued the NSR program for the past 15 years. Rather than replow this ground, EPA should eliminate the threat of protracted NSR litigation and provide a clear statement that any upgrades necessary to implement Building Block 1 do not trigger NSR. EPA failed to do so, however, instead waving generally in the direction of “flexibility” to suggest that the states themselves can provide the necessary relief from the efforts of EPA’s enforcement arm and citizen groups for their sources as a part of their state plans under section 111(d). 79 Fed. Reg. at 34,928. EPA’s first suggestion that states could somehow adjust statewide demand side measures and renewable energy requirements so as to counterbalance the supposed impact of unit-specific efficiency improvements on unit utilization is speculative at best. EPA’s second suggestion of essentially imposing synthetic minor limits on all coal-fired sources (because, at least under the NSR enforcement initiative view of NSR, most, if not all, Building Block 1 projects likely trigger NSR) goes against the “flexibility” EPA touts in the Proposed Guidelines. Imposing such synthetic minor limits across the board would, for example, prevent a state plan from relying on the shutdown of smaller, less efficient units and increased utilization of larger, more efficient units. 176 In sum, at least Building Block 1 of EPA’s Proposed Guidelines are based on measures that EPA’s enforcement arm has declared illegal, as EPA has at least implicitly acknowledged. Because EPA’s justification of the state emission goals relies on the ability of sources to implement efficiency improvement measures, and because EPA has failed to propose any credible regulatory provisions to otherwise address this issue, EPA has failed to demonstrate the achievability of its goals as required by section 111. See, e.g., Essex Chem. Corp. v. Ruckelshaus, 486 F.2d 427, 433 (D.C. Cir. 1973) (a “standard . . . must be achievable” under section 111); Nat’l Lime Ass’n v. EPA, 627 F.2d 416, 433 (D.C. Cir. 1980) (EPA bears the burden of explaining “how the standard proposed is achievable under the range of relevant conditions which may affect the emissions to be regulated”). Similarly, EPA’s failure to account for the potential cost of NSR—and NSR uncertainty—for Building Block 1 projects in the Proposed Guidelines is arbitrary and capricious. See, e.g., Star Enter. v. EPA, 235 F.3d 139, 142 (3d Cir. 2000) (EPA must “tak[e] into account the cost of achieving [emission] reduction” under section 111 in setting standards) (internal quotation marks and citation omitted); Pub. Citizen v. Fed. Motor Carrier Safety Admin., 374 F.3d 1209, 1216 (D.C. Cir. 2004) (“an agency’s rule normally is arbitrary and capricious if it entirely failed to consider . . . [a] statutorily mandated factor”) (internal quotation marks and citations omitted). XII. The Proposed Guidelines Are Not Justified by Their Purported Benefits When Those Benefits Are Characterized Properly. If the benefits of the Proposed Guidelines are properly characterized—that is, without inappropriately claiming benefits from the U.S. Government’s social cost of carbon (“SCC”) estimates and the benefits of reducing pollutants that are regulated to protective levels by other rules—there are no benefits to weigh against the exorbitant costs. The Proposed Guidelines therefore cannot be justified and are thus unlawful. In fact, EPA never claims the Proposed 177 Guidelines will mitigate climate change, despite describing the issue of climate change at length as a purported rationale for the rule. While the Agency quantifies the estimated emission reductions that the Proposed Guidelines will achieve, it never makes the connection between those emission reductions and the resulting impact on climate change, raising questions about the purpose of the rule. A “standard of performance” for purposes of section 111 is defined as the degree of emission limitation achievable through the best, adequately demonstrated system of emission reduction, “taking into account the cost of achieving such reduction . . . .” CAA § 111(a)(1) (emphasis added). A section 111 standard that imposes greater costs than benefits is therefore unlawful. EPA’s reliance on the SCC and public health benefits to claim benefits from the Proposed Guidelines is arbitrary and capricious. Any regulation whose costs outweigh its benefits is unlawful and is also ineffective and misguided. See CAA § 301(a) (authorizing EPA to adopt “such regulations as are necessary to carry out [its] functions under [the Act]”; regulations that yield no benefits cannot be “necessary”). A. Costs EPA’s estimate of the costs of the Proposed Guidelines appears to be significantly understated. As discussed elsewhere in these comments, EPA’s Building Block assumptions are in most cases unrealistic and unachievable. Implementing the Building Blocks as designed will almost certainly cost more than EPA has estimated because of the Agency’s unrealistic assumptions about achievability, discussed in Section XIV below. EPA states that the purpose of the rule is to “continue progress already underway to lower the carbon intensity of power generation in the United States.” 79 Fed. Reg. at 34,832. The Agency makes the case that climate change exists and is affecting the United States, id. at 34,841-43; RIA at 1-2, 4-1 to 4-6, but EPA does not even attempt to claim that the rule will have any impact on climate change. The preamble asserts that the Proposed Guidelines will reduce 178 CO2 emissions, 79 Fed. Reg. at 34,931, and claims that emission reductions will be “meaningful,” see e.g., id. at 34,878, but does not assert that those CO2 emission reductions will actually accomplish anything in terms of climate change mitigation. The rule covers only a portion of the power generation sector in a single country—emissions representing less than 5 percent of global GHG emissions. 33 And the rule would reduce only a portion of the emissions of that sector. EPA estimates that the rule would result in a reduction in CO2 emissions of about 555 million tons of CO2 in 2030 for the Option 1 State Compliance Approach. RIA at 3-20, Tbl. 3-5. The IPCC has stated recently that 2010 global anthropogenic emissions of GHGs amounted to 49 gigatons (49 billion) tons per year. IPCC 2014 at 6. That is 1.13 percent of global emissions, assuming full compliance with the rule as proposed—not enough to significantly impact climate change. If the purported rationale of the Proposed Guidelines is to address climate change, but the rule would not in fact do hardly anything to mitigate climate change, EPA should either discard the rule entirely as being unfounded, or be more forthright about the real reasons for the rule. EPA understates the Proposed Guidelines’ likely disruption to the national economy, the negative impact on the reliability of the U.S. energy system, and the adverse impact to U.S. energy security. None of these impacts is easy to quantify. It is difficult to measure the multiplier effects of a wide-ranging regulatory effort that seeks to strand a significant portion of 33 The U.S. State Department reports that 2010 U.S. power sector emissions were 2,342 Teragrams CO2 equivalent. See U.S. Department of State, Projected Greenhouse Gas Emissions, at 82, Tbl. 5-5 (undated), available at http://www.state.gov/documents/organization/140007.pdf. That equates to 2.342 billion metric tons, or about 4.78 percent of global anthropogenic emissions in 2010. See IPCC, Climate Change 2014, Mitigation of Climate Change: Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Summary for Policymakers, at 6 (2014), available at http://report.mitigation2014.org/spm/ipcc_wg3_ar5_summary-for-policymakers_approved.pdf (“IPCC 2014”). 179 the country’s existing electricity generating sector. In an attempt to look at some of these issues, NERA Economic Consulting recently conducted an extensive analysis of the projected impacts of the Proposed Guidelines on energy markets. See NERA Economic Consulting, Potential Energy Impacts of the EPA Proposed Clean Power Plan (Oct. 2014), available at http://www.nera.com/publications/archive/2014/potential-impacts-of-the-epa-clean-powerplan.html. NERA found that implementing EPA’s Option 1, State Compliance Approach, with all four Building Blocks would cause 45 GW of coal-fired EGUs to retire, id. at S-6; cause coalfired generation to decline 29 percent, with natural gas-fired generation increasing 5 percent, id.; and cause gas prices to increase 2 percent and electricity prices to increase 12 percent over the 2017-2031 period, id. NERA also found that the highest energy system costs from the Proposed Guidelines would stem from state energy efficiency programs which would cost about $560 billion in present value from 2017 through 2031. Id. at 21. NERA found that “total electricityrelated costs . . . increase by an average of $34 billion per year from 2017 through 2031 across all sectors,” with the largest of the increases going to residential and commercial consumers. Id. at 26. NERA also examined the potential to implement EPA’s Option 1 utilizing only Building Blocks 1 and 2, and found much more significant negative impacts, including a 71 percent decrease in coal-fired generation, a 55 percent increase in gas-fired generation, “169 GW of coal retirements, a 29% average increase in natural gas prices and a 17% increase in average delivered electricity prices.” Id. at S-6, S-7. In this scenario, “the additional costs of providing electricity services would be about $335 billion,” and “[t]he net result would be an increase in energy system costs by about $479 billion in present value terms over the period from 2017 through 2031.” Id. at S-8. 180 NERA’s attempt to quantify some of the broader effects of the Proposed Guidelines reveals that EPA has significantly underestimated the costs of the proposal, to say nothing of the disruption to the national economy or the effect on the reliability of the U.S. energy system or energy security. It must be stressed that NERA looked only at the impacts of the Proposed Guidelines to the energy system rather than to the macroeconomy at large. A full accounting of the costs of the Proposed Guidelines would examine the multiplier effects of these drastic increases to the energy system on other sectors of the economy, including manufacturing and services, and would almost certainly result in much higher estimated costs. B. Social Cost of Carbon EPA used SCC estimates “to analyze CO2 climate impacts” of the Proposed Guidelines in monetary terms. 79 Fed. Reg. at 34,839 n.12. As a threshold matter, UARG repeats its strong objections to the U.S. government’s development and use of SCC estimates to justify regulation. In response to the Office of Management and Budget’s (“OMB”) request for comments on the U.S. government’s development of SCC values, UARG and others submitted extensive comments on the many legal, procedural, and other shortcomings associated with the development, derivation, and application of the SCC, and urged the U.S. government to discontinue the use of SCC values in the regulatory context. See generally Utility Air Regulatory Group, Comments to OMB on the Notice of Availability and Request for Comments on the Social Cost of Carbon Technical Support Document and Updates, 78 Fed. Reg. 70,586 (Nov. 26, 2013), (Feb. 26, 2014), Docket ID No. OMB-2013-0007-0091 (“UARG Comments to OMB”); Letter from Mary L. Frontczak, Sr. Vice President & Gen. Counsel-Americas, State Gov’t Relations & Commc’ns, Peabody Energy, to Office of Information & Regulatory Affairs, OMB (Feb. 24, 2014), Docket ID No. OMB-2013-0007-0122. UARG’s Comments to OMB included a study conducted by Dr. Anne E. Smith of NERA Economic Consulting cataloguing 181 the many uncertainties associated with the SCC. See Anne E. Smith, Ph.D., Uncertainties in Estimating a Social Cost of Carbon Using Climate Change Integrated Assessment Models (Feb. 26, 2014) (“NERA Social Cost of Carbon”). The UARG Comments to OMB, including Dr. Smith’s study, are incorporated herein by reference and are included in the Supplemental Materials for the Rulemaking Record that UARG is filing with these comments. Among the issues UARG and Dr. Smith identified with the U.S. government’s SCC estimates are that: • They were derived in a nontransparent manner, without being subject to notice-andcomment rulemaking; • The government’s choice of inputs to the Integrated Assessment Models (“IAMs”), and the choice of IAMs themselves, that were used to derive the SCC estimates did not undergo peer review or public comment and are highly uncertain; • There is no rational connection between the inputs and assumptions in the IAMs and the conclusions drawn from the results of the model runs; • The IAMs and SCC cannot account for threshold effects or nonlinear changes that might be ascribed to additional emissions or emission reductions; • The wide range of SCC values that the U.S. government developed is too broad to be useful, but still does not reflect the full range of uncertainty associated with measuring the impacts (both positive and negative) of climate change; • There are serious questions regarding the usefulness of a single dollar amount to represent the asserted climate benefits of rulemakings, particularly given the severe limitations of climate science and the associated uncertainties with estimating the future costs and benefits of a regulation’s impact on climate change; • It is impossible to measure accurately the contribution of a regulation to climate change in isolation, as if one could realistically hold other emissions and change factors constant; and • The IAMs do not recognize the possibility that we will adapt to climate change, or that a reduction in GHG emissions in the United States might result in an increase in GHG emissions elsewhere. See generally UARG Comments to OMB. The Electric Power Research Institute (“EPRI”) also recently undertook a comprehensive review of the U.S. government’s derivation of SCC values and found similar results, including 182 substantial inconsistencies and uncertainties among the three IAMs that detract from the government’s “multi-model approach.” See generally EPRI, Understanding the Social Cost of Carbon: A Technical Assessment, Report No. 3002004657 at xiv, 5-2 to 5-4, 7-1 (Oct. 2014), available at http://epri.co/3002004657. EPRI found that these inconsistencies “raise questions about the statistical comparability of the results, which, along with independence, is statistically required for averaging results as is currently done in the [U.S. government’s] approach.” Id. at xiv. EPRI suggested that the U.S. government’s results therefore “may not be robust,” and that they suggest “[e]xperimental design challenges” and suffer from “a lack of transparency.” Id. at xv. Instead of continuing down the current path, EPRI recommends that the U.S. government undertake “a new framework” subject to external peer review to develop SCC estimates, which would improve transparency. Id.; see also id. at 7-4. EPRI also recommends that the U.S. government undertake “a fuller characterization and discussion of uncertainty and a subsequent analysis of robustness.” Id. at xv. For all these reasons, in particular the non-transparency and the uncertainties involved in the choice of IAMs and the choice of inputs to the IAMs, the lack of public and peer review of the process, and the resulting risk for gamesmanship, the U.S. government’s deployment of SCC values is premature and should not be used in this rulemaking. EPA improperly uses the SCC to justify the Proposed Guidelines. EPA estimates the monetized global climate benefits of Option 1 in 2030 to be $30 billion for the Regional Compliance Approach and $31 billion for the State Compliance Approach. 79 Fed. Reg. at 34,943-44, Tbl. 19. Two reasons in particular demonstrate why EPA should not use the SCC for regulatory analysis: (1) the SCC measures benefits on a global instead of domestic basis while costs are measured on only a domestic basis; and (2) EPA’s use of discount rates is misleading. 183 First, the climate benefits figure of Table 19 reflects the purported global benefits of the Proposed Guidelines, even though the Guidelines would govern only domestic activities. This is inappropriate. Cost-benefit analysis for domestic policy should examine domestic costs and benefits. It is impossible to measure the contribution of a regulation to global climate change in isolation. One cannot know with any certainty how much reducing or continuing to emit a single ton of CO2 will impact global climate change. Future global levels of GHG emissions must be factored into such an analysis, as must a firm sense of how marginal increases in emissions interact with the climate system (e.g., climate sensitivity). Moreover, it is inappropriate for the EPA to amass in its calculation climate benefits that largely will accrue outside the United States, when EPA has no jurisdiction to hold worldwide emissions constant, or to control the climate costs and benefits that may result outside of the United States. The U.S. government has estimated that a domestic SCC value represents about 7 to 23 percent of the global SCC value. Interagency Working Group on Social Cost of Carbon, United States Government, Technical Support Document: Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866 at 11 (Feb. 2010), available at http://www.whitehouse.gov/sites/default/ files/omb/inforeg/for-agencies/Social-Cost-of-Carbon-for-RIA.pdf (“February 2010 TSD”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Thus, assuming such approaches were valid, which they are not, the $30 and $31 billion climate benefit estimates for the Proposed Guidelines, measured more properly on a domestic basis, are closer to the range of $2 billion to $7 billion, which is less than EPA’s estimated compliance costs of that rule ($7.3 billion and $8.8 billion for the Regional and State Compliance Approaches for Option 1, respectively). 79 Fed. Reg. at 34,943-44, Tbl. 19. 184 Second, EPA presents climate benefits figures in the preamble of the Proposed Guidelines at only the 3 percent discount rate, rather than all four discount rates recommended by U.S. government TSDs on the SCC. February 2010 TSD at 1; Interagency Working Group on Social Cost of Carbon, United States Government, Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866 at 2 (May 2013, revised Nov. 2013) (“May 2013 TSD”), available at http://www.whitehouse.gov/sites/default/files/omb/assets/inforeg/technical-update-social-costof-carbon-for-regulator-impact-analysis.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). EPA’s RIA reveals that, using a 5 percent discount rate, the estimated global climate benefits of the Regional Compliance Approach to proposed Option 1 in 2030 are significantly lower—$9.3 billion. RIA at ES-18, Tbl. ES-6. Domestic climate benefits would be only $650 million to $2.1 billion (7% and 23% of $9.3 billion). Compared to EPA’s estimated compliance costs for proposed Option 1 of $7.3 billion and $8.8 billion, it seems clear why EPA chose to omit the 5 percent discount rate figures in the preamble, even though it decided to include estimates at a 7 percent discount rate for air pollution health cobenefits, compliance costs, and net monetized benefits. Had EPA used the 7 percent discount rate OMB recommends, even the global climate benefits of the Proposed Guidelines would almost certainly have been lower than the compliance costs. 34 And converting that global figure to a domestic figure would have reduced the estimate 34 OMB, Circular A-4 Regulatory Analysis at 34 (Sept. 17, 2003), available at http://www.whitehouse.gov/sites/default/files/omb/assets/omb/circulars/a004/a-4.pdf (citing OMB Circular A-94) (“For regulatory analysis, you should provide estimates of net benefits using both 3 percent and 7 percent. An example of this approach is EPA’s analysis of its 1998 rule setting both effluent limits for wastewater discharges and air toxic emission limits for pulp and paper mills. In this analysis, EPA developed its present-value estimates using real discount 185 to de minimis levels. Even if one accepts EPA’s SCC methodology, therefore, any thorough analysis of this issue would demonstrate that the benefits of EPA’s Proposed Guidelines do not justify their costs. In summary, the SCC values are speculative, selective, based on untested and inconsistent economic models, based on climate models that grossly over-predict, and are not tied to domestic policy and results. These issues in both the IAMs used to derive SCC values and the derivation and application of the SCC make SCC values far too speculative to be useful in trying to ascertain the real costs and benefits of the Proposed Guidelines. See UARG Comments to OMB at 4, 8-19; NERA Social Cost of Carbon at 9-29. In addition, because the SCC is currently the subject of an OMB rulemaking to consider these uncertainties and other issues with the SCC, EPA should at a minimum refrain from utilizing SCC values until OMB has finalized its rule on the use of SCC among federal agencies. Without the inflation provided by the SCC, its global measure of benefits, and EPA’s use of low discount rates, the climate benefits of the Proposed Guidelines do not come close to their costs. C. Public Health Benefits EPA asserts that the Proposed Guidelines will lead to health co-benefits of at least $12 billion in 2020, 79 Fed. Reg. at 34,939, Tbl. 16 (based on EPA’s Option 2, Regional Compliance scenario at a 7 percent discount rate); see also RIA at 4-31, Tbl. 4-13 (increasing to at least $23 billion in 2020); 79 Fed. Reg. at 34,937, Tbl. 14 (based on EPA’s Option 1, Regional rates of 3 and 7 percent applied to benefit and cost streams that extended forward for 30 years. You should present a similar analysis in your own work.”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). In the May 2013 TSD, the SCC for the year 2014 is calculated at $11 at a 5% discount rate, but at $56 for a 2.5% discount rate, a nearly five-fold difference with just a 2.5% decrease in discount rates. May 2013 TSD, App. A, Tbl. A1. 186 Compliance scenario at a 7 percent discount rate); RIA at 4-33, Tbl. 4-15. These estimates are complete fiction. Even if it were appropriate to use coincidental decreases in independentlyregulated pollutants other than CO2 to justify the Proposed Guidelines (which it is not), the Agency improperly credits the Proposed Guidelines for purported health benefits from reducing concentrations of fine particulate matter (“PM2.5”) and ozone (“O3”) below levels at which EPA has determined public health is protected. RIA at 4-15. Furthermore, the asserted benefits of the Proposed Guidelines include reductions in PM2.5 and O3 that are already required by existing regulatory programs. Such reductions simply cannot be attributed to the Proposed Guidelines. PM2.5 and O3 are both regulated under the NAAQS program. 40 C.F.R. §§ 50.7, 50.10, 50.13, 50.15, 50.18. EPA sets NAAQS at the level “requisite to protect the public health.” CAA § 109(b). NAAQS incorporate “an adequate margin of safety.” Id. Thus, in establishing NAAQS for PM2.5 and O3, EPA has found that exposure to these pollutants does not endanger public health in areas where the NAAQS are attained. 35 These levels have been chosen after due consideration by the Clean Air Scientific Advisory Committee and EPA of the science. It is disingenuous for EPA to turn around and claim benefits from concentrations determined to be protective of public health, especially to justify a rule targeting a different pollutant. EPA, however, goes even further to claim there are health benefits from reducing PM2.5 and O3 to levels below the NAAQS. 36 RIA at 4-15. Virtually all of the health benefits that EPA 35 78 Fed. Reg. 3086, 3164 (Jan. 15, 2013) (“The Administrator concludes that this suite of standards would be requisite to protect public health with an adequate margin of safety against health effects potentially associated with long- and short-term PM2.5 exposures.”); 73 Fed. Reg. 16,436, 16,483 (“[A] standard set at 0.075 [parts per million (“ppm”)] would be sufficient to protect public health with an adequate margin of safety, and [EPA] does not believe that a lower standard is needed to provide this degree of protection.”). 36 NAAQS must be reviewed and revised, as appropriate, at least every five years. CAA § 109(d). In the event newer scientific information indicates significant health risks remain after 187 claims from the Proposed Guidelines would result from reduced exposure to PM2.5. 37 In projecting the health benefits related to PM2.5, EPA attributes benefits to the lowest modeled PM2.5 levels, including levels well below the NAAQS. 38 RIA at ES-17. Because EPA’s 2016 baseline modeling predicts that no more than five percent of the population will be exposed to PM2.5 levels as high as the 12 µg/m3 annual NAAQS, 39 a high percentage of the purported health benefits must come from exposures at levels below the NAAQS. Because, as discussed above, NAAQS are set at the level EPA has determined is protective of public health, it is improper to claim benefits below that level. Many of the emission reductions—and any associated co-benefits—that EPA attributes to the Proposed Guidelines would occur even without them. Those purported co-benefits cannot existing NAAQS are attained, EPA must revise those standards to address those risks. Thus, the NAAQS process will address even newly-recognized health risks. The Proposed Guidelines are not a necessary or appropriate tool for this purpose. 37 For example, in Table 14, Summary of the Monetized Health Co-Benefits for the Proposed Standard Option 1 Regional Compliance Approach in the U.S., 79 Fed. Reg. at 34,937, the predicted health benefits due to reduced emissions of the O3 precursor NOx range from $0.63 billion to $2.7 billion in 2020 (at a 3 percent discount rate) compared to total predicted PM2.5 health benefits of $15.7 to $34.4 billion in 2020 (3 percent discount rate) (based on adding the benefits of reductions in emissions of the PM2.5 precursors sulfur dioxide (“SO2”) ($12-26 billion), elemental and organic carbon ($0.75-1.7 billion), crustal matter ($0.77-1.7 billion), and NOx ($2.2-5.0 billion)). A similar preponderance of PM2.5 co-benefits is reported in Tables 15 to 17. Id. at 34,938-40. 38 A similar assumption was applied in estimating health co-benefits attributable to O3. See RIA at ES-16. 39 See id. at 4-46, Fig. 4-5 (illustrating graphically that only about 5 percent of the population is exposed to an annual mean PM2.5 level of 12 µg/m3 or greater in EPA’s baseline modeling). Indeed, because (a) the baseline modeling used to generate Figure 4-5 did not reflect emission reductions anticipated from the MATS Rule that would drive baseline PM2.5 concentrations still lower, id. at 4-43, n.100; (b) other analyses by EPA has shown attainment in 2020 nationwide except a few counties in California, see infra note 40 and accompanying text; and (c) states are required to deal with the residual non-attainment, see CAA § 110(a)(1), an even smaller percentage of the population than shown on the Figure would be expected to be exposed to concentrations above the NAAQS. In fact, exposures to 12 µg/m3 could be zero. 188 therefore be properly attributed to the Proposed Guidelines. EPA projects that the vast majority of the country will attain the most stringent existing PM2.5 and O3 NAAQS by 2020, without the Proposed Guidelines. 40 States are required to develop plans to bring into attainment all areas determined to exceed the NAAQS, including those that EPA projects will not be in attainment in 2020. CAA § 110(a)(1). EPA acknowledges that it may be double-counting benefits. RIA at 415 (acknowledging the possibility that some of the estimated benefits of the Proposed Guidelines “may account for the same air quality improvements as estimated in the illustrative NAAQS RIAs”). 41 In fact, the modeling that EPA performed using its IPM 42 to characterize the health co-benefits of the Proposed Guidelines failed to take into account air quality improvements 40 See EPA, Projected Fine Particle Concentrations for Counties with Monitors in 2020 (Dec. 14, 2012) (projecting attainment of the µg/m3 PM2.5 NAAQS everywhere in 2020 except in a handful of counties in California), available at http://www.epa.gov/airquality/particlepollution/ actions.html; EPA, Counties Projected to Violate the 2008 Ozone Standard in 2020 Using Air Quality Model Projections (Mar. 12, 2008) (projecting all but 28 counties will attain the 0.075 ppm O3 NAAQS in 2020), available at http://www.epa.gov/airquality/ozonepollution/ actions.html. 41 EPA acknowledges that the benefits analyses for the section 111(d) proposal and for the PM2.5 and O3 NAAQS are both “illustrative,” but indicates that it has greater confidence in the projected emission reductions for the Proposed Guidelines because they are focused on “one well-characterized sector (i.e., the EGU sector).” RIA at 4-15 to 4-16. The most recent RIAs for both the PM2.5 and O3 NAAQS, however, also assumed additional controls on EGUs in the modeled control strategy. See EPA, Regulatory Impact Analysis for the Final Revisions to the National Ambient Air Quality Standards for Particulate Matter at ES-7 (Dec. 2012), available at http://www.epa.gov/ttn/ecas/ria.html; EPA, Final Ozone NAAQS Regulatory Impact Analysis at 3-10 (Mar. 2008), available at http://www.epa.gov/airquality/ozonepollution/actions.html. Although states would, in the case of NAAQS, have the discretion to focus on another sector for implementation, the reality is that states will likely follow EPA’s lead in focusing on controlling emissions from EGUs under a broad range of rules. 42 According to EPA, the IPM is a “multi-regional, dynamic, deterministic linear programming model of the U.S. electric power sector that the EPA has used for over two decades to evaluate the economic and emission impacts of prospective environmental policies.” GHG Abatement Measures TSD at 3-20. IPM is able to project least-cost capacity expansion and electricity dispatch while also accounting for constraints due to fuel supply, transmission, dispatch, and reliability. Id. 189 associated with several rules for which specific implementation controls have yet to be specified. 43 Thus, it does not take into account many emission reductions from EGUs that have occurred and will occur from implementing the 2006 PM2.5 NAAQS, 71 Fed. Reg. 61,144 (Oct. 17, 2006); the 2008 O3 NAAQS, 73 Fed. Reg. 16,436 (Mar. 27, 2008); the 2010 NAAQS for SO2, 75 Fed. Reg. 35,520 (June 22, 2010); and the 2013 PM2.5 NAAQS, 78 Fed. Reg. 3086 (Jan. 15, 2013). These emission reductions will occur before the Proposed Guidelines are fully implemented. Thus, the NAAQS program and other existing rules—not the Proposed Guidelines—will redress any public health danger from PM2.5 and O3. In short, EPA claims that the health co-benefits from the Proposed Guidelines are equal to or greater than the climate benefits. 79 Fed. Reg. at 34,840-41, Tbls. 1 & 2. Those cobenefits are, however, illusory. Most of them are attributable to exposures to pollutants at levels below those EPA has found protective of public health. Even those fictional benefits are largely due to emission reductions that will occur even in the absence of the Proposed Guidelines in order to address any residual pollutant levels above those found to be protective of public health. Thus, the health co-benefits of the Proposed Guidelines are essentially zero. D. Summary EPA underestimates the costs of the Proposed Guidelines by failing to account for the almost certain multiplier effects of the rule on the macroeconomy, as well as the disruption to, reliability of, and security of the U.S. energy system. Although the purported rationale for the Proposed Guidelines is based on climate change, EPA never makes the case that the proposed 43 In the IPM modeling used by EPA to estimate the benefits attributable to the Proposed Guidelines, EPA accounted only for the MATS Rule and CAIR, both of which focus on EGUs and “other state and Federal regulations to the extent that they contain measures, permits, or other air-related limitations or requirements.” RIA at 3-5. 190 rule would actually mitigate climate change, raising questions about the true purpose of the rule. Moreover, EPA improperly estimated the benefits of the rule by using SCC values, which are speculative, selective, based on untested and inconsistent economic models, based on climate models that grossly over-predict, and are not tied to domestic policy and results. EPA’s use of health “co-benefits” to claim benefits from the Proposed Guidelines is improper given that NAAQS for those pollutants are already set at levels protective of public health. If the benefits of the Proposed Guidelines are properly characterized, there are no benefits to weigh against the costs of the rule. The Proposed Guidelines are therefore unjustified and unlawful. XIII. EPA’s Calculation of the State Goals Impermissibly Limits the States’ Discretion and Is Error-Ridden. A. Despite EPA’s Claims, States Have Little to No Flexibility in Their Ability To Meet the State Goals Contained in the Proposed Guidelines. In calculating the state goals, EPA applied all four Building Blocks to each state’s 2012 electric generation to determine the extent to which the state could reduce its CO2 emissions rate by 2030. 79 Fed. Reg. at 34,863. Although EPA’s methodology in setting the goals is not transparent, in many cases it appears that EPA just assumes that each state can in fact feasibly implement all four Building Blocks, without having done the analysis itself and without having analyzed the effects of implementing all four Building Blocks at the same time, rather than in isolation. EPA says that if a state believes that one of the Building Blocks is technically infeasible or too costly, then it should alert EPA to this fact, implicitly acknowledging that the Agency adopted a one-size-fits-all approach without considering individual state characteristics. Id. at 34,893. But EPA also states that because it believes it is imposing “reasonable” levels of each Building Block (“rather than the maximum”) that a state should be able to increase its use of another Building Block in such case to compensate for a lower or no use of another Building Block. Id. Notably, because the Agency has published only parsed IPM modeling results for the 191 year 2025 and not for 2030—the year for ultimate compliance with the proposed final and interim goals—EPA has not provided commenters with the data necessary to determine whether the selected Building Block implementation levels are actually feasible at a reasonable cost. As discussed in Section III.A, section 111(d) provides that states are to develop performance standards for individual sources as part of state plans, after considering state and local factors and other factors particular to individual emitting units, including the remaining useful life of individual units. The Proposed Guidelines are unlawful because EPA proposes to establish emission goals that it says states have no authority to modify. See Section IX.A. Even if EPA had authority to establish emission standards for the states, because EPA has simply assumed each Building Block can be implemented fully without ascertaining whether that is in fact the case, the touted flexibility of the rule is illusory. See Initial NERC Report at 2, 22 (noting that Proposed Guidelines would require substantial fossil fuel-fired EGU retirements by 2020 and raise significant reliability concerns). In many cases, states cannot realistically implement one or more Building Blocks and do not have the flexibility to increase the stringency of other Building Blocks or measures to achieve their mandated goals. The specific assumptions EPA makes in calculating state targets are also questionable. EPA assumes in the setting of state targets emissions from and capacity for existing units that do not currently exist, that have not yet been fully developed, and that may not ever be completed. In North Carolina, for example, EPA assumed the full operation in 2012 of two combined cycle units that were not in fact in operation. See James Marchetti, “Review of EPA’s State-Specific CO2 Emission Rate Goals: Building Blocks 2 and 3, at 11, Tbl. 6 (Nov. 2014) (Attachment C) (“Marchetti Report”). Duke Energy’s Dan River NGCC plant started operating in December 2012. Another Duke Energy NGCC facility, Lee, started reporting data in January 2013 but both 192 the Lee and Dan River facilities were assumed by EPA in its North Carolina goal calculations to have been operating for all of 2012. These units should more appropriately be credited as “under construction” units rather than as part of the baseline of fully operational units. In addition, EPA identified 2,249 MW of new NGCC under construction capacity when in fact only 625 MW of NGCC was under construction. That was Duke Energy’s Sutton NGCC plant that was under construction but not yet operational by the end of 2012. It began operation in late 2013. Id. at 15. In calculating Tennessee’s goal, EPA included TVA’s Watts Bar Nuclear Unit 2, which is currently under construction. Tennessee cannot meet its state goal without this unit. Because of the many potential issues that could delay or otherwise negatively impact the startup and commissioning of nuclear units like this one, EPA should not include nuclear units under construction in the state goal assumptions. If the Agency continues to assume completion of nuclear units in state goals, then it should allow for an off-ramp or other adjustment if nuclear units are not completed or do not operate as expected. South Dakota has only one NGCC plant, Basin Electric Power Cooperative’s Deer Creek Station. That plant began operating late in 2012, running at about a 1 percent capacity factor. South Dakota also has only one coal-fired plant, Big Stone, owned by Otter Tail Power Company, Montana-Dakota Utility Resources Co., and NorthWestern Energy. The two South Dakota plants are located in different RTOs, however. Deer Creek will be in the SPP as of October 2015, and Big Stone is in MISO. Thus, for South Dakota at least, Building Block 2, or any form of redispatch, is not available, because these two plants have different owners with no contractual relationship, and they are committed and dispatched by two separate entities with unique commitment and dispatch processes. Id. at 18-19. Even if Building Block 2 were 193 feasible, it was also incorrect for EPA to assume Deer Creek was in operation for 2012; instead, EPA should have put that unit into the “under construction” category for purposes of calculating South Dakota’s target. Id. at 11. Building Block 1 is also not an option for South Dakota because Big Stone is undergoing a $384 million upgrade to comply with BART Regional Haze Program requirements. 44 The new station service required by this upgrade—estimated to be as much as 8 MW out of 475 MW—will cancel out any heat rate improvements that the plant might achieve. J. Edward Cichanowicz & Michael C. Hein, “Evaluation of Heat Rate Improving Techniques for Coal-Fired Utility Boilers as a Response to Section 111(d) Mandates” at 2-7 (Oct. 13, 2014) (Attachment D) (“Cichanowicz Heat Rate Report”). In fact, as discussed at greater length in Section XIV.A below, the pollution control upgrades required by MATS and other rules are in many cases substantially lowering the operating efficiencies of existing units that are covered by the Proposal. Id.; see also Julie Wernau, Illinois coal plant owners say they’ve done their part, CHICAGO TRIBUNE (Aug. 19, 2014), available at http://www.chicagotribune.com/business/ct-carbon-rule-coal-0819-biz20140819-story.html (quoting managing director of regulatory affairs at Dynegy as saying that the installation of scrubbers to lower sulfur dioxide emissions reduced plant efficiencies by 1.8 percent). Furthermore, in some cases, significant investments that have been recently made due to MATS and other rules will become stranded by the Proposed Guidelines if they are finalized. Under EPA’s analysis, the Proposed Guidelines will result in the closure of all coal-fired power plants in Mississippi. TVA has the right to purchase all of the output of the 440 MW lignitefired Red Hills Power Plant in Choctaw County, Mississippi, through a power purchase 44 See Big Stone Plant Air Quality Control Systems, http://www.bigstoneplantaqcs.com/ about-aqcs (“Big Stone AQCS”). 194 agreement and has historically purchased most of the facility’s output. The plant began operating in 2002. The plant is currently installing dry-sorbent injection in order to comply with MATS. If Red Hills is required to be shut down due to the Proposed Guidelines, the cost of this injection system would be stranded. All of the examples above—describing several incorrect assumptions made by EPA in computing state goals—are not isolated and run counter to EPA’s assertion that states have compliance flexibility. See Marchetti Report. Finally, EPA appears to have neglected to account for the significant amount of electricity that is imported to the United States from Canada. The overwhelming majority of this power is generated by non-renewable sources, supply that helps to meet U.S. load at emission rates lower than that which would be achieved without the imports. See Canadian Electricity Association, Reducing GHG Emissions Under EPA’s Section 111(d) Guidelines; Comments of the Canadian Electricity Association at 2-3 (Jan. 2014), Docket ID No. EPA-HQ-OAR-20130602-0007. Several U.S. states have formally recognized the significant role this imported power plays in reducing state electricity sector emissions. Vermont, Minnesota, Wisconsin, and Connecticut all credit Canadian imported hydropower for compliance in state renewable portfolio standards. See id. at 4. If EPA fails to recognize the positive role imported Canadian power plays in the United States by allowing for such power to be used to meet state goals, the Agency runs the risk of instituting a non-tariff barrier or other trade distortion by disincentivizing the current free trade of electricity that occurs across the U.S.-Canada border. Under Chapter 9 of the North American Free Trade Agreement (“NAFTA”), technical regulations and standards must be structured to avoid serving as obstacles to trade. North American Free Trade Agreement, U.S.-Can.-Mex. art. 195 904(4), Dec. 17, 1992, 32 I.L.M. 289, 387 (1993) (“No Party may prepare, adopt, maintain or apply any standards-related measure with a view to or with the effect of creating an unnecessary obstacle to trade between the Parties.”). For instance, if the final guidelines essentially mandate a significant ramp-up of natural gas-fired generation, without properly recognizing or crediting Canadian power imports, it is possible that zero-emitting generation from Canada will be displaced in favor of higher-emitting gas-fired generating in the U.S., thus resulting in higher emissions, lower imports from Canada, and greater difficulty for a state in complying with its state targets. If EPA finalizes the Proposed Guidelines, it must explicitly address the issue of Canadian imported power. B. EPA Should Not Include Hypothetical Generation From Units that Are Under Construction When Implementing Building Block 2 in State Goal Calculations. In calculating each state’s goal, EPA assumes that NGCC units that were under construction in 2012 have available generating capacity that may be utilized to displace coalfired generation under Building Block 2. Specifically, EPA assumes that each NGCC unit under construction will be completed and that it would operate at a 55 percent annual capacity factor under a “business as usual” scenario in the absence of the proposed emission guidelines. Goal Computation TSD at 12. This 55 percent annual capacity factor reflects EPA’s calculation of the “average performance of NGCCs that came online in the past 5 years,” although EPA did not provide any data analysis to support that calculation. Id. The Agency claims it “conservatively designated the generation associated with this 55% capacity factor as unavailable for redispatch to reduce CO2 (i.e., not qualifying for building block 2), instead, reserving that amount of generation potential to meet other system needs presumed to have motivated the construction” of those units. Id. Accordingly, EPA assumes that an additional 15 percent of the under- 196 construction unit’s nameplate capacity will remain available for redispatch, bringing the unit to an overall annual capacity factor of 70 percent. EPA’s approach is arbitrary and unreasonable. As a general matter, EGUs that commenced construction before January 8, 2014—even those that were “under construction” during the 2012 baseline year that EPA has proposed to use—would be considered “existing units” for the purposes of section 111(d) once completed, and thus could properly be subject to standards of performance contained in a state plan if they meet the relevant applicability criteria. See CAA § 111(a)(6). But EPA should not rely on generating capacity from “under construction” units when applying Building Block 2 to calculate the proposed state goals. EPA has no basis to assume what generating capacity will be available for redispatch from NGCC units that had no operating history during or prior to the 2012 baseline year. The average annual capacity factor of all recently completed NGCC units, see Goal Computation TSD at 12, is an inadequate predictor of the future generation performance of the individual units EPA includes in the state goals. Such an average does not account for the fact that some units are constructed for the purpose of supplying base load, while others are constructed as load-following units. Recent experience suggests that EPA’s methodology grossly overestimates the generation capacity available from units that were under construction in 2012. For example, the Lee Combined Cycle unit in North Carolina (which EPA mischaracterized as “existing” in 2012 instead of under construction) operated at an 81 percent annual net capacity factor during 2013, its first year of commercial operation. Marchetti Report at 11. Tbl. 6. Yet EPA’s methodology would require this unit to contribute an additional 15 percent of its nameplate capacity to displacing coal-fired generation, effectively pushing the Lee unit above a 96 percent capacity factor based on net capacity. 197 Given EPA’s inability to provide a rational justification for predicting the hypothetical operating duty of units that were not operating in the baseline year of 2012, EPA should exclude “under construction” NGCC units from its application of Building Block 2 for the purposes of calculating state goals. Because these units are technically “existing sources” under section 111, however, they would remain subject to appropriate state plan requirements and would count towards compliance with a state’s interim and final goals if they satisfy the relevant applicability criteria. C. EPA Should Not Penalize States that Have Invested in Nuclear Generation by Including “At Risk” and Under-Construction Nuclear Capacity in State Goals. Under Building Block 3, EPA also includes generation from nuclear units that are under construction and nuclear generation that the Agency considers to be “at risk” in calculating each state’s goal. 79 Fed. Reg. at 34,870. EPA identified five “under construction” nuclear units in Georgia, South Carolina, and Tennessee that it claims will be completed at no “incremental cost,” and that the Agency therefore believes “should be factored into the state goals for the respective states where these new units are located.” Id. Likewise, EPA claims that states can “increase the amount of available nuclear capacity” by “preserv[ing] existing nuclear EGUs that might otherwise be retired.” Id. EPA attempted to quantify this “at risk” nuclear capacity by citing an EIA report that projected nationwide nuclear capacity reductions of 5.7 GW due to “continued economic challenges.” Id. at 34,871. The Agency distributed this amount among the states based on their existing nuclear capacity, “without making any judgment about the likelihood that any individual EGU will retire.” Id. EPA should not include generation from under construction or “at risk” nuclear capacity in the state goals. Although it appears that EPA’s intent is to promote the construction and preservation of nuclear EGUs, incorporating these values into the state goals only penalizes 198 states that have already invested significantly in this low-emitting energy source, while doing nothing to promote nuclear capacity development in states that have not done so already. Marchetti Report at 37-38. The under construction nuclear units substantially lower the state goals for Georgia, South Carolina, and Tennessee by incorporating an assumed 17,392 gigawatt hours (“GWh”) of generation into the goals for Georgia and South Carolina and 8,870 GWh for Tennessee. Id. at 38. Incorporating these under construction nuclear units does nothing to promote the completion of these EGUs—but as EPA recognizes, if unexpected circumstances prevent the units from being completed by 2020 as the Agency projects, it could make it impossible for those states to meet their goals. 79 Fed. Reg. at 34,870. Likewise, EPA’s inclusion of “at risk” nuclear capacity is arbitrary and serves only to penalize states with existing nuclear capacity. The Agency has made no attempt to evaluate whether any nuclear capacity is actually at risk of retirement in any state, instead extrapolating from generalized estimates of nationwide nuclear retirements. The result is that EPA overestimates the amount of nuclear capacity that is at risk (if any) in most, if not all, states. Marchetti Report at 37. And even if some nuclear capacity is at risk in a state (whether at the level EPA estimates or at some lower level), including this capacity in a state’s goal does nothing to incentivize its preservation because states have no control over the primary factor threatening nuclear EGUs: federal license renewal. Thus, because states have little ability to promote nuclear capacity and can only be harmed by including under construction and “at risk” capacity in the goals, EPA’s approach is simply a penalty and not an incentive. To the extent preserving and increasing nuclear capacity needs to be incentivized, that needs to be done by the relevant federal agency (the Nuclear Regulatory Commission) to the extent Congress has given that agency that authority. It cannot be done, however, by EPA through the CAA. 199 D. EPA’s Goal Calculation Methodology Is Riddled With Errors. Even if the Proposed Guidelines were lawful, which for all of the reasons discussed above they are not, EPA’s approach to calculating each state’s interim and final emission goals is fatally flawed. The Agency’s goal calculation methodology contains numerous errors, inconsistencies, and oversights that substantially impact each state’s goals. See generally Marchetti Report at 3-16. These errors do not affect just a handful of states: they undermine the proposed goals for every state in the nation. These basic and widespread errors demonstrate the fundamental inadequacy of the “analysis” EPA has performed to support this rulemaking. EPA must withdraw its Proposed Guidelines and, if the Agency so chooses, can re-propose a revised set of emission guidelines for public comment. EPA cannot simply finalize a new and different set of goals in this rulemaking. States and other interested parties, such as UARG, must have an opportunity to comment on any revised goals. 1. EPA Applies the Building Blocks to Non-Affected Units. EPA’s goal calculation methodology is defective because it applies the measures identified as BSER to sources that do not qualify as “affected EGUs” for the purposes of the Proposed Guidelines. In other words, the proposed state goals assume that state plans will constrain emissions from sources that EPA cannot lawfully regulate under section 111(d). This error affects the state goals because it overstates the number of sources that are available to implement the BSER building blocks—particularly Building Block 2, which is based on shifting generation from higher- to lower-emitting affected EGUs. Under section 111(d), state plans may establish standards of performance only for “any existing source . . . to which a standard of performance under this section would apply if such existing source were a new source.” CAA § 111(d)(1)(A)(ii). In this case, the Proposed Guidelines may be used only to establish standards of performance for existing EGUs that 200 otherwise meet the eligibility criteria for EPA’s proposed NSPS for GHG emissions from new EGUs. See 79 Fed. Reg. at 1430. The NSPS for new EGUs applies to any stationary combustion turbine that, inter alia, has a base load rating greater than 73 MW (250 MMBtu/h) and was constructed for the purpose of supplying, and supplies, one-third or more of its potential electric output and more than 219,000 MWh as net-electrical sales on a 3 year rolling average basis. Id. at 1506, Proposed 40 C.F.R. § 60.4305(c). Therefore, EPA’s Proposed Guidelines may deal only with regulation of existing Subpart KKKK stationary combustion turbines that meet these same criteria. See CAA § 111(d); 40 C.F.R. § 60.21(b); Section V. The preamble to the Proposed Guidelines suggests that EPA intended to follow these applicability criteria and exclude from the state goal calculations any unit that would not be subject to the Proposed Guidelines. See 79 Fed. Reg. at 34,895 n.260. It is clear from the supporting documentation for the Proposed Guidelines, however, that EPA disregarded these applicability criteria and applied the BSER building blocks to include ineligible Subpart KKKK units when determining each state’s obligations. In particular, EPA made no effort to exclude from the Proposed Guidelines NGCC units that were not “constructed for the purpose of supplying, and suppl[y], one-third or more of [their] potential electric output and more than 219,000 MWh net-electric output to a utility distribution system.” 79 Fed. Reg. at 1511, Proposed 40 C.F.R. § 60.5509(a)(1); 79 Fed. Reg. at 34,954, Proposed 40 C.F.R. § 60.5795(b)(2) (defining “affected EGUs” for purposes of Proposed Guidelines). Instead of using available data to determine which units met this one-third sales exclusion, the Agency pooled excluded units together with affected EGUs and applied the building blocks to them. Data available in the docket suggest that a substantial number of NGCC units used in EPA’s goal calculation would qualify for this exclusion. For example, of the 464 plants that EPA examined to determine what 201 capacity factor is achievable for existing NGCC units, 162 had an annual plant-level capacity factor in 2012 that was less than 33 percent, with some plants operating as low as 0 percent. 2012 NGCC Plant Capacity Factor (“2012 NGCC Spreadsheet”) Docket ID No. EPA-HQ-OAR2013-0602-0250. Many of these plants, or individual units within these plants, would likely qualify for the one-third sales exclusion on a three-year rolling average. EPA can, and must, examine historical data for existing EGUs in order to determine which units would be exempt from regulation under the one-third sales exclusion before calculating each state’s goal. This historical data is readily available to EPA. Indeed, shortly before the end of the comment period, EPA included generated data from 2010 and 2011 in the docket for this rulemaking, although it has not performed any analysis of these data to determine which existing EGUs would be exempt and how this would affect each state’s goals. Likewise, in calculating state goals, EPA should apply the measures identified as BSER to Subpart Da units only if they “combust[] fossil fuel for more than 10.0 percent of the heat input during any 3 consecutive calendar years” and “suppl[y] more than one-third of [their] potential electric output and more than 219,000 MWh net-electric output to a utility power distribution system for sale on an annual basis.” 79 Fed. Reg. at 1502, Proposed 40 C.F.R. § 60.46Da(a). The NSPS for new Subpart Da units applies only to sources that meet these (and several other) criteria. Id. EPA proposed to eliminate these applicability criteria in its proposed NSPS for modified and reconstructed EGUs and in its definition of “affected EGUs” for these Proposed Guidelines. See 79 Fed. Reg. at 34,954, Proposed 40 C.F.R. § 60.5795(b)(1); 79 Fed. Reg. at 34,979. As noted in Section VI and in UARG’s comments on EPA’s proposed GHG standards for modified and reconstructed EGUs, however, EPA cannot broaden the scope of the Proposed Guidelines to include units that, if they were newly constructed, would not be subject 202 to the proposed NSPS for new EGUs. UARG Modified/Reconstructed Comments at 69-76. Therefore, EPA must exclude such units from its goal calculation. Excluding non-affected EGUs from the state goal calculation is critical because including these units in the 2012 generation and emission levels affects the stringency of the state goals as they will eventually be applied to affected EGUs. The state goals, as currently calculated, assume that some of the burden of achieving the goal will fall on EGUs that do not fit within the regulated source category. Because states cannot impose standards of performance on these units, the additional burden associated with these units will ultimately fall on affected EGUs that do meet EPA’s applicability criteria. This is particularly important in implementing Building Block 2, where EPA’s purported “system of emission reduction” requires transferring energy generation from coal- and oil/gas-fired steam EGUs to NGCC units with available generating capacity. Including NGCC units that meet the one-third sales exclusion in the goal calculation artificially inflates the amount of NGCC generating capacity that is available for redispatch, which thus inflates the amount by which coal-fired units must reduce generation under Building Block 2. Once this inflated redispatch is incorporated into a state’s goal, affected NGCC units will be forced to operate at capacity factors significantly above 70 percent in order to accommodate the expected generation that states cannot require from non-affected EGUs. Paradoxically, including these units in the goal calculation also assumes that states can take measures to require NGCC units that formerly were exempt by virtue of the one-third sales exclusion to ramp up their utilization to 70 percent, effectively forcing non-affected EGUs to become affected EGUs. This is plainly arbitrary and capricious. The proposed state goals are fundamentally flawed due to EPA’s failure to apply its definition of BSER only to the source category subject 203 to regulation. Because this defect pervades EPA’s entire methodology for calculating state emission goals, EPA must withdraw the Proposed Guidelines. If it so chooses, EPA can repropose a corrected version of the guidelines and allow states and the regulated community to analyze and comment on the revised goals. 2. Errors Regarding 2012 Generation Data As discussed in a report by UARG’s consultant, James Marchetti, EPA’s 2012 data regarding existing and under-construction EGUs contain numerous errors and invalid assumptions that affect its state goal calculations. Marchetti Report at 3-16. In particular, EPA has: (1) improperly used nameplate capacity to calculate generation available for redispatch from NGCC units; (2) double-counted some NGCC units as both existing and underconstruction; (3) identified some under-construction NGCC capacity that does not appear to exist; and (4) improperly counted as “existing” some EGUs that retired or began operation in 2012 or shortly thereafter. Each of these is discussed below. EPA improperly relied on nameplate capacity instead of more realistic measures of unit generating capacity, such as net capacity, in order to determine the level of existing NGCC capacity that the Agency considers to be available for redispatch to displace coal- and oil/gasfired steam EGU generation. Id. at 3-5. Net capacity reflects the maximum output that generating equipment can supply to system load, and it is typically lower than nameplate capacity because it reflects capacity reductions due to electricity use for station service or auxiliary loads such as emission control technology. Further, net capacity for a unit may differ significantly between periods of peak summer and winter demand, reflecting the effects of ambient temperature and other factors on unit capacity. See “Net summer capacity” & “Net winter capacity,” U.S. Energy Information Administration (“EIA”) Glossary, http://www.eia. gov/tools/glossary/index.cfm?id=N. The difference between nameplate capacity and net 204 capacity can be substantial. For example, in Texas, the difference between the total nameplate capacity and summer net capacity of NGCC units in the state is roughly 5,000 MW. Marchetti Report at 4, Tbl. 1. Net capacity is a better indicator of the amount of NGCC capacity available for redispatch purposes because it reflects the amount of energy that these units will actually be able to supply to the power grid. As discussed in the Marchetti Report, using nameplate capacity in the state goal calculations adds substantial levels of NGCC generation that are not actually available for redispatch in Building Block 2. For example, in Florida, the difference in available generation from 70 percent utilization of the actual summer net capacity of 23,784 MW versus the existing nameplate capacity of 29,485 MW is 35,054,309 MWh. Id. In order to support the levels of redispatch assumed in EPA’s proposed state goals, NGCC units would have to operate at statewide average capacity factors in excess of 80 percent and above—much higher than the already-inflated utilization level of 70 percent that EPA assumes. For example, the average annual capacity factors would need to reach 87.0 percent in Florida, 84.2 percent in Arkansas, 81.1 percent in North Carolina, and 80.3 percent in Texas. Id. at 5, Tbl. 2. Because these values are statewide averages and many units cannot reach such high levels of utilization, many NGCC units would have to operate at even higher annual capacity factors. Correcting for these inflated values to use net capacity would significantly increase the state emission goals. Yet in the Proposed Guidelines, EPA has failed even to acknowledge the implications of its use of nameplate capacity rather than other capacity measures, despite the fact that an accurate assessment of the existing NGCC capacity actually available for redispatch is essential to developing an achievable emission guideline using EPA’s building blocks. EPA should use net capacity when analyzing and applying Building Block 2 in order to reflect more 205 accurately existing NGCC generating capacity and the level of NGCC utilization that would be necessary to compensate fully for displaced generation from other sources. In addition, it appears that EPA has included some units’ temporary supplemental generating capacity from the use of power augmentation modes, such as duct burners, in the existing NGCC capacity available for long-term redispatch. Id. at 6. Most NGCC units constructed since the early 2000s include duct burners. Id. Duct burners increase power throughput, but at the cost of substantially increasing heat rate and accelerating component wear. Id. Consequently, continual operation of duct burner capacity is not a feasible option for displacing coal-fired generation and should not be considered as part of a strategy to reduce CO2 emissions. EPA must examine the NGCC capacity values used in its goal calculation and eliminate any duct burner capacity. EPA’s data also contain numerous errors in the Agency’s treatment of certain existing and under-construction EGUs. Id. at 7-13. At least one NGCC unit—Ohio’s Dresden Energy Facility—is double-counted as both existing and under-construction in 2012, meaning that a portion of its generation is redispatched twice in the goal calculation. The amounts of underconstruction capacity listed for California, Colorado, Florida, Mississippi, and North Carolina are also questionable, and the fact that EPA is unable to identify any under-construction units in those states by name suggests that development of those units is too uncertain to rely on those units for redispatch. Likewise, many of the EGUs that EPA has classified as “existing” for the purposes of redispatch under Building Block 2 did not commence commercial operation in 2012 at all. Marchetti Report at 8, Tbl. 3. As UARG’s consultant notes, EPA incorrectly assumes that these units were available to provide generation throughout the year but operated at a zero percent 206 capacity factor, meaning that their entire capacity was available for redispatch under Building Block 2. Id. at 8. The Agency’s faulty analysis added substantial amounts of NGCC generation to the state goal calculation for Alaska, California, Florida, Louisiana, and Utah—and as a result, inappropriately reduced generation from Subpart Da units in equal amounts. Id. at 7, 8, Tbl. 3. Similar mistakes are likely to have occurred for other states. These units should have been considered to be “under construction” and unavailable for redispatch in state goal calculation. 45 EPA also errs in its treatment of “existing” units that either commenced commercial operation, were out of service, or retired in 2012. Units that commenced commercial operation or were out of service during 2012 may report artificially low capacity factors because they did not accumulate a full year of operating data by the end of that year. Conversely, units that retired in 2012 would appear to have capacity available for redispatch that will not actually be available to displace coal generation in the future. For example, South Dakota’s Deer Creek Station appears to have a 1 percent capacity factor based on nameplate capacity, but in fact did not begin commercial operation until August 2012. Goal Computation TSD, App. 7, Docket ID No. EPA-HQ-OAR-2013-0602-0256; Marchetti Report at 11, Tbl. 6. Similarly, the state goals for California, Colorado, Connecticut, Georgia, Idaho, Mississippi, North Carolina, Ohio, and Tennessee (and potentially others) all include artificially low 2012 generation levels from NGCC units that commenced operation during that year. Id. For nearly all of these units, the annual capacity factor increased significantly during the following full year of operation. For example, North Carolina’s Dan River facility commenced operation in December 2012 and reported an annual capacity factor of 45 As discussed in Section XIII.B, no generation from NGCC units that are under construction should be considered available for redispatch. 207 only 1 percent, but reported a 73.2 percent annual capacity factor the following year. Id. In addition, several units (including Washington’s 322 MW Transalta Centralia Generation NGCC plant) were out of service for all or part of 2012, and their reported annual capacity factors may not reflect generation levels under normal conditions. Id. at 9, Tbl. 4. Units that came online or were out of service during 2012 should not be included in their states’ goals so that EPA does not assign generation for redispatch from those units that is actually needed to meet generation obligations from those units that were not reflected in the partial year data. Similarly, EPA should not include generation in its state goal calculations from NGCC units that retired in 2012 or are scheduled to retire in the near future. For example, EPA’s goal calculation for Florida includes generation from the Hansel facility, which retired in October 2012, compare Goal Computation TSD, App. 7, with Marchetti Report at 9, Tbl. 4, and its goal calculation for New Jersey includes generation from the Gilbert facility, which is scheduled for retirement in 2015, compare Goal Computation TSD, App. 7, with 2012 NGCC Spreadsheet. Generation from these NGCC units will not be available for redispatch to displace lost generation from coal- and oil/gas-fired EGUs. Two states provide illustrative examples of the serious errors in EPA’s treatment of existing and under-construction NGCC units, as described in the Marchetti Report. For Florida, EPA’s value for existing NGCC capacity incorrectly includes the 1,210 MW Cape Canaveral NGCC facility (which did not begin commercial operation until April 2013), Marchetti Report at 8, Tbl. 3, and the 122.4 MW Orlando Cogen facility (a compressed storage facility). It also includes the 55 MW Hansel NGCC facility, which retired in October 2012. Id. at 9, Tbl. 4. At the same time, EPA lists 1,157 MW of unspecified under-construction capacity. Correcting these errors would substantially increase Florida’s state goals. Likewise, EPA’s goal calculation 208 for North Carolina includes “existing” NGCC capacity from the 691.1 MW Dan River facility, which only commenced operation in the last month of 2012, and the 1068 MW Lee Combined Cycle Plant, which did not provide any generation to the grid until 2013. Id. at 11, Tbl. 6. The Agency’s widespread errors in its treatment of existing and under-construction NGCC units, and the effects of these errors on the state goals, cast doubt on the accuracy and reasonableness of EPA’s entire goal calculation methodology. EPA must withdraw the Proposed Guidelines and correct its methodology to account for these errors. 3. EPA Assumes Unrealistic Emission Rates for NGCC Units and Subpart Da Units. In calculating the state goals, EPA assumes that each state’s coal-fired units will emit CO2 at an average rate that is 6 percent lower than their average rate in 2012, and that NGCC units will emit CO2 at an average rate that is the same as 2012, even after generation from those affected EGUs is redispatched under Building Block 2. This assumption is unreasonable and implausible. In practice, implementation of EPA’s proposed BSER will increase the average CO2 emission rate of coal-fired EGUs and may also increase the average emission rate of NGCC units. For coal-fired EGUs, operation at low loads increases a unit’s heat rate and thus its CO2 emission rate. As discussed in the Cichanowicz Heat Rate Report and in Section XIV.A below, operating a unit at 50 percent load can negatively impact a unit’s heat rate compared to full load by an amount greater than EPA claims units can improve heat rate through operating practices and equipment upgrades. Cichanowicz Heat Rate Report at 2-1, 2-6. EPA itself acknowledges that EGUs have higher heat rates (and thus a higher rate of CO2 emissions) when operating as load following units and during periods of startup and shutdown. GHG Abatement Measures TSD at 2-5, 2-21. In the related NSPS rulemaking for reconstructed coal-fired EGUs, EPA 209 observed that subcritical units experience a 10 percent increase in heat rate when operating at 50 percent load. Memorandum from EPA OAQPS to EGU NSPS Docket, “Best System of Emissions Reduction (BSER) For Reconstructed Steam Generating Units and Integrated Gasification Combined Cycle (IGCC) Facilities” at 4 Fig. 2 (June 2014), Docket ID No. EPAHQ-OAR-2013-0603-0046 (“2014 Reconstructed EGU TSD”). A primary component of EPA’s proposed BSER in the Proposed Guidelines is to reduce utilization of coal-fired EGUs directly through redispatch to NGCC units under Building Block 2 and indirectly by increasing renewable energy and decreasing demand under Building Blocks 3 and 4. As a result, more coal-fired EGUs will convert from base load to load-following units and operate at lower loads, increasing their heat rates and counteracting any steps those units take to lower heat rate under Building Block 1. Therefore, each state’s average CO2 emission rate for coal-fired EGUs in the state goal calculations should be higher—not lower—than its 2012 level. Similarly, it is unreasonable for EPA to assume without analysis that the average CO2 emission rate for NGCC units in a state will be identical before and after application of the proposed BSER. Under Building Block 2, the Agency expects states to increase utilization of NGCC units to provide more base load generation. Depending on the design of these units, increased utilization may actually increase the overall fleet’s average CO2 emission rate. As EPA noted in its related NSPS rulemaking for modified and reconstructed Subpart KKKK units, some units “are designed to be highly efficient when operated as load-following units, but these design characteristics reduce the efficiency at base load.” 79 Fed. Reg. at 34,980. Many of the NGCC units that have historically operated at lower capacity factors than EPA’s target level of 70 percent may have done so because they operate more efficiently at those levels than when serving base load. 210 In order to develop a defensible goal calculation methodology, the Agency must use realistic CO2 emission rates for affected EGUs that reflect the impact of all four building blocks on those units’ operations and emission rates. The emission rates that EPA uses in the Proposed Guidelines are arbitrary. XIV. EPA’s Building Blocks A. Building Block 1 Is Not Achievable. EPA is proposing to find that overall heat rate improvements of 6 percent (or 4 percent under the alternate goals EPA solicits comment on) are achievable at existing coal-fired EGUs under Building Block 1. 79 Fed. Reg. at 34,861. EPA based its estimates of achievable heat rate improvements primarily on measures described in a 2009 report by Sargent & Lundy, along with the Agency’s own limited analysis of historical generating data. Id. at 34,859; GHG Abatement Measures TSD. EPA’s estimates are arbitrary and capricious, and they demonstrate the Agency’s poor understanding of the nature, cost, and availability of heat rate improvement measures. EPA’s “analysis” of historical data from coal-fired units fails to provide any support for its claim that heat rate improvements of 4 to 6 percent are achievable, and the Agency has failed to consider several critical factors that will make it impossible for affected EGUs to reach this goal individually or on average across any state. As a threshold legal and regulatory issue, many of the operating practices and equipment upgrades that are the basis of the Agency’s assertion that 4 to 6 percent heat rate improvements are achievable are not included in a Subpart Da affected facility. For this reason, EPA lacks authority to regulate such equipment under section 111(d). Subpart Da is the codification of the NSPS for electric utility steam generating units, and these pieces of complicated equipment, logically enough, generate steam. The “affected facility” is “each electric utility steam generating unit.” 40 C.F.R. § 60.40Da. The steam turbines that extract thermal energy from 211 pressurized steam and are used to drive an electrical generator do not generate steam, and for 40 years EPA has never considered them to be part of the affected facility for purposes of section 111. In 1986, EPA had occasion to specify what other equipment is included in the Subpart Da affected facility. Memorandum from John B. Rasnic, Acting Dir., Stationary Source Compliance Div., EPA OAQPS, to James T. Wilburn, Chief, Air Compliance Div., EPA (Nov. 25, 1986) (“Rasnic Memorandum”) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Attachment C to the Rasnic Memorandum includes a drawing that literally draws a line around the boiler island. Equipment within the boiler island is the Subpart Da affected facility, while equipment beyond the boiler island is not. The following equipment is specifically beyond the purview of this rule: steam turbines, water purification equipment, water-supply systems, air cleaning and cooling apparatus, condensors, main exhaust and main steam piping, water screens, motors, and moisture separator for turbine steam. Rasnic Memorandum at 3. EPA’s attempt to require improvements in these components’ efficiency for Building Block 1 is contrary to law. The Supreme Court’s recent caution regarding the Agency’s attempts to pound square pegs into round holes in GHG regulation is on point. Any revision by policymakers of EPA’s long-standing engineering and regulatory understanding of what equipment constitutes the “affected facility” under Subpart Da “should have alerted EPA that it had taken a wrong interpretive turn.” UARG v. EPA, 134 S. Ct. at 2446. Even if EPA had authority to regulate these pieces of equipment under section 111, its technical conclusions are misguided, flawed, and clearly erroneous. As discussed in the Cichanowicz Heat Rate Report, EPA does not account for the fact that the efficiency benefits associated with the measures identified in the 2009 Sargent & Lundy report are highly variable by unit, are not cumulative, and degrade over time. Cichanowicz Heat Rate Report at 3-1 to 3-2. 212 A recent report by the National Coal Council, a federal advisory committee to the U.S. Secretary of Energy, highlighted these basic facts, noting that “[t]he opportunity to apply these efficiency improvements across the existing fleet will vary significantly.” National Coal Council, “Reliable & Resilient – The Value of Our Existing Fleet: An Assessment of Measures to Improve Reliability & Efficiency While Reducing Emissions” at 4 (May 2014) (“NCC Report”), available at http://www.nationalcoalcouncil.org/NEWS/NCCValueExistingCoalFleet.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Measures that may improve heat rate at an individual plant by as much as 1 percent may yield only negligible or nonexistent benefits at many others that have already implemented similar measures or that are otherwise operating in a highly efficient manner. Id. For example, the NCC Report estimates that the benefit of deploying variable speed drives to reduce auxiliary load may vary from 0.05 percent to 0.50 percent, depending on the state of a unit’s existing drive motors. Id. at 65-66. Large benefits from steam turbine upgrades (the highest-payoff measure) are only possible for units that are already severely degraded; for most units, the available payoffs would lie at the low end of the possible range. Id. at 62; Cichanowicz Heat Rate Report at 3-6. Many of the available heat rate improvements for which the efficiency benefits outweigh the costs have already been implemented by EGU owners for economic reasons and would not yield substantial benefits if they were implemented again. Likewise, many of the available actions to improve heat rate do not provide cumulative benefits and thus cannot be added together to estimate the potential efficiency gains at coal-fired EGUs. Cichanowicz Heat Rate Report at 4-4. In particular, measures that increase heat removal from the boiler, such as economizer modifications, improved air heater performance, and low 213 temperature heat recovery, do not provide additive efficiency benefits because any heat that is recovered by an individual project cannot be recovered a second time. NCC Report at 69. In addition, to the extent that any measures to improve heat rate are available at a given unit, the long-term payoffs of many of these measures are significantly smaller than the immediate reduction in heat rate observed after implementation. Id. at 70. Upgraded components begin to incur wear as soon as they return to operation, and will need to be replaced themselves eventually in order to maintain the improved heat rate. For example, while a steam turbine upgrade may improve a unit’s heat rate below its design level, gradual degradation of the turbine blades will reduce the magnitude of that improvement over time from the “new” state without periodic overhauls. Cichanowicz Heat Rate Report at 4-3 to 4-4. Field data indicate that the efficiency of a steam turbine retrofit may decline by 0.5 percentage points in the first 6 months, and by about 4 percentage points after 10 years. Id. at 4-4. Accordingly, EPA cannot rely on the immediate, short-term payoff from efficiency improvements that will degrade over time as the foundation for an emission standard that a source must meet continuously into the foreseeable future. EPA’s Environmental Appeals Board (“EAB”) recently recognized as much in its order denying review of a PSD permit issued by Massachusetts to a new NGCC facility that accounted for efficiency degradation in setting the facility’s GHG BACT limit. In re Footprint Power Salem Harbor Development, LP, PSD Appeal No. 14-02, slip op. at 14-15 (EAB Sept. 2, 2014), available at http://yosemite.epa.gov/ oa/EAB_Web_Docket.nsf/Filings%20By%20Appeal%20Number/9331677B331CFA1285257D4 700716CDB?OpenDocument. Like coal-fired EGUs, the equipment in NGCC facilities can degrade over time and lose efficiency, leading to increased CO2 emission rates. The EAB found that the BACT limit in the permit “appropriately accounted for the degradation of turbine 214 equipment over time that can lead to efficiency losses that directly impact greenhouse gas emissions.” Id. at 15; see also In re Pio Pico Energy Ctr., PSD Appeal Nos. 12-04 through 1206, 2013 WL 4038622, at *49 (EAB Aug. 2, 2013) (upholding GHG BACT limit that accounted for “unrecoverable losses in efficiency over the life of the plant”). Numerous other analyses have noted the same issues regarding the practical limits on benefits from heat rate improvement measures. Notably, a new analysis by Sargent & Lundy (prepared in support of the National Rural Electric Cooperative Association’s (“NRECA”) comments on the Proposed Guidelines) is highly critical of the Agency’s use of the 2009 Sargent & Lundy report to support the Proposed Guidelines. In the new analysis, Sargent & Lundy conclude that EPA greatly exaggerated the amount of heat rate improvements that can realistically be expected at coal-fired EGUs. See Sargent & Lundy LLC, “Coal Fired Power Plant Heat Rate Reduction – NRECA” (Nov. 21, 2014) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). This report emphasizes many of the same issues noted by UARG’s consultants in the Cichanowicz Heat Rate Report, including the fact that actual benefits from many heat rate improvement methods are non-additive, temporal, and highly site-specific. Among other conclusions, the 2014 Sargent & Lundy analysis specifically disavows EPA’s reliance on the earlier report, stating that “Sargent & Lundy’s 2009 report does not conclude that any individual coal-fired EGU or any aggregation of coal-fired EGUs can achieve 6% heat rate improvement . . . or any broad target, as assumed by the EPA.” Id. at C-1. Sargent & Lundy also note that the potential benefits cited in their 2009 report “were estimated at a conceptual level and were not based on detailed site-specific analyses.” Id. The conclusions reached by UARG’s consultants in the Cichanowicz Heat Rate Report and the 2014 Sargent & Lundy report are further corroborated by similar analyses performed by 215 EPRI and by the National Coal Council. EPRI, “Range and Applicability of Heat Rate Improvements,” at 5-1, 6-1 (Apr. 2014) (noting benefits are non-additive and not available at many units) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record); NCC Report at 69-70 (noting benefits are non-additive, site-specific, and may degrade over time). In addition, in its Building Block 1 analysis, EPA ignores the ways in which EGUs’ heat rates (and, thus, their CO2 emissions) are negatively impacted by operating load and by auxiliary power requirements needed to run associated equipment and emission controls, despite paying lip service to these concepts elsewhere. See, e.g., 2014 Reconstructed EGU TSD at 4. In particular, EPA has overlooked the negative impact on coal-fired EGU efficiency of EGUs’ obligations under the remaining components of its proposed BSER and under other CAA programs. Under Building Block 2 of the Proposed Guidelines, EPA is requiring states to redispatch generation from coal-fired EGUs to NGCC units, which will result in more coal-fired EGUs being dispatched as load-following units rather than providing base load at high capacity factors. As EPA itself acknowledges, EGUs have higher heat rates when operating as load following units and during periods of startup and shutdown. GHG Abatement Measures TSD at 2-5, 2-21. Figure 2 below, which is analyzed in greater detail in the Cichanowicz Heat Rate Report, displays representative trends in 2012 heat rate data from Delaware’s Indian River Generating Station Unit 4 at various capacity factors. Operating at its full load of 445.5 MW, Unit 4 achieves a heat rate of approximately 8,250 Btu/kWh, but at 50 percent load this heat rate decreases by about 9 percent to roughly 9,000 Btu/kWh. Efficiency decreases even more dramatically at lower capacity factors. See Cichanowicz Heat Rate Report at 2-3 to 2-6. 216 Figure 2: Gross Heat Rate for Indian River Unit 4 at Varying Loads The Cichanowicz Heat Rate Report evaluated 28 coal-fired EGUs with recent flue gas desulfurization (“FGD”) retrofits and found that operation at 50 percent load would increase unit heat rate by 7.4 percent (for subcritical boilers) and 4.3 percent (for supercritical boilers) on average compared to full load. Id. at 2-6. Thus, the changes in coal-fired EGU capacity factor required by Building Block 2 will most likely eliminate any gains in efficiency achieved through Building Block 1. In addition, parasitic load at an EGU lowers net heat rate because it reduces the amount of gross plant generation that can be supplied to the grid. Auxiliary power loads are often needed to provide power to coal pulverizers, water pumps, and pollution control equipment, among other systems. NCC Report at 65, 70. As discussed in the Cichanowicz Heat Rate Report, retrofitting an existing coal-fired unit with SCR can increase auxiliary power 217 consumption by 1.5 percent, while adding FGD may increase auxiliary load by 1 to 2 percent depending on coal type. Cichanowicz Heat Rate Report at 2-7. In assessing the achievability of Building Block 1, EPA has also failed to consider the fact that other provisions of the CAA will require existing coal-fired EGUs to install additional emission controls that will reduce those units’ heat rates in ways that are not reflected in the 2012 generation data that is the basis for EPA’s proposed state goals. This further reduces the potential gains in EGU efficiency. Many existing coal-fired EGUs will be required to install additional emission controls in order to comply with the 2012 MATS Rule, while others must install further retrofits in coming years to comply with regional haze requirements. For example, South Dakota’s Big Stone Plant is currently undergoing a $384 million upgrade in order to comply with regional haze requirements that is estimated to consume as much as 8 MW of auxiliary power from the plant. See Big Stone AQCS. Likewise, as discussed in Section XI above, EPA and citizen plaintiffs have long targeted efficiency-improving measures like steam turbine upgrades in NSR enforcement suits alleging that such measures constitute “major modifications” triggering strict permitting requirements. To the extent that these groups can successfully argue that NSR requirements apply to Building Block 1 measures, NSR review could result in coal-fired EGUs having to install even more stringent emission controls, further increasing parasitic load. EGU owners typically perform the kinds of upgrades discussed in the 2009 Sargent & Lundy report in conjunction with emission control projects in order to minimize the resulting efficiency losses, but the heat rate improvements are generally outweighed by the new auxiliary load. Because EPA failed to consider the effect of looming emission control requirements on EGU heat rate, affected EGUs will be forced to overcome the energy penalties associated with these controls, and then achieve an additional 4 to 6 percent improvement in heat 218 rate on average in order to comply with the Proposed Guidelines. EPA is not only ignoring the CAA, but the laws of thermodynamics as well. In addition to these fundamental oversights, EPA has failed to provide any reasonable analysis of whether its heat rate improvement targets are achievable. For several reasons, the “assessment of heat rate improvement potential” in the GHG Abatement Measures TSD does not support EPA’s claim that state-wide heat rate improvements of 4 to 6 percent are achievable. See GHG Abatement Measures TSD at 2-30 to 2-34. First, EPA’s so-called “statistical analysis” of heat rate improvements available through “best practices” is fundamentally flawed and does not support EPA’s chosen target of 4 percent. Indeed, it does not support any estimate of heat rate improvements at individual units or for the fleet as a whole. In particular, EPA’s “model” assessing potential heat rate improvements via “best practices” is unfounded, arbitrary and capricious, and misleading. In this “model,” EPA grouped hourly heat input data into “bins” representing ranges of load and ambient temperature, then performed crude calculations to reduce the variation among observations in each bin by arbitrary increments of 10, 20, 30, 40, and 50 percent. Id. at 2-30 to 2-31. The Agency used these adjusted values to calculate an adjusted heat rate for each unit, then compared the study population’s average adjusted heat rate under each scenario to the population’s actual average heat rate in order to determine the “improvement” under each scenario, with a roughly 30 percent reduction in each unit’s variation corresponding to a 4.0 percent improvement in the overall study population’s heat rate. Id. at 2-31. Accordingly, EPA “conservatively” proposed to find that a 4 percent improvement in heat rate is achievable by applying “best practices” of EGU operation because that level “is in the middle of the range of options” that EPA explored. Id. at 2-34. 219 As detailed in the attached report by UARG’s consultants, this conclusion is arbitrary and totally unfounded. J. Edward Cichanowicz & Michael C. Hein, “Critique of EPA’s Statistical Evaluation Defining Feasible Heat Rate Improvements,” at 8-9 (Dec. 1, 2014) (Attachment E) (“Cichanowicz Statistical Critique”). Conspicuously absent from EPA’s “analysis” is any discussion linking the 4 percent “best practices” estimate to any identifiable, technically feasible heat rate improvement measures. See id. at 9. At most, EPA’s discussion shows that a 4 percent reduction in overall heat rate is mathematically feasible if sources can find a way to reduce variation in hourly heat input. The Agency makes no attempt to show whether the measures it identifies as “best practices” are technically capable of reducing heat input variation at individual units at all, let alone reducing variation to a sufficient degree to reach the target efficiency level. Indeed, as UARG’s consultants note, if EPA had undertaken such an analysis it would have realized that the 9 “No Cost or Low Cost Options” the Agency identified as “best practices” are incapable of achieving such reductions as a technical matter. Each individually offers 0.1 to 0.35 percent improvement at most, and several of these options have benefits that are not additive. Id. at 10. Moreover, the 10, 20, 30, 40, and 50 percent reductions in heat input variation that EPA calculates at each unit do not represent varying implementation levels of “best practices” measures: they were apparently selected without any technical basis. Thus, EPA’s decision to “conservatively” select a level of reduction that is “in the middle of the range of options” it arbitrarily chose is itself arbitrary. 46 GHG Abatement Measures TSD at 2-34. 46 EPA states that its target of 4 percent heat rate improvements through “best practices” is equivalent to asking each unit to operate at its “best three-year moving average” heat rate. GHG Abatement Measures TSD at 2-34. This explanation also does not support the 4 percent “best practices” target. EPA does not discuss whether the difference between any unit’s current heat rate and its best three-year moving average is due to operational factors within the unit’s 220 Moreover, on a more fundamental level, EPA’s analysis incorrectly assumes that the existence of variation in heat rate at individual units means there is “significant variation in the operation of EGUs” indicating that “significant potential for heat rate improvement is available through the application of best practices.” Id. at 2-30. In fact, heat rate variability at individual units is driven by the design, duty cycle, fuel type, size, cooling conditions, and location of each unit, and it cannot be ameliorated by changes in operating practices. Cichanowicz Statistical Critique at 5-8. EPA did not assess differences in variability based on these factors or control for these factors in its achievability analysis. Rather, the Agency’s assessment was based on a simplistic and deeply flawed regression analysis of hourly unit capacity factor and ambient temperature on gross heat rate and on the Agency’s observations regarding “residual” heat rate variation not accounted for by this regression. Id. at 3. The report by UARG’s consultants shows that much of the “unexplained” heat rate variability that EPA assumed could be reduced by implementing best practices actually reflects the impact on heat rate of differences in individual units’ age, capacity factor, fuel type, boiler design, and cooling conditions, id. at 6-8, and that EPA’s regression itself is flawed because its results defy known technical relationships between heat rate, operating duty, and ambient temperature, id. at 3-6. In any event, many of the “best practices” that EPA identifies in the TSD are not operating practices at all, but instead involve replacing or upgrading physical components of the EGU or related power plant equipment. See GHG Abatement Measures TSD at 2-33, Tbl. 2-13 (listing “No-Cost and Low-Cost Options” EPA equates with “best practices”). These actions have no discernible connection to variability in a unit’s heat rate. Although EPA quantifies the control, or whether it would be achievable for a unit to consistently achieve a heat rate equivalent to its best three-year moving average using “best practices.” 221 potential benefits of these measures in terms of reduced heat rate variability, the Agency presents no evidence that the “best practices” measures actually have the effect of reducing heat rate variability (as opposed to improving a unit’s average heat rate while maintaining a similar range of variation in individual observations). Second, EPA’s assessment of available heat rate improvements through “equipment upgrades” is similarly arbitrary. EPA set this component of the target by identifying the four most costly heat rate improvement methods listed in the 2009 Sargent & Lundy report based on average estimated $/kW costs, then simply adding together the average estimated Btu/kWh improvements for each. Id. at 2-33. Based on this calculation, EPA concluded that the four specified equipment upgrades could provide a 4 percent heat rate improvement if all were applied on an EGU that has not already made them, but “conservatively” reduced the target to 2 percent because “some units may have applied at least some of the upgrades.” Id. at 2-34, 2-35; 79 Fed. Reg. at 34,860 (“We therefore propose to use as a data input for purposes of developing state goals an estimate that, on average across the fleet of affected EGUs, only half of the full equipment upgrade opportunity just described remains.”). As discussed above, not only does this methodology include turbine and other upgrades that are beyond the scope of section 111, it erroneously assumes that the heat rate improvements from these upgrades are cumulative and that they provide consistent long-term benefits. In reality, the combined payoff from these four upgrades will be less than the sum of each measure’s payoff would be when applied alone to an EGU, and these heat rate benefits will begin to degrade immediately once the unit returns to service. Also, EPA’s reliance on the average Btu/kWh improvement of each measure is flawed because the potential payoff of each upgrade is highly unit-specific and is skewed towards lower values. See NCC Report at 71 222 (“The benefits and cost are highly variable and depend on the specifics of any one site.”). For example, the NCC Report notes that large gains from turbine overhauls are possible only for “units that are severely degraded,” and potential benefits from variable frequency drives depend heavily on existing equipment. Id. at 62, 65-66 (emphasis added). Moreover, one of the measures EPA identifies—acid dew point control—is not broadly available and has not even been proven for most of the generating fleet. Cichanowicz Statistical Critique at 10 (discussing achievability of 2 percent heat rate improvement through “equipment upgrades”). In addition, EPA’s attempt to account for individual characteristics by “conservatively estimat[ing]” a potential improvement of 2 percent instead of 4 percent is arbitrary and unreasonable. GHG Abatement Measures TSD at 2-35. Reducing the already-inflated target by half without further analysis is not “conservative,” it is merely guesswork. The actual potential for efficiency improvements via equipment upgrades is likely much lower, given that “[m]any of these measures have been already applied on units in the existing inventory.” NCC Report at 71; Cichanowicz Statistical Critique at 10. EPA cannot require affected EGUs to duplicate emission reductions they have achieved using measures they have already taken. Therefore, EPA must assess what efficiency upgrades have already been implemented by affected coal-fired EGUs in order to determine what level of additional heat rate improvements is achievable. The Agency has simply failed to adequately prepare the necessary background information for this rule. Likewise, EPA must also assess other factors that will determine what level of heat rate improvements is achievable for individual coal-fired EGUs across the source category as a whole, such as whether the upgrades are compatible with the design and hardware present at a unit and whether the upgrades may present operational reliability issues at a particular unit. See NCC Report at 71. 223 Finally, the individual EGUs to which EPA points as examples provide no support for the Agency’s claim that Building Block 1 is achievable. EPA relies heavily on its analysis of historical CEMS-derived gross heat rate data from “16 EGUs that reported a single year-to-year heat rate improvement of 3-8%.” GHG Abatement Measures TSD at 2-32; see also “GHG Abatement Measures Units with Heat Rate Changes,” Docket ID No. EPA-HQ-OAR-20130602-0236 (listing the 16 identified units). According to EPA, significant year-to-year improvements at all other units in the study population were due to “changes in capacity factor, reporting method, or other events.” GHG Abatement Measures TSD at 2-32. For two of the sixteen units, the Agency claims to have identified “equipment upgrades” that EPA speculates are “responsible for 2-3% heat rate improvement” based on estimates contained in the 2009 Sargent & Lundy report, while concluding that unspecified equipment upgrades were “the most likely cause of some of the observed heat rate improvements” at the remaining fourteen units. Id. EPA apparently never contacted the owners of these units to determine what factors were responsible for the apparent reductions. UARG’s consultants, however, contacted the owners of the sixteen units EPA singled out in the TSD in order to determine the cause of these reported year-to-year heat rate improvements. J. Edward Cichanowicz & Michael C. Hein, “Critique of EPA’s Use of Reference Units To Select Heat Rate Reduction Targets” (Nov. 25, 2014) (“Cichanowicz Reference Unit Report”) (Attachment F). Based on discussions with the owners of these units, in virtually every case the EPA-calculated heat rate improvement was due to a change in stack flow monitor calibration or other unrelated sources of normal variability in CEMS measurements and were not the result of purposeful efforts to improve unit efficiency. Id. at 1; Ralph L. Roberson, “Real Heat Rate Improvement or Measurement 224 Variability/Uncertainty” at 8 (Nov. 25, 2014) (“Roberson Report”) (Attachment G). Figure 3 below summarizes the reported causes of observed CEMS-derived heat rate improvements at each of the sixteen units. Figure 3: Causes of EPA-Calculated Heat Rate Improvements at 16 EPA Reference Units Owner Unit Name Cause of Apparent Heat Rate Improvement AES Petersburg 1 Gross heat rate reduction by ~3% in 2003 due to steam turbine “dense-pack” retrofit. Other reductions due to CEMS changes. Net heat rate increases 2-3% prior to 2004 due to SCR retrofit and operating variables. City of Springfield MO Southwest 1 Almost all gross heat rate reductions appear to be due to unrelated variability in CEMS measurements. Steam turbine dense-pack derived a 2% reduction in the year following retrofit. Duke Energy Gibson 1 Unrelated CEMS measurement changes noted with retrofit of FGD and dedicated stack. Net heat rate based on coal use exceeds CEMS-derived values by 5-10%. The CEMS gross heat rate data lower than unit design basis. Dynegy Newton 1 Dynegy evaluating data. Georgia Power Gorgas 1 Georgia Power states most changes in reported gross heat rate are due to the variability and relative accuracy associated with using CEMS-based heat input data. CLECO/ Lafayette Utilities System Rodemacher 2/ Brame Energy Center All reported gross heat rate reductions due to unrelated variability in CEMS measurements. Modest decrease in net heat rate (~2%) noted over 11 years but not due to targeted actions. NV Energy North Valmy All reported gross heat rate reductions due to unrelated variability in CEMS measurements. Nebraska Public Power Sheldon 1 Gross heat rate when calculated using plant monitoring equipment increases over the reporting period. A reported gross heat rate reduction of 10% in 2006 is due to CEMS recalibration. 225 Owner Unit Name Cause of Apparent Heat Rate Improvement Pacificorp Dave Johnston 2 All changes CEMS-related; most due to calibration of stack gas temperature probe in 2005 and 2007. 24 month rolling average of net heat rate increases ~3%. Pacificorp Dave Johnston 4 Steam turbine upgrade in 2009. Replacement in 2011 of: legacy FGD with dry FGD; new stack; rebuilt boiler. CEMS location changed. The reported gross heat rate reduction of 10% is due to unrelated variability in CEMS measurements. Net heat rate lower ~1% vs. 2009. Pacificorp Jim Bridger 3 All reported reductions due to unrelated variability in CEMS measurements; notably, replacement of the stack flow monitor in 2005. 24 month rolling average of net heat rate decreased by 1.7% due to changes in operation. TVA Colbert 1-3 EPA conclusions regarding unit gross heat rate are compromised as heat input data is reported from common stack CEMS and apportioned by load, based on assumed identical heat rate. Yearly differences in unit operating conditions influence results. Any actual changes in reported gross heat rate are largely due to coal blend. Wisconsin Public Weston 3 Service Heat rate varies between 11,000 and 9,000 Btu/kWh over four years. All reductions in reported heat rate due to unrelated variability in CEMS measurements. Xcel Refurbishment: boiler cyclones and heat exchangers; coal handling, steam turbine internals; cooling tower. Replaced legacy FGD and ESP with dry FGD and fabric filter; also SCR. Net heat rate actually increases vs. 2003 due to higher station load for environmental controls. Allan King 1 In other words, the reported heat rate improvements do not reflect actual changes in unit efficiency or reductions in unit CO2 emission rates. The owners of eleven units report that all changes are due to unrelated variability in the CEMS-based measurements. Cichanowicz Reference Unit Report at 1. Some units did report significant initial heat rate improvements 226 from equipment upgrades, such as upgrading the steam turbine, but these improvements were generally more than offset by increases in net heat rate due to the retrofit of environmental controls. Id. Indeed, the only conclusion that EPA’s analysis of the 16 reference units supports is that CEMS-derived gross heat rate data contain too much unrelated normal variability to reliably quantify relatively small changes in unit heat rate, particularly when averaged over a year. Id. at 2. This conclusion is underscored in the attached report from UARG’s consultant, which details numerous reasons why CEMS data reported under Part 75 are inappropriate for the purpose to which EPA applies them in these Proposed Guidelines, despite Part 75’s stringent quality assurance criteria. Roberson Report at 3-8. Rather than reflecting actual changes in unit heat rate, variation in these values may reflect changes in flow monitor calibration, calibration reference methods, bias adjustment factors, stack diameter measurement, or monitoring technology, and physical changes in the flue gas handling system. See generally id. In addition, where units share a common stack, EPA’s data handling procedures permit the owner to apportion the observed data between those units based on load—effectively assuming that those units’ heat rates are equal. Id. at 5-6. EPA has failed even to acknowledge, let alone account for, these sources of variability in CEMS-derived gross heat rate data. In sum, EPA has failed to satisfy its obligation to demonstrate that its proposed Building Block 1 target of 4 to 6 percent improvement in heat rate for coal-fired EGUs statewide is achievable “under the range of relevant conditions which may affect the emissions to be regulated.” Nat’l Lime Ass’n v. EPA, 627 F.2d at 433. The Agency failed to consider important factors affecting the achievability of its proposed targets, including operating load, auxiliary power consumption, source-specific variability in available upgrades and associated benefits, and 227 the non-additive and temporal nature of some heat rate improvements. EPA’s analysis of historical data from coal-fired EGUs is deeply flawed and does not support the Agency’s claims. B. Building Block 2 Is Not Achievable. EPA is proposing to find that it is achievable for affected EGUs in each state to shift generation from existing coal- and oil/gas-fired steam EGUs to existing NGCC units until those NGCC units reach a statewide maximum capacity factor of 70 percent (or 65 percent under the alternate goals EPA solicits comment on). 79 Fed. Reg. at 34,864-66. EPA bases this conclusion on its observation that of 464 NGCC plants it identified with generation data in 2012, 10 percent had a capacity factor of 70 percent or greater. GHG Abatement Measures TSD at 3-8 to 3-9. Using a different set of data, EPA also observed that some units are capable of operating at greater than 70 percent capacity factor on a seasonal basis, with 16 percent and 19 percent of units operating at or above that level in the 2012 winter and summer seasons, respectively. 79 Fed. Reg. at 34,863. Relying on this data, EPA “assumed that 70% was a reasonable fleet-wide ceiling for each state” on an annual average basis. GHG Abatement Measures TSD at 3-9. This conclusion is not supported by EPA’s analysis, which fails to demonstrate that the 70 percent target for Building Block 2 redispatch is achievable by affected EGUs in the source category. In addition to the data errors described in Section XIII above, which misstate the amount of NGCC generation available for redispatch in each state, EPA’s Building Block 2 analysis is insufficient because it fails to account for factors limiting the ability of EGU owners to increase utilization of their NGCC units. For example, some units are unable to increase generation to any level approaching 70 percent due to technical limitations, permit limits, or gas and transmission infrastructure constraints. Other NGCC units will not be able to generate sufficient amounts above 70 percent to make up for these lower-utilization units. 228 EPA’s assumption that each state’s entire fleet of existing NGCC units can match the operational level of its top 10 percent of units is unsupported and arbitrary. EPA did not undertake any assessment of the differences between high- and low-capacity factor NGCC units that may have led a small subset of those units to operate above a 70 percent capacity factor in 2012. The Agency acknowledged that units operating above 70 percent on an annual basis were “largely dispatched to provide base load power,” and that units operating above 70 percent on a seasonal basis typically “were idled or operated at lower capacity factors” during periods of lower demand. Id. But EPA did not examine whether the NGCC units providing base load power have different characteristics from the other existing NGCC units that are expected to provide generation for redispatch, or whether units that were idled during periods of relatively low demand did so because of economic, technical, or regulatory constraints on their operations. Instead, EPA assumed that all NGCC units are identical. This is plainly unreasonable. In order to demonstrate that a standard is achievable, EPA must “establish that the test data relied on by the agency are representative of potential industry-wide performance, given the range of variables that affect the achievability of the standard.” Sierra Club v. Costle, 657 F.2d 298, 377 (D.C. Cir. 1981). This determination cannot be based on “mere speculation or conjecture.” Lignite Energy Council v. EPA, 198 F.3d 930, 934 (D.C. Cir. 1999) (per curiam). EPA has failed to establish that the 10 percent of existing NGCC units operating at 70 percent capacity factor or higher are representative of the remainder of NGCC units in the source category. Indeed, many of these high-utilization units are likely not representative of the source category, given that EPA excluded a number of them from its calculations of each state’s existing NGCC capacity in the Goal Computation TSD. Compare 2012 NGCC Spreadsheet (including, inter alia, Martinez Refining units in assessment of achievable NGCC capacity factor); with Goal 229 Computation TSD, App. 7 (excluding, inter alia, Martinez Refining units from California’s existing NGCC capacity). In reality, many existing NGCC units face constraints that will prevent them from increasing their utilization to a 70 percent capacity factor. Some units may be located in areas that are designated as in non-attainment for a NAAQS, and as a result would likely have operating permits imposing mass limits on carbon monoxide (“CO”) or NOx emissions that would effectively establish a cap on those units’ operations. Marchetti Report at 26. EPA is currently considering whether to revise the primary NAAQS for ozone to a more stringent level, which could lead to additional restrictions on EGUs operating in certain areas and would further limit states’ ability to increase generation from NGCC units. See EPA, Proposed Rule: National Ambient Air Quality Standards for Ozone (Nov. 25, 2014) (pre-publication version), available at http://www.epa.gov/airquality/ozonepollution/pdfs/20141125proposal.pdf. In addition, some units were financed, designed, and maintained for the specific purpose of operating in cycling duty rather than as base load. Many of these units would not be able to achieve the target utilization rate without significant upgrades and testing to ensure that they are technically capable of operating near full load on a continuous annual basis. In any event, their permitted emission limits may not allow them to operate at a 70 percent capacity factor. For example, the Jackson Power Facility in Jackson, Michigan currently operates under a Title V permit containing enforceable restrictions sufficient to limit operation to well under a 70 percent capacity factor—likely 50 percent or less. Marchetti Report at 23. The Oswald and Fitzhugh NGCC facilities in Arkansas faces similar constraints. Id. at 24. Other units were specifically designed and permitted to provide dedicated backup generation for intermittent sources of renewable energy, such as wind and solar. See id. at 36. 230 For these units, a significant portion of their capacity would need to be preserved for reliability purposes, limiting the capacity available for redispatch to displace coal-fired generation and serve base load. Further, for many existing NGCC units there is simply insufficient pipeline, gas availability, or transmission infrastructure to permit operating year-round at a 70 percent capacity factor. Id. at 19-23. These infrastructure limitations are exacerbated by the fact that the proposed state goals are calculated assuming that states will fully implement Building Block 2 by 2020, giving states and utilities only a few years or less in which to expand gas and transmission infrastructure as needed to reach the target level. As discussed below, such rapid expansion is impossible. EPA does not even acknowledge, let alone adequately address, the constraints preventing existing EGUs from operating at a 70 percent or higher capacity factor. EPA completely ignores potential technical or permit limits on NGCC unit operation and dismisses infrastructure concerns. The Agency’s primary response—that the allowance for “emission rate averaging across multiple units and across time” within a state in the proposed emission guidelines will compensate for such constraints—does not demonstrate that a 70 percent overall capacity factor is achievable. GHG Abatement Measures TSD at 3-15. EPA makes no attempt to determine whether each state will have a sufficient number of existing NGCC units that are able to generate sufficient electricity above a 70 percent capacity factor to make up for units that are not able to approach this level. As a result, the Agency’s reliance on “averaging flexibility” is “mere speculation or conjecture.” See Lignite Energy Council, 198 F.3d at 934. EPA also points to historic trends in an attempt to justify its assumption that sufficient infrastructure already exists to support Building Block 2. For example, the Agency notes that NGCC generation increased by 22 percent between 2011 and 2012, supposedly demonstrating 231 that existing natural gas infrastructure can accommodate rapid increases in existing NGCC utilization. GHG Abatement Measures TSD at 3-9 to 3-10. But EPA fails to note that even after this expansion—which was largely a function of historically low natural gas prices in 2012 due to the widespread development of hydraulic fracturing technology—over half of the existing NGCC units that EPA examined still operated at a capacity factor below 50 percent, presumably due to the kinds of constraints that EPA attempts to wish away. See 2012 NGCC Spreadsheet. EPA also points out that nationwide NGCC generation increased by approximately 81 percent between 2005 and 2012 (compared to the estimated 47 percent increase by 2020 that would be required in order to implement Building Block 2), while failing to acknowledge that this figure includes new NGCC units that were constructed from 2005 to 2012 and does not simply reflect increased utilization of existing units, which is the centerpiece of the Building Block 2 approach. GHG Abatement Measures TSD at 3-11. In addition, EPA relies on trends in hourly capacity factor to claim that a 70 percent annual capacity factor is an achievable goal. According to the Agency, the nationwide NGCC capacity factor during peak hours of the day averages 11 percentage points higher than the overall annual average, suggesting that the current system is able to support national average capacity factors “in the mid to high 50’s for NGCC for peak” (i.e., 11 percentage points higher than the 2012 national average capacity factor of 45.8 percent). 47 Id. at 3-15 to 3-16. EPA does 47 EPA is confused concerning its entire justification for Building Block 2. The Agency cites different values for this analysis in the preamble than in the GHG Abatement Measures TSD. In the preamble, EPA claims that national average capacity factors are “60 percent or higher” during peak hours, and that these values are “15 to 20 percentage points above their average utilization rates.” 79 Fed. Reg. at 34,863. In contrast, EPA states in the GHG Abatement Measures TSD that national average capacity factors are “approximately 50%” during peak hours, and that these values are “approximately 11% above the average capacity factor.” GHG Abatement Measures TSD at 3-15. EPA does not explain this discrepancy, but 232 not explain, however, why it believes it is reasonable to expect that the current system can accommodate an additional 10-15 percentage points in order to reach a 70 percent average capacity factor over all hours of the day. The viability of EPA’s Proposed Guidelines depends on the increased availability of both natural gas and renewable energy. See 79 Fed. Reg. at 34,864; RIA at ES-24, 3-26 (implying that new natural gas generating capacity will increase 20 to 22 percent by 2020 to replace retiring coal-fired capacity under Option 1, and stating that new natural gas generating capacity will increase 9 percent relative to the base case in 2020 under Option 2). Most significant new renewable capacity will be constructed far from load centers because of the unique siting needs and sensitivities of large-scale wind and solar projects. Ramping up both renewable energy and natural gas will require significant infrastructure improvements, most notably, more transmission lines and natural gas pipelines. Initial NERC Report at 2. Both of these forms of infrastructure require many layers of permitting and years of regulatory approvals, often at federal, state, and local levels. EPA is unrealistic in assuming that this infrastructure will be in place by 2020 for states to begin implementing all four Building Blocks. A large number of permits, consultations, and approvals are needed from multiple government bodies to get a new transmission line permitted and constructed. The timeline for a transmission project depends on completion of planning studies and technical analysis, real estate availability (negotiating rights-of-way or exercising eminent domain authority), procurement of long lead time equipment, environmental permitting requirements, public opposition, regulatory approval, and opportunities for tie in outages. A relatively simple project the values in the TSD appear to match the EIA sources EPA cites in both the preamble and the TSD, so this analysis assumes that the values in the TSD govern. 233 that will not traverse an environmentally sensitive area, require the exercise of eminent domain, or involve significant public opposition will take up to three years prior to construction. More complicated projects that will traverse federal lands, environmentally sensitive areas, or will generate public opposition may take as much as 10 years to complete. Among the many permits that may be required for a new transmission line are the following: an Environmental Assessment or Environmental Impact Statement (required under both federal and state law in some cases), if the project involves significant state or federal government action of any kind; a Section 404 permit from the Army Corps of Engineers if dredge or fill material will be placed in “waters of the United States”; Section 7 consultation with the U.S. Fish and Wildlife Service under the Endangered Species Act if the project has the potential to impact threatened or endangered species; a Special Use Authorization under the National Forest Management Act if the project will traverse federal lands managed by the U.S. Forest Service; a right-of-way grant under the Federal Land Policy and Management Act from the U.S. Department of Interior Bureau of Land Management (“BLM”) if the project traverses federal lands managed by BLM; a state water quality permit (if required by a state water quality statute); fish, game, and other wildlife related permits, if the project will divert natural flow of water bodies or otherwise affect fish and game; Section 106 National Historic Preservation Act consultation if the project might impact cultural or historic resources; a right-of-way lease agreement; and an air quality permit if disturbed acreage will exceed certain thresholds. See California Public Utilities Commission, Federal, State, and Local Permitting Processes Likely to be Required for Electric Transmission Projects (June 2009), http://www.cpuc.ca.gov/NR/ rdonlyres/D896C1EA-BD35-4BC8-83C8-D332BAE959BF/0/GenericTransmissionLine 234 Permit.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). The Utah Office of Energy Development has estimated that a typical schedule for siting, permitting, and constructing a new transmission line in that state could take up to seven years, assuming the line is at least 100 miles long, crosses federal lands, and requires a federal environmental impact statement. See Utah Office of Energy Development, Guide to Permitting Electric Transmission Lines in Utah, at iii, Fig. ES-1 (Aug. 2013), available at http://energy. utah.gov/download/reports/PermittingGuide_Final%20080913%20(cd%20ver).pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). In Colorado, siting transmission lines can take even longer because local governments (counties and municipalities) have jurisdiction over the siting of transmission lines and each may impose different permit or other authorization requirements and conditions. As just one example of the time that may be required to carry out a transmission line project, in Arizona—a state that is likely to require significant amounts of new transmission infrastructure, given that the state’s proposed goal assumes that all coal-fired generation is eliminated by 2020—the Morgan to Sun Valley Transmission Line Project is expected to take 8 to 9 years from its conception in 2007 to final approval, before the design phase even begins. See APS, “Current siting projects,” available at https://www.aps.com/en/ourcompany/generation transmission/transmissionfacilitysiting/Pages/current-siting-projects.aspx. The project will extend 38 miles and hold a 500 kilovolt transmission line and a 230 kilovolt line, crossing roughly 9 miles of federal public lands. The State of Arizona has diligently carried out its functions in the regulatory process: the process simply involves numerous steps, such as siting and public participation, that require a significant amount of time to complete. In addition, 235 because the project will cross federal land (a common requirement for Western projects), it was required to undergo analysis under the National Environmental Policy Act (“NEPA”), which took 4 years and culminated in an environmental impact statement analyzing the potential impacts of the entire project and a record of decision approving the project utility’s corridors and necessary right-of-way. See BLM, “APS Sun Valley to Morgan 500/230kV Transmission Line Project Record of Decision,” (Jan. 2014), available at http://www.blm.gov/style/medialib/blm/ az/pdfs/energy/aps.Par.38965.File.dat/ROD.pdf. The project’s owner anticipates that the project will receive final approval and a notice to proceed from the state in 1 to 2 years. Design and construction of the project will not begin until it receives a notice to proceed. Moreover, although significant gains have been made in the past several years in locating new sources of natural gas, little has been done to improve the infrastructure needed to connect this new gas to power plants. The Interstate Natural Gas Association of America (“INGAA”) predicts that the natural gas industry will need to construct 2,000 miles per year for the next twenty years of new gas transmission mainlines and laterals to meet demand. See The INGAA Foundation, Inc., Jobs & Economic Benefits of Midstream Infrastructure Development: US Economic Impacts Through 2035, at ES-1 (Feb. 15, 2012), available at http://www.ingaa.org/ File.aspx?id=17744 (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record); see also The INGAA Foundation, Inc., North American Midstream Infrastructure through 2035: Capitalizing on Our Energy Abundance, at 38 (Mar. 18, 2014), available at http://www.ingaa.org/Foundation/Foundation-Reports/2035Report.aspx (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). INGAA and other industry groups also stated recently that more than 338,000 miles of pipeline will need to constructed between 2014 and 2035. See Comments of American Petroleum Institute, 236 Association of Oil Pipe Lines, International Association of Geophysical Contractors, Interstate Natural Gas Association of America, Utility Air Regulatory Group, and Utility Water Act Group on Three Endangered Species Act Critical Habitat Proposals of the U.S. Fish and Wildlife Service and the National Marine Fisheries Service Published on May 12, 2014, at 5 (Oct. 9, 2014), available at http://www.ingaa.org/File.aspx?id=22680 (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). For perspective, there were about 306,000 miles of gas pipeline in use in the United States at the end of 2008. See EIA, “Estimated Natural Gas Pipeline Mileage in the Lower 48 States, Close of 2008,” available at http://www.eia.gov/pub/oil_gas/natural_gas/analysis_publications/ngpipeline/ mileage.html. EPA has not adequately addressed in the record these significant barriers that exist to ramping up the use of natural gas and renewables in the electricity generating sector by 2020 and 2030. In order to meet EPA’s interim goals between 2020 and 2030, this capacity realistically needs to be in place by 2020. EPA has not demonstrated how the sector will be able to ramp up natural gas and renewable capacity in time to meet the deadlines in the Proposed Guidelines given the many years needed to plan, permit, and build the infrastructure to support these forms of generation. This lack of demonstration and record support is an additional factor rendering the rule arbitrary and capricious. The Agency suggests that pipeline infrastructure will have sufficient time to expand if necessary to accommodate Building Block 2 because the Proposed Guidelines “provide[] for flexible implementation that will permit efficient scheduling of infrastructure upgrades as needed.” GHG Abatement Measures TSD at 3-16. Presumably, EPA is suggesting that states may delay full implementation of Building Block 2 because compliance with the proposed state interim goals is based on average emissions from 2020 to 2029. The “flexible implementation” 237 timeline EPA claims to have provided states is largely illusory, however, particularly with respect to Building Block 2. For most states, redispatch under Building Block 2 accounts for the most significant emission reductions of any building block, and EPA calculated the proposed state goals with the assumption that Building Blocks 1 and 2 would be fully implemented by 2020. 79 Fed. Reg. at 34,905-06; see Goal Computation TSD at 18 (noting that in calculating state goals, only “the RE and EE assumptions change for each year from 2020 through 2029”). Delaying full implementation of Building Block 2 in order to allow for necessary infrastructure development would significantly increase a state’s emissions in early years of the 2020 to 2029 compliance period, forcing the state to achieve more drastic emission reductions in later years in order to comply with the interim goal on average. Thus, most states will need to implement necessary infrastructure improvements to accommodate Building Block 2 by 2020 or very soon thereafter. This timeline is likely infeasible, given that gas pipeline projects often require a decade to complete, in part because the timeline for environmental permitting can be lengthy. In addition, by focusing its achievability analysis for Building Block 2 on annual capacity factors, EPA has failed to consider the effect of widespread redispatch to NGCC on states’ ability to meet demand during periods of peak load. During peak demand, many utilities are already operating all of their available generating capacity to the fullest extent, including NGCC and coal-fired units, while also purchasing power from merchant generators. Marchetti Report at 24-26. But in several states, the limits on coal utilization that follow from redispatch to existing NGCC will eliminate or severely limit coal-fired units as an option for meeting peak demand. In Arizona and other states, the interim and final goals are based on eliminating coal-fired generation from the state’s generation mix completely. Without access to this coal-fired capacity, utilities would be forced to purchase substantial additional amounts of electricity from 238 NGCC merchant generators on the short-term market. Id. at 25. Yet many of these merchant generators already operate at or near full capacity during periods of peak demand, and in any event, their capacity may be committed to serve load in neighboring states under other long- or short-term agreements. Id. Therefore, utilities cannot rely on merchant generation to meet longterm demand requirements during peak summer months in lieu of generation from existing coalfired units. Id. at 25-26. EPA’s Building Block 2 is also based on the faulty premise that generation from any NGCC unit located within a state may be redirected by a utility or state within the state’s borders to offset reduced generation from coal-fired units elsewhere in the state. In many cases, this is simply not true. In most states, electric generation is dispatched by an RTO or ISO, which calls up generation from EGUs within its system such that the lowest cost generating sources are dispatched first. See id. at 16-19. Thus, as an initial matter, neither a utility nor any state agency controls the relative dispatch of coal-fired EGUs and NGCC units in most states. In addition, the areas covered by RTO and ISO systems typically do not coincide with state borders. In most cases, an ISO’s footprint includes generating units in several states, and some states are divided between multiple ISOs. As a result, it is common for EGUs within a state to be dispatched based on direction from different managing authorities, and for EGUs to serve load outside of the states in which they are located. There may be limited interconnection between the different dispatch systems in a state, and EGUs in one part of a state may lack the cross-state transmission rights needed to serve load in a different part of the state. One straightforward example may be found in South Dakota, which contains only one coal-fired unit (Big Stone, operating in the MISO) and one NGCC unit (Deer Creek, soon to join the SPP). Id. at 18-19. Because these units operate in different ISOs, they do not share a 239 common dispatch signal, and reduced generation from Big Stone will not necessarily be replaced by increasing generation from Deer Creek. In addition, each of these units was designed, sited, and permitted to serve the needs of their utilities’ respective customer bases, not those of other utilities. Thus, even though these units are located in the same state, each utility owns firm transmission rights from its unit to its own retail load—and not to the other utility’s customers. Id. at 19. The issue becomes even more complicated in states with a more complex generation mix—for example, Texas contains at least three separate RTOs (the ERCOT, SPP, and MISO), with the majority of its NGCC generating capacity concentrated in just one (the ERCOT). See FERC, Market Oversight, Regional Transmission Organizations, http://www.ferc.gov/marketoversight/mktelectric/overview/elec-ovr-rto-map.pdf. Even where electricity dispatch is governed by individual utilities and not by an ISO or RTO, similar limitations on in-state redispatch arise where a state is served by multiple utilities or where a utility spans multiple states. For example, TVA generates power from fossil fuels in 4 states and distributes power to Tennessee and portions of 6 other states. It also buys and sells some power from neighboring systems and optimizes the dispatch of generating assets and operation of transmission infrastructure as a system (not necessarily within individual states) to maintain reliability and minimize cost. Three of TVA’s five NGCC units are located in Mississippi, but all of its coal-fired units are located outside of Mississippi (TVA does purchase power through a power purchase agreement from the Red Hills coal-fired unit in Mississippi). Increasing the capacity factor of TVA’s Mississippi NGCC plants would have minimal effect on coal-fired generation in Mississippi: virtually all of the generation displaced would be from TVA’s coal-fired units in other TVA states. 240 In sum, EPA has failed to demonstrate that its Building Block 2 target of redispatching generation from existing coal- and oil/gas-fired steam EGUs to existing NGCC units up to an overall NGCC capacity factor of 70 percent is achievable. The Agency failed to assess whether the subset of NGCC units currently operating at 70 percent capacity factor represents the remainder of existing NGCC units, and it did not properly address economic, technical, regulatory, or infrastructure constraints preventing some units from operating at the target level. Above all else, Building Block 2 demonstrates that EPA fundamentally miscomprehends how energy is generated and distributed in the United States. C. The Renewable Energy Component of Building Block 3 Is Not Achievable. EPA proposes to find that it is achievable for states to increase in-state generation from renewable energy sources to an amount that is based on application of a regional annual growth factor to the state’s 2012 renewable energy generation, gradually approaching a maximum regional renewable energy generation target. 79 Fed. Reg. at 34,867. EPA expects states to begin taking measures to increase renewable energy generation in 2017, three years before the beginning of the compliance period. Id. The regional maximum renewable energy target for each state is based on the average 2020 primary RPS goal of states within each region that have RPS goals. Id. The renewable energy generation goals that EPA has calculated for each state are unachievable. EPA’s methodology for determining these goals is fatally flawed because it arbitrarily ties each state’s goal to the RPS goals of other states, which often do not require as much renewable energy generation as they appear to and reflect each state’s unique mix of renewable energy potential. Marchetti Report at 31-34. For example, EPA relied on state RPS goals that are based on capacity rather than generation, despite claiming that it excluded such goals. See GHG Abatement Measures TSD at 4-9 n.108; Marchetti Report at 32 (noting Kansas 241 RPS is capacity-based). EPA’s renewable energy generation targets for some states are unreasonably aggressive and do not take into account factors affecting the actual renewable energy growth potential in each state. Even EPA’s IPM modeling results indicate that predicted renewable energy generation by the end of the compliance period will not come close to the renewable energy targets EPA used in its goal calculation. Accordingly, EPA must withdraw the Proposed Guidelines and reduce its renewable energy generation targets to more realistic levels. 1. EPA’s Method for Calculating the Renewable Energy Targets for Building Block 3 Is Flawed. EPA’s approach to determining regional renewable energy generation targets is flawed in several respects. First, it is arbitrary for EPA to assume that each state is capable of achieving similar levels of renewable energy generation as its neighboring states. The Agency acknowledges that state RPS goals reflect “policy objectives including both feasibility and costs” and not merely renewable energy potential. 79 Fed. Reg. at 34,866. These policy objectives may vary wildly from state to state within individual regions. Moreover, renewable energy potential is highly state-specific and depends on each state’s natural resources (e.g. insolation, availability of wind energy, biomass resources, etc.), existing transmission infrastructure, and current renewable energy development. The proposed “regional” approach is simply based on NERC regions and does not reflect similarities between states based on renewable energy potential. EPA’s failure to account for the precise nature of existing state RPS goals leads to unreasonable results. For example, the renewable energy target for the South Central region 48 is based entirely on Kansas’s 2020 RPS target of 20 percent renewable energy capacity. See 48 The South Central region includes the states of Arkansas, Kansas, Louisiana, Nebraska, Oklahoma, and Texas. GHG Abatement Measures TSD at 4-15. 242 generally Kansas Renewable Energy Standards Act, KAN. STAT. ANN. §§ 66-1256 through 661262. Because actual generation is significantly less than capacity, a 20 percent target based on capacity would not come close to approximating a 20 percent generation-based renewable energy target. Moreover, in adopting its RPS, Kansas was building upon its experience with 24.2 percent average annual growth in wind from 2008 to 2012. See Marchetti Report at 32. No other states in the South Central region have a generation-based RPS; the Texas RPS is also capacity-based. Thus, EPA erroneously imputed a renewable energy generation-based target to the entire region that does not exist in any state in the South Central region. EPA also failed to account for factors limiting certain renewable generation in particular states—factors that would prevent certain states from meeting their goals. In Arkansas, industrial combined heat and power (“CHP”) facilities burning wood or wood-derived fuels made up over 95 percent of the state’s renewable energy generation and only grew at a 1.6 percent average annual rate from 2008 to 2012. Id. Arkansas cannot reasonably quadruple this growth rate, as it would be required to do to meet its renewable energy target. Similarly, Kentucky has very low wind and solar generation potential. See EIA, Kentucky Profile Analysis (Dec. 18, 2013), available at http://www.eia.gov/state/analysis.cfm?sid=KY. Likewise, the renewable energy target for the entire Southeast region 49 is based solely on North Carolina’s 2020 RPS goal of 10 percent and assumes a renewable energy growth rate of 13 percent per year for each state. But in Georgia and Alabama, industrial CHP sources burning wood or wood-derived products (such as pulp and paper mills) constitute over 90 percent of each state’s renewable energy generation. Marchetti Report at 34. Between 2008 and 2012, these 49 The Southeast region includes Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, and Tennessee. GHG Abatement Measures TSD at 4-15. 243 industrial CHP sources grew at an annual rate of only 2.1 percent in Georgia and actually declined by 5.1 percent annually in Alabama. These sources have little potential for growth, so in order for Georgia to achieve its state renewable energy target, other renewable energy sources would need to increase by over 44 percent per year. Accordingly, EPA’s approach is inherently defective because it does not account for these state-specific limits on achievable renewable energy potential. Second, EPA’s analysis of state RPS goals is oversimplified and reflects EPA’s limited understanding of these goals. Some state RPS goals are voluntary, while others establish requirements only for large, investor-owned electric utilities or apply less stringent “secondary” or “tertiary” RPS goals to smaller utilities, municipal utilities, or cooperative utilities. See GHG Abatement Measures TSD at 4-10. But EPA applies the more stringent primary RPS goals to all in-state generation and treats all goals as mandatory, artificially leading to more aggressive renewable energy generation targets. Many state RPS goals include hydroelectric power generation, but EPA does not include existing hydroelectric power towards compliance with its Proposed Guidelines. Marchetti Report at 33. For example, New York counts hydroelectric power towards its 29 percent RPS target, and this source accounted for 83 percent of the state’s renewable generation in 2012. Id. Including this component of the RPS in the Building Block 3 goals without allowing states to rely on existing hydropower for compliance will substantially increase the amount of new renewables required in each state, even in states that have already reached their own RPS goals. Moreover, some state RPS goals allow utilities to comply through mechanisms other than actually developing in-state renewable energy generation. For example, as described in the Marchetti Report at 31-32, North Carolina’s RPS goal allows utilities to meet 25 percent of their 244 requirements by reducing energy consumption through energy efficiency measures, and another 25 percent of their requirements by purchasing renewable energy certificates from out-of-state facilities. Thus, the 10 percent RPS target EPA assumed for North Carolina effectively requires in-state renewable energy generation to reach only 5 percent by 2020. Because North Carolina’s RPS alone sets the regional target for the entire Southeast region, EPA’s target is roughly 100 percent too high for eight states. This is not a minor error. North Carolina and other states have built cost-control mechanisms into their RPS goals in order to reduce the requirements for utilities if compliance becomes too burdensome. Id. EPA did not account for any of these flexible compliance mechanisms when dictating regional renewable energy generation targets. Third, EPA’s methodology is flawed because it requires states to begin increasing renewable energy generation by 2017, three years before the compliance period begins for the Proposed Guidelines. To put this timeline in perspective, initial state plan submissions are not due to EPA until June 30, 2016, and EPA is allowing itself until June 30, 2017, to take final action approving or disapproving these plans. 79 Fed. Reg. at 34,915-16. Thus, EPA’s proposed rule would require states to begin implementing their plans before they have even received the Agency’s approval. The disconnect is even greater in light of the fact that states may delay plan submission until June 30, 2017 (for various reasons), or June 30, 2018 (for multi-state plans), pushing final EPA approval of these plans to mid-2018 or -2019. Id. at 34,915. In addition, building new renewable energy generation capacity requires several years to allow for planning, obtaining funding, seeking regulatory approval, and construction. Increasing Renewable Energy capacity will also require new transmission infrastructure, which can require 8-10 years to complete, as described in Section XIV.B above. In effect, states would have needed to begin efforts to implement Building Block 3 several years ago. This is plainly unreasonable. 245 Due to these issues, EPA’s proposed renewable energy generation goals are drastically higher than the states can achieve within EPA’s proposed timeline. Comparing present-day estimates of the expected baseline renewable energy growth to the renewable energy totals EPA applies in Building Block 3 reveals just how unreasonable, arbitrary, and capricious the Agency’s estimates of potential renewable energy generation are. The Marchetti Report compares estimated renewable energy generation under Building Block 3 assumptions in the EIA’s Annual Energy Outlook (AEO) 2013 baseline analysis, and in EPA’s IPM runs projecting the results of state-focused implementation of the Proposed Guidelines (State Option 1). Marchetti Report at 28-30, Tbl. 13 & Fig. 2. Under Building Block 3, EPA proposes to find that states can increase nationwide renewable energy generation from 217,868 GWh in 2012 to 522,723 GWh in 2029, with an annual growth rate from 2020 to 2029 of roughly 7.1 percent. Id. at 28. In stark contrast, the EIA projects that under the status quo, nationwide renewable energy generation will reach only 385,433 GWh in 2029, with an annual growth rate of only 1.5 percent from 2020 to 2029. Id. at 29, Tbl. 13. EPA’s own “Base Case” modeling predicts even less renewable energy growth under status quo conditions, with only about 350,000 GWh in 2029 and an annual growth rate of 1.7 percent. RIA at 3-27. Although states may have some available measures at their disposal that could lead to some incremental increase in renewable energy generation from the status quo over this timeframe, expecting the nation as a whole to nearly quintuple its annual growth rate for renewables and increase its expected renewable energy generation by more than one third over projected levels is simply unreasonable. This same pattern holds true for individual states. The Marchetti Report also contrasts EPA’s estimated renewable energy generation under Building Block 3 for Georgia and Alabama 246 with a comparable baseline regional estimate from the EIA’s AEO 2013. 50 Marchetti Report at 30-31 & Tbl. 14, Fig. 3. Under Building Block 3, EPA expects Georgia and Alabama together to increase renewable energy generation from 6,055 GWh in 2012 to 26,523 GWh in 2029, for an annual growth factor from 2020 to 2029 of 11.4 percent. But EIA’s baseline projection forecasts renewable energy generation of only 18,700 GWh in 2029, for an annual growth factor of 2.0 percent. EPA cannot establish that increases of this magnitude are achievable— particularly because, as discussed in the next section, the Agency has failed to provide any parsed analysis of its IPM model runs for 2030 that would allow states to assess the impact of these required increases. Indeed, EPA has provided data for only 4 of the 25 IPM model runs that it performed as support for the Proposed Guidelines. Finally, in addition to the regional RPS-based approach to developing Building Block 3 goals that EPA used to calculate and justify its Proposed Guidelines, EPA also solicits comment on an “alternative approach” to quantifying achievable renewable energy generation. 79 Fed. Reg. at 34,869-70; Alternative RE Approach Technical Support Document, Docket ID No. EPAHQ-OAR-2013-0602-0458 (“Alternative RE Approach TSD”). EPA has not applied this alternative renewable energy approach to calculate an alternative set of proposed state CO2 emission goals, and therefore it cannot promulgate final state CO2 emission goals based on this alternative renewable energy approach without first issuing a supplemental proposal publishing such alternative state goals for comment. UARG reserves the right to comment more extensively on the alternative renewable energy approach in any supplemental proposal that EPA would publish. 50 The AEO 2013 SERC Reliability Corporation – Southeastern includes most of Alabama and Georgia, plus a small section of eastern Mississippi and the Florida panhandle. 247 UARG notes, however, that the alternative approach to quantifying state renewable energy potential is also arbitrary and unreasonable. The alternative approach would set each state’s renewable energy generation target at the lesser of two calculated values. The first calculation is based on a National Renewable Energy Laboratory (“NREL”) GIS-based study estimating state renewable energy technical potential by technology type (solar, wind, etc.). EPA would: (1) identify each state’s technology-specific renewable energy potential for each technology; (2) determine each state’s current “development rate” for each type of technology based on 2012 generation data; (3) establish a “benchmark rate” of development for each technology type based on the top 16 states; (4) apply this benchmark rate uniformly to states for each technology type to develop target generation levels for each technology; and finally (5) add the state’s target generation level for each technology type to determine the state’s overall renewable energy generation target. Alternative RE Approach TSD at 1-2. The second calculation is based on the results of IPM runs that assume the cost of new renewable builds of each renewable energy technology type is reduced by “the avoided cost of other actions that could be taken instead to reduce power sector CO2.” Id. at 2. Neither calculation provides a reasonable method to determine achievable renewable energy generation targets for states. The approach that relies on the NREL GIS-based study would arbitrarily apply a set of uniform benchmark development rates to all states across the country without considering whether the states used to set the benchmark rates for each technology type are representative of the rest of the country. EPA would not consider factors that may lead some states to develop specific renewable energy technologies more than others. In addition, this approach adds estimated renewable energy potential across various technology types, even though the NREL study warns (and EPA acknowledges) that these estimates cannot 248 be added together because they may double count available land for use by multiple renewable energy technologies. See id. at 2 (estimates “may overstate electricity production potential because a given site cannot produce RE simultaneously from multiple technology types”); Anthony Lopez, et al., NREL, U.S. Renewable Energy Technical Potentials: A GIS-Based Analysis, at 2 (July 2012), available at http://www.nrel.gov/gis/re_potential.html (“the same land area may be the basis for estimates of multiple technologies”). Likewise, the IPM-based approach is arbitrary because EPA does not provide any justification for its assumptions regarding reduced costs for new renewable energy generation. 2. EPA Has Not Provided All of the Data Needed To Evaluate the Proposed Guidelines in Violation of Section 307(d)(3) of the CAA. EPA must provide the basis and purpose for the Proposed Guidelines under section 307(d)(3) of the CAA. The basis and purpose must include “a summary of . . . the factual data on which the proposed rule is based; . . . the methodology used in obtaining the data and in analyzing the data; and . . . the major legal interpretations and policy considerations underlying the proposed rule.” CAA § 307(d)(3). EPA’s failure to include the data for 21 of its modeling runs has affected UARG’s ability to comment meaningfully on the Proposed Guidelines. In response to a request from UARG under the Freedom of Information Act, EPA responded that it had not parsed the 21 modeling runs, which is why those data are not included in the docket. Letter from Karen Orehowsky, Deputy Dir., Clean Air Markets Div., EPA to Craig S. Harrison, Hunton & Williams LLP (Aug. 12, 2014) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Without such parsed data, however, EPA cannot have conducted a reasonable inquiry into the achievability of Building Block 3. This is an egregious failure to comply with section 307(d)(3), and will render any final rule unlawful. EPA is “play[ing] hunt the peanut with technical information, hiding or disguising the information that it 249 employs,” and is “treat[ing] what should be a genuine interchange as mere bureaucratic sport.” Conn. Light & Power Co. v. Nuclear Regulatory Comm’n, 673 F.2d 525, 530 (D.C. Cir. 1982). The four parsed IPM runs that are available demonstrate that the Agency’s Building Block 3 targets are far too costly and do not represent realistic estimates of the potential renewable energy generation available to achieve the Proposed Guidelines. EPA performed IPM modeling runs to assess how states would comply with the Proposed Guidelines under both statefocused and region-focused compliance approaches. Thus, the results of these IPM runs constitute EPA’s estimate of representative scenarios in which the power sector is operating in compliance with the Proposed Guidelines at the least cost possible, within the constraints EPA imposes on the model. IPM model runs predict that actual renewable energy generation growth will meet only a small fraction of EPA’s target under the least-cost approach to compliance with the Proposed Guidelines. IPM predicts that under both the state-focused and region-focused compliance approaches, renewable energy generation will reach only 356,063 GWh by 2029, and will grow at an annual rate of 1.1 percent from 2020 to 2029, far short of EPA’s target of 522,723 GWh by 2029. Marchetti Report at 29, Tbl. 13. This level of generation is only 1.7 percent greater than the renewable energy generation that EPA predicts in 2029 under its own Base Case analysis, and is actually 7.6 percent lower than the EIA predicts will occur in 2029 in the absence of any emission guidelines. See id.; RIA at 3-27. EPA’s own analysis severely undermines the Agency’s prediction of the level of increased renewable energy generation that is achievable, and demonstrates that the Agency has failed to accurately consider factors such as cost and feasibility. The IPM results suggest that rapid growth in renewable energy generation is so costly that states are able to achieve only negligible incremental increases in renewable energy 250 generation above the status quo (at best), and they must instead rely more heavily on other components of EPA’s selected BSER in order to comply with the Proposed Guidelines. This, in turn, will increase the cost of those other Building Blocks in ways that EPA has failed to analyze in the Proposed Guidelines. EPA must withdraw the Proposed Guidelines because they impose arbitrary, capricious, and unachievable renewable energy generation targets. D. EPA Correctly Excluded Natural Gas Conversion and Co-Firing From Its Proposed BSER Determination. EPA states that it “does not propose to consider [natural gas conversion or co-firing at coal-fired utility boilers] part of the best system of emission reduction adequately demonstrated for existing EGUs.” 79 Fed. Reg. at 34,875. This conclusion is due largely to the high costs of implementing such a conversion. However, EPA “solicit[s] comment on whether natural gas cofiring or conversion should be part of the BSER.” Id. at 34,876. UARG agrees that natural gas conversion and co-firing are not BSER for coal-fired EGUs and should not be included in the Proposed Guidelines. Although sources should have the option to voluntarily use these measures to comply with CO2 emission standards, natural gas conversion and co-firing are extremely costly and are appropriate only for certain EGUs based on site-specific factors. See Lowell L. Smith, “Overview of Natural Gas Conversion Options To Meet Emission Requirements on Coal-Fired Electric Generating Units” (Nov. 2014) (“Smith Report”) (Attachment H). XV. Issues Raised by EPA’s October 30, 2014 Notice of Data Availability Late in this rulemaking process EPA issued a “notice of data availability” (“NODA”), 79 Fed. Reg. 64,543-53 (Oct. 30, 2014). For the most part, NODA is a misnomer; it contains 251 virtually no discussion of any relevant new “data,” 51 but instead is described as a means to “provid[e] additional information on several topics raised by stakeholders.” Id. at 64,543. The NODA identifies three topics that the Agency says have been raised either in public or private meetings with certain unnamed stakeholders: (1) “the emission reduction compliance trajectories created by the interim goal for 2020 to 2029”; (2) “certain aspects of the building block methodology”; and (3) “the way state-specific [CO2] goals are calculated.” Id. Although some of the stakeholders’ suggestions regarding EPA’s Proposed Guidelines might have profound effects on the requirements that states and regulated entities might have to meet in any final guidelines, the Agency has not budged on its December 1, 2014 deadline for comments on the Proposed Guidelines or on the NODA itself. Id. In any event, the NODA contains no EPA proposals or proposed regulatory language but rather is musings about different policies that the Agency might prefer. These seem more appropriate for an advance notice of proposed rulemaking rather than a NODA on the eve of a comment deadline. Most of the suggestions that are discussed in the NODA cannot be pursued without EPA undertaking additional analyses after which it could re-propose the rule for comment. See CAA § 307(d)(3). A. General Issues As UARG stated in a November 13, 2014 letter to the EPA Administrator, EPA should provide the public with more time to comment on the issues raised in the NODA. 52 Letter from Allison D. Wood, Hunton & Williams LLP, to Regina A. McCarthy, Adm’r, EPA at 2 (Nov. 13, 2014). Although the Agency has not proposed any specific changes to its original June 18, 2014 51 By way of contrast, EPA released valuable data in its Translation TSD, which it noticed in 79 Fed. Reg. 67,406 (Nov. 13, 2014). 52 EPA rejected this request. Letter from Janet McCabe, Acting Assistant Adm’r, EPA to Allison D. Wood, Hunton & Williams LLP (Nov. 18, 2014) (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). 252 Proposed Guidelines, UARG members need additional time to prepare fully informed and wellconsidered comments on the issues that have been raised for the first time in the NODA. The October 30, 2014 Federal Register notice was published barely a month before the due date for comments. The Proposed Guidelines are extraordinarily complex and, as discussed below, the issues raised in the NODA on which EPA solicits comment are very broad, seem to contain some approaches that are either new or that contradict its previous positions, and ultimately obscure to the point of incoherence what approach EPA might ultimately take. Obscuring what action an agency might take late in a comment period does not provide adequate notice to the public, is not reasoned decisionmaking, and provides no grounds for EPA to assert that its unsupported policy musings are a logical outgrowth of the original proposal. EPA acknowledges that the various alternative approaches the Agency is now contemplating in the NODA could fundamentally alter the proposed emission guidelines and lead to more stringent goals for all or some states. 79 Fed. Reg. at 64,544. EPA also recognizes that many issues in the NODA—and the four Building Blocks generally—“interact with each other.” Id. Realistically there is no way to assess how stringent any state’s goals will actually be until EPA provides greater certainty on its preferred approach. Because of this fundamental ambiguity, EPA cannot finalize any of the changes discussed in this NODA without issuing a supplemental proposal that clarifies the Agency’s position, provides the basis and purpose for its approach pursuant to CAA section 307(d)(3), provides specific details of how each state would be affected, provides sufficient economic analysis of the changes, and provides proposed regulatory text. None of these things are present in the NODA. UARG is pleased that EPA appears to recognize some of the legal flaws and irrational consequences inherent in the Proposed Guidelines. For example, EPA now appears to recognize 253 that the Proposed Guidelines do not consider the remaining useful lives of existing EGUs, despite the explicit statutory command to do so in section 111(d)(1)(B) of the CAA. Id.; see also id. at 64,549. The Agency has discovered that if finalized the Proposed Guidelines would create an enormous stranded asset problem with the required shutdown of many coal-fired EGUs that have recently been constructed or have recently received significant capital investment to comply with other EPA regulations, such as the MATS Rule and measures to reduce emissions of visibility-impairing pollutants. Id. at 64,546. EPA now concedes that its requirement for redispatch in Building Block 2 will in many situations distort regional electricity markets. Id. at 64,549. Finally, the Agency now seems to acknowledge that the Proposed Guidelines render illusory its oft-stated purpose of providing states with flexibility with respect to time and emission reduction strategies. Id. at 64,544. For all of these reasons and for the reasons stated elsewhere in these comments, EPA should withdraw the Proposed Guidelines, resolve the many problems that it has finally recognized, and propose a fundamentally different approach. EPA’s scattershot efforts to date fail to satisfy the basic elements of notice and comment rulemaking. See CAA § 307(d)(3); 5 U.S.C. § 553. B. Building Block 2 Issues: New NGCC and Co-Firing as Components of BSER The NODA solicits comment “on ways that building block 2 could be expanded to include new NGCC units and natural gas co-firing in existing coal-fired boilers.” 79 Fed. Reg. at 64,545. EPA’s discussion suggests that BSER should be based on “the cost and feasibility of the total amount of natural gas used, as opposed to the extent to which the gas is used for particular types of generation (i.e., existing NGCC generation, new NGCC generation, or co-firing).” Id. at 64,549. EPA claims that requiring new NGCC and co-firing as part of the BSER would be “more consistent with historic NGCC deployment, better reflect growing geographic availability of natural gas supply, contribute to expanded generation fuel diversity in states that currently 254 have relatively little NGCC capacity, and offer more cost-effective emission reductions.” Id. Under this new approach, raised for the first time in the NODA but not formally proposed, in addition to shifting generation from existing coal-fired units to existing NGCC units until the NGCC units reach a 70% capacity factor, Building Block 2 would also assume “some minimum value as a floor for the amount of generation shift for purposes of building block 2, whether that shift takes the form of redispatch from steam generation to existing NGCC units, re-dispatch to new NGCC units, or co-firing natural gas in existing coal-fired boilers.” Id. at 64,550. The minimum floor for redispatch away from coal-fired units would be expressed as a percentage of the state’s fossil steam generation. EPA specifically solicits comment on a minimum value of 12 percent (the shift observed by the lower quartile of states with some redispatch under the current Proposed Guidelines), but requests comment on a vast range of possibilities between 0 and 55 percent. Id. This approach is both unlawful and unreasonable. As discussed in Section X, EPA has no authority to impose federally enforceable obligations on new sources (such as new NGCC units) under section 111(d). Nor does EPA have authority under section 111(d) to require the construction of any new source. By its explicit terms, section 111(d) (entitled “Standards of performance for existing sources; remaining useful life of source”) applies only to existing sources, which the CAA defines to be mutually exclusive of new sources. A state plan “establishes standards for performance for any existing source . . . .” CAA § 111(d)(1)(A) (emphasis added). EPA erroneously suggests that constructing a new NGCC unit can be considered to be a “system of emission reduction” under section 111. See 79 Fed. Reg. at 64,550. As discussed in Section III, a “system of emission reduction” is limited to activities that can be implemented at the source itself. The approach EPA discusses would require owners to 255 go beyond the source itself and construct an entirely new source to replace it. EPA’s glib claim that a source owner “may invest in new NGCC units,” id., does not mean that replacing an affected EGU’s generation with generation from a new facility can be required under CAA section 111. Moreover, EPA’s suggestion that natural gas co-firing at existing boilers can be designated as BSER, id., is wrong both as a matter of fact and of rulemaking procedure. Cofiring has not been adequately demonstrated for the source category as a whole, and EPA has failed to provide any justification for any such determination as required by section 307(d)(3) of the CAA (EPA must provide the basis and purpose for the proposed guidelines). The fact that owners or operators of certain individual units have determined co-firing to be feasible and economic does not imply by any stretch of the imagination that co-firing is an adequately demonstrated system of emission reduction. EPA for decades has implemented section 111 such that any NSPS “establishes what every source can achieve, not the best that a source could do.” Letter from Gary McCutchen, Chief, New Source Review Section, EPA OAQPS, to Richard E. Grusnick, Chief, Air Div., Ala. Dep’t of Envtl. Mgmt. at 1 (July 28, 1987), available at http://www.epa.gov/region7/air/nsr/nsrmemos/crucial.pdf (included in UARG’s separately filed Supplemental Materials for the Rulemaking Record). Natural gas co-firing is far too expensive to qualify as BSER (or as a component of the combined BSER EPA is proposing) for coal-fired EGUs. EPA recognized this in the Proposed Guidelines when it specifically rejected co-firing and conversion as BSER. The Agency noted that that: [T]here are more cost effective opportunities for coal-fired utility boilers to reduce their CO2 emissions than through natural gas conversion or co-firing. As a result, the EPA has not proposed at this time to include this option in the BSER and has not incorporated implementation of the option into the proposed state goals. 256 79 Fed. Reg. 34,875. EPA similarly rejected proposing natural gas co-firing as BSER for modified and reconstructed coal-fired EGUs. The Agency noted that: While conversion to or co-firing with natural gas in a utility boiler is a technically feasible option to reduce CO2 emission rates, it is an inefficient way to generate electricity compared to use of an NGCC and the resultant CO2 reductions are relatively expensive. The EPA found costs for natural gas co-firing to range from approximately $83/ton to $150/ton of CO2 avoided. Even for cases where the natural gas could be co-fired without any capital investment or impact on the performance of the affected facility (e.g., an existing IGCC facility that already has a sufficient natural gas supply), the costs of CO2 reduction would still be approximately $75/ton of CO2 avoided. 79 Fed. Reg. at 34,982 (citation omitted). Even at the low end of the range cited by the Agency these costs are clearly too high for an NSPS. Indeed, if EPA has determined that co-firing is inappropriate as BSER for modified and reconstructed Subpart Da units, which pursuant to section 111(a)(2) are deemed to be new units, it certainly cannot be deemed to be BSER for existing Subpart Da units. Some environmental groups and states have asked EPA to include in its consideration of the costs with regard to BSER “some benefits associated with the co-firing of natural gas with coal” such as “reduc[ing] emissions of nitrogen oxides (NOx); sulfur dioxide (SO2); particulate matter; and hazardous air pollutants, including mercury.” 79 Fed. Reg. at 64,550. As discussed in Section XII.C., EPA’s purported co-benefits cannot be attributed to the Proposed Guidelines, and in any event are inappropriate in its making a BSER determination. Finally, EPA solicits comment “on other factors or variables that might affect the decision to use natural gas in co-firing at a particular unit (e.g., type, age, or size of a boiler), as well as factors that could limit the amount of co-firing that could be done.” Id. There are technical challenges in converting a coal-fired boiler to burn natural gas, whether to co-fire both coal and natural gas through separate burners, dual fuel fire both coal and gas in the same burner, 257 fire either coal or gas, or fire only natural gas. These technical issues must be evaluated on a case-by-case basis, taking into consideration the remaining anticipated useful life of the unit. Potential necessary alterations can concern the following: (1) initial boiler design; (2) burner design; (3) gas piping; (4) burner management; and (5) boiler and combustion controls. Smith Report at 2. Depending upon the specific boiler design, changes to a coal-only designed boiler to accommodate natural gas firing may include alterations to the following: (1) convective pass (superheat, reheat, attemperation, etc.); (2) fuel supply system; (3) control system; (4) burners; and (5) emission control systems. Id. at 4. With respect to the availability of natural gas, the following must be considered: (1) proximity of a major gas pipeline; (2) capacity of the source pipeline; (3) regional breakdown between residential and industrial natural gas use; (4) long-term natural gas supply availability; (5) potential for interruptible natural gas supply; and (6) costs and benefits of alternative emission reduction methods. Id. If EPA wishes to pursue the possibility of including natural gas co-firing at existing boilers as BSER, it must undertake a thorough engineering and cost evaluation, as required by section 307(d)(3) of the CAA, and allow UARG a sufficient period of time to evaluate and comment on those studies. C. Issues Regarding an Alternative “Glide Path” to Compliance EPA’s NODA seems to recognize for the first time that the Proposed Guidelines’ interim goals (to be met on average from 2020 to 2029) deny states the implementation flexibility and safeguards for reliability that the CAA requires. This result is the opposite of what EPA states it intended to preserve in the Proposed Guidelines. See 79 Fed. Reg. at 64,545. The Agency claims that its intent in proposing to require compliance with an interim goal is to “provide states with a reasonable glide path to compliance with their final goals by 2030.” Id. at 64,548. Yet EPA fails to explain why interim goals are necessary to achieve the Proposed Guidelines’ ultimate goal. For the reasons discussed in Sections XIII and XIV of these comments, EPA 258 should eliminate entirely any interim goals for states. Such requirements can severely limit a state’s flexibility and ability to respond to reliability concerns, which EPA disingenuously claims is an overriding goal of the Proposed Guidelines. This NODA is the first time that EPA has discussed the use of a “glide path” for states to comply with the 2030 final goals, and the Agency has neither formally nor informally defined that term. The Agency, however, often uses the phrase “glide path” in other CAA contexts. For example, “glide path” is commonly employed with respect to the “reasonable progress goals” in state implementation plans addressing regional haze requirements. See 40 C.F.R. § 51.308(d)(1). In the regional haze context, the states, rather than EPA, play the primary role in determining whether their own glide path toward attaining natural visibility conditions may deviate from a uniform rate of improvement, based on considerations of “costs of compliance, the time necessary for compliance, the energy and non-air quality environmental impacts of compliance, and the remaining useful life of any potentially affected sources.” Id. § 51.308(d)(1)(i)(A). UARG believes that if any final emission guidelines were to include interim goals, that they should be flexible and allow states to comply using a reasonable glide path. Each state should have the authority to define its own glide path using criteria such as costs, energy impacts, and the remaining useful life of affected sources. Final guidelines that allowed such a glide path would be a logical outgrowth of the Agency’s June 2014 Proposed Guidelines. Regarding Building Block 2, EPA suggests ways to reduce the negative consequences of requiring an interim goal: (1) allowing credit for early CO2 emission reductions; and (2) phasing in Building Block 2 over time rather than assuming full redispatch by 2020. 79 Fed. Reg. at 64,545. If EPA does not abandon interim goals, UARG urges the Agency to adopt both of these changes in the final guidelines. First, UARG supports providing full credit for early actions that 259 result in reduced CO2 emissions before the compliance period begins in 2020. This relaxation from the proposed approach alone, however, will not significantly alleviate the negative consequences of EPA’s proposed interim goals. The timeline between final state plan approval (which, for many states, may not occur until mid-2018 or later) and 2020 is far too short for states to implement significant action to encourage implementation of early emission-reducing measures that would counteract the Agency’s stringent interim goal requirements. EPA should provide for redispatch under Building Block 2 to be phased in, rather than be fully implemented by 2020. As EPA now appears to recognize, requiring full redispatch by 2020 makes it impossible for some states to meet the interim goal because of “technical, engineering, and infrastructure limitations.” Id. at 64,546. One consequence of EPA not phasing in redispatch would be the required shutdown of many coal-fired units that “have recently been constructed” or “have received significant capital investment (e.g., in the form of pollution control retrofits).” Id. Full redispatch by 2020 is simply unachievable and bad public policy. EPA should allow the states to determine how quickly Building Block 2 may be phased in. The Agency suggests that an appropriate phase-in schedule could be based on “whether, and to what extent, any additional infrastructure improvements (e.g., natural gas pipeline expansion or transmission improvements) are needed to support more use of existing natural gas-fired generation,” or on avoiding stranded investments considering the “book life” of generating assets. Id. at 64,548-49. States are in a much better position than EPA to assess these factors and determine an appropriate phase-in schedule. EPA gives no indication as to how it would go about assessing these factors and developing a phase-in schedule on its own for the states. UARG is pleased that EPA seems to have discovered that section 111(d)(1) requires EPA to “take into consideration, among other factors, the remaining useful life of the existing source 260 to which [a] standard applies.” CAA § 111(d)(1). UARG strongly objects, however, to the Agency’s apparent suggestion that book life is an appropriate means of implementing this statutory mandate. Book life is an accounting and ratemaking concept that concerns depreciation. It does not describe the end of useful life of an asset, which should be determined on a case-by-case basis. Moreover, the statutory mandate to consider the remaining useful life by implication includes the discretion to exempt such sources from regulation. UARG also supports altering the Proposed Guidelines to provide for the gradual phase-in of heat rate improvements under Building Block 1. 79 Fed. Reg. at 64,548. Efficiency improvement measures require significant time and investment to implement, including planning, contracting with vendors to perform this work, and scheduling outages in which it can be performed. Under the Proposed Guidelines, every coal-fired EGU that is not retired would be expected to implement these measures by 2020. This would place an enormous strain on the energy grid as numerous units go down for outages at once in order to implement these measures by 2020. Units would be reluctant to begin this work before a final state plan is approved, which in many states may not occur before mid-2018 (and in states with multi-state agreements may not occur before mid-2019). NERC just issued a report that emphasizes that EPA’s Proposed Guidelines pose a serious obstacle to the reliable supply of electricity in this nation. Initial NERC Report at 9-10, 22. NERC is an international regulatory authority established to evaluate and improve the reliability of the BPS in North America. Among its findings in its Executive Summary are that under EPA’s Proposed Guidelines: (1) essential reliability services may be strained; and (2) more time for implementation may be needed to accommodate reliability enhancements. Id. at 2-3; see also Section VII.B.4 (discussing Initial NERC Report). 261 D. Regionalized Approach to Building Blocks 2 and 3 EPA solicits comment on a third “regionalized” approach to RE in Building Block 3. The Agency’s new approach was not included in the June 14 Proposed Guidelines and is not actually proposed in the NODA. Under a “regionalized” approach, EPA would “adjust[] each state’s RE target based on the RE potential available across a multi-state region in which the state is located” rather than attempting to determine what level of renewable energy could be generated within the state itself. 79 Fed. Reg. at 64,551. This approach would “group states into regions; aggregate RE generation potential across states within each region; and then reapportion the aggregate identified RE generation to individual states according to criteria that assume regional RE development in which parties in multiple states participate, regardless of the specific state where the generation occurs.” Id. EPA believes this approach would “better align RE targets with the proposal to allow the use of certain out-of-state renewables for compliance.” Id. Similarly, EPA solicits comment on using a regional approach to redispatch under Building Block 2 that would include requiring redispatch across state lines. Again, EPA does not actually propose this approach and merely seeks comment on an approach whereby “regional availability of NGCC generation would be considered rather than just in-state availability of NGCC generation in setting building block 2 targets.” Id. at 64,550. Under this approach, the levels of required generation shift under Building Block 2 could be determined in at least three different ways—using six regions whose borders are informed by NERC regions, RTOs, or some alternative regional structure. Id. at 64,551. The Agency explains that using regional approaches could “mitigate the concerns expressed” by certain unnamed stakeholders “that building block 2 has little or no effect on certain states with large amounts of coal-fired generation and limited excess NGCC capacity.” Id. 262 As noted, EPA has neither proposed any of the regional approaches on which it requests comment nor has it provided sufficient information such as technical analyses that UARG needs to allow it to comment in a meaningful way. The Agency provides no direction on how states might be grouped into regions, instead listing four significantly different options and also soliciting comment on “some other approach.” Id. EPA does not specify how the aggregate RE generation potential across states within each region might be calculated. The Agency does not specify what criteria would be used to reapportion aggregate RE generation into state goals, instead asking commenters for direction on “a simple state-specific quantitative characteristic that reflects interstate patterns to develop RE potential at reasonable cost across a region.” Id. And EPA does not specify what components of a state’s RE targets would be “regionalized” under this approach. Id. at 64,551. The questions posed in the NODA reveal the incomplete nature of the entire Proposed Guidelines. EPA asks a long series of questions concerning many different approaches, the permutations and combinations of which are staggering in number. The Agency seems to try to shift the burden of justifying its Proposed Guidelines to the regulated community. EPA’s musings and requests for thoughts on a variety of different approaches essentially show that the Agency has little if any idea how its Proposed Guidelines might work. The Agency’s failure to complete the necessary background work suggests that a more appropriate course of action at this time would have been to issue an advance notice of proposed rulemaking designed to elicit information to use in formulating a proposal, rather than to propose an approach that is both poorly-conceived and has not been fully analyzed by the Agency and then issue a NODA designed to elicit information. The process EPA has employed here is backwards. 263 Although UARG appreciates the Agency’s interest in obtaining information necessary to develop a sound proposal, the myriad questions posed in the NODA confirm that EPA has failed to collect the data and perform the technical analyses needed to identify and analyze the impacts of the Proposed Guidelines. Most of EPA’s questions with regard to various regionalized approaches cannot be answered without detailed analyses of emissions and other data that the Agency should have undertaken and placed in the docket as part of its basis and purpose of the Proposed Guidelines. CAA § 307(d)(3). In this regard, the APA defines a rule as a statement of “future effect” designed to “prescribe law or policy.” 5 U.S.C. § 551(4). Because a rule is intended to prescribe the legal standards that govern the future conduct of regulated parties, a proposed rule must be more than a vague set of criteria for future applications with a series of questions regarding how the Agency might formulate a final rule. See Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506, 549 (D.C. Cir. 1983) (“Agency notice must describe the range of alternatives being considered with reasonable specificity” to allow for adequate public comment and “better-informed agency decisionmaking”; unspecified, general notice is inadequate.). The Agency’s Proposed Guidelines are entirely inadequate in this regard, and thus EPA should terminate this proceeding until it has had time to gather and analyze the data needed to formulate a proper proposed rule. E. Goal-Setting Methodology EPA’s NODA solicits comment on whether EPA should alter its goal computation methodology in order to reduce generation more drastically from existing fossil fuel-fired EGUs. The Agency neither proposes this change in the NODA nor did it propose such an approach in the June 2014 Proposed Guidelines. Instead, the Proposed Guidelines implement Building Block 2 redispatch from coal to natural gas by “subtract[ing] 1 MWh of fossil steam generation and corresponding emissions from the 2012 baseline levels for every 1 MWh of incremental NGCC 264 generation.” 79 Fed. Reg. at 64,552. This leaves the total level of overall generation in a state unchanged after implementation of Building Block 2. By way of contrast, the formula for Building Blocks 3 and 4 “incremental RE and EE to 2012 baseline generation levels . . . but does not reduce the 2012 baseline levels of fossil generation . . . by that incremental RE and EE, or remove the corresponding emissions.” Id. EPA now solicits comment on an alternative approach in which “incremental RE and EE explicitly replaces generation from fossil fuel-fired sources in the goal calculation.” Id. According to EPA, this approach would “reflect the full potential, under the BSER, for incremental RE and EE to replace fossil steam generation.” Id. EPA forthrightly recognizes that adopting this alternative approach would necessarily “increase the collective stringency of the state goals” and subsequently increase the costs of the Proposed Guidelines. Id. at 64,553. UARG agrees that if EPA were to adopt this alternative goal computation approach it would necessarily reduce (i.e., make more stringent) states’ final and interim goals across the board, significantly increasing the costs of the Proposed Guidelines. This approach would further reduce the share of a state’s generation that could come from existing fossil fuel-fired EGUs, and in states where that generation would already be significantly reduced due to other components of the Proposed Guidelines, the alternative approach could eliminate coal-fired generation altogether—just as Building Block 2 would already eliminate coal-fired generation in several states. UARG opposes this approach for three reasons. First, as discussed in Section III.B, EPA has no authority to establish a standard of performance under section 111 based on reduced utilization (or retirement) of a source. Second, if EPA wishes to pursue such a dramatic change in its Proposed Guidelines, it must undertake a thorough evaluation of costs and other implications, as required by section 307(d)(3) of the CAA, and allow UARG a sufficient period 265 of time to evaluate and comment on those studies. UARG suspects there will be significant costs, including additional stranded assets, as well as additional threats to the reliable supply of electricity. It cannot intelligently and meaningfully evaluate these issues, however, without the additional analyses that EPA has failed to undertake. Third, beyond the likely very high additional costs, this alternative approach is nonsensical as a method to achieve EPA’s apparent overarching goal of reducing generation from all fossil fuel-fired EGUs because it would have the counterintuitive effect of encouraging the construction of additional new fossil fuel-fired EGUs relative to EPA’s June 2014 Proposed Guidelines. According to EPA, the rationale for its proposed goal computation approach (which adds incremental RE and EE to the goal formula without simultaneously reducing fossil generation) was that the additional RE and EE would replace expected generation increases from fossil sources that otherwise occur after 2012. Id. at 64,552. In other words, RE and EE would be used to meet future generation needs, while leaving the existing fossil EGUs to continue serving existing (i.e., pre-2012) load demand. The alternative approach on which comment is requested in the NODA would require that incremental RE and EE be used first to satisfy existing (i.e., pre-2012) demand by displacing generation from existing fossil EGUs. Id. If the incremental RE and EE spurred by the Proposed Guidelines is dedicated to meeting historical demand, any post-2012 increases in demand will need to be met by other sources. Because the proposed goals for RE and EE under Building Blocks 3 and 4 are already extremely aggressive, the most likely source for additional future generation is new fossil fuel-fired EGUs. Therefore, one likely outcome of the alternative approach discussed in the NODA would be to encourage the use of new fossil generation to satisfy future demand rather than RE—a policy outcome that seems to be at odds with EPA’s overall goals. 266 XVI. Monitoring, Recordkeeping, and Reporting Requirements for EGUs EPA proposes to require that state plans include monitoring provisions for EGUs that are no less stringent than those in its Proposed Guidelines at Subpart UUUU § 60.5805. According to EPA, monitoring under Part 75 and reporting to EPA’s ECMPS would generally satisfy those requirements, with only a few exceptions. 79 Fed. Reg. 34,913. Part 75 requires monitoring and reporting of CO2 mass emissions (in tons per hour or tons per day) and hourly gross unit load from fuel combustion (in gross megawatt electrical or “MWge”). Although Part 75 does not specify a method for measuring unit load, EGUs routinely measure gross load with a watt meter (or meters). For CO2 mass emissions, Part 75 provides four options. 53 First, any EGU may use a CO2 CEMS and a stack flow monitoring system. For that option the CO2 CEMS consists of a CO2 concentration monitor and an automated data acquisition and handling system (“DAHS”). 40 C.F.R. §§ 72.2, 75.10(a)(3)(i), 75.13(a). Second, any EGU may use a CO2 CEMS and a stack flow monitoring system, where the CO2 CEMS consists of an oxygen (“O2”) concentration monitor and a DAHS. Id. §§ 75.10(a)(3)(iii), 75.13(a). EGUs that use an O2 concentration monitor convert O2 concentration to CO2 concentration using fuel factors (“F-factor”) 54 provided in Part 75, Appendix F § 3.3.5. Id. § 75.13(c) and pt. 75, App. F § 4.4.1. Appendix F, Table 1 provides F- and Fc-factors for most 53 These methods also are described in detail in EPA’s Technical Support Document for Stationary Fuel Combustion Emissions: Proposed Rule for Mandatory Reporting of Greenhouse Gases at 5-9 (Jan. 30, 2009), Docket ID No. EPA-HQ-OAR-2008-0508-0004 (“EPA GHG MRR TSD”). 54 F-factors represent the ratio of gas volume to the calorific value of the fuel combusted. Identical F-Factors are used to calculate emission rates in all NSPS affecting EGUs. See 40 C.F.R. pt. 60, App. A-7, Method 19. An Fc-factor specifically represents the volume of CO2 generated. 267 commonly used fuels, but also allows calculation of a site-specific F- or Fc-factor in the event one is not provided for a specific fuel. Id. pt. 75, App. F §§ 3.3.5., 3.3.6. For options one and two, the CO2 (in ppm or percent) and the volumetric stack flow (in standard cubic feet per hour) are used, along with a constant (or “K-factor”) for CO2 and, if CO2 or O2 concentration is measured on a dry basis, an hourly average stack moisture value (in percent H2O) to determine hourly CO2 mass emissions in tons per hour. Id. §§ 4.1 and 4.2. According to EPA, in 2006, eighty-five percent (85%) of total CO2 mass emissions reported under Part 75 were reported using one of these two methods. EPA GHG MRR TSD at 6, Tbl. 4. That constitutes thirty-three percent (33%) of the total number of units reporting under Part 75. Id. Third, any EGU that combusts only fossil fuel may determine CO2 mass emissions based on the measured carbon content of the fuel and the procedures in Appendix G to estimate CO2 in tons per day. 40 C.F.R. §§ 75.10(a)(3)(ii), 75.13(b). Under that procedure, carbon content based on analysis of a fuel sample collected at the specified period (e.g., per shipment or delivery for oil and gas delivered in lots, monthly for natural gas, and daily for other gases) and fuel feed rates using a fuel flow meter (for oil and gas) are used to determine CO2 mass emissions in tons per day. 55 Id. pt. 75, App. G § 2.1. EGUs using wet FGD systems or a fluidized bed boiler must calculate and include CO2 from sorbent. Id. § 75.13(b) and pt. 75, App. G § 3. According to EPA, in 2006, only one percent (1%) of CO2 mass emissions reported under Part 75 was reported 55 Coal-fired EGUs also are allowed to use this option under Part 75 using weekly coal sampling and company records of the amount of coal combusted. If this option were chosen, the EGU could adjust the estimated CO2 for carbon retained in ash. 40 C.F.R. § 75.13(b) and pt. 75, App. G § 2.2. However, no Part 75 coal-fired EGU has opted to use this method. EPA, Plain English Guide to the Part 75 Rule at 13 (June 2009), available at http://www.epa.gov/airmarkets/ emissions/docs/plain_english_guide_par75_final_rule.pdf (“Part 75 Plain English Guide”). 268 using this method. EPA GHG MRR TSD at 6, Tbl. 4. That constitutes five percent (5%) of the total number of units reporting under Part 75. Id. Fourth, any EGU that meets the definition of gas-fired or oil-fired 56 may determine CO2 mass emissions using hourly heat input and the same F-factors referenced above (for EGUs using an O2 concentration monitor). 40 C.F.R. pt. 75, App. G § 2.3. Heat input is determined using fuel feed rate and the gross calorific value of the fuel determined under Part 75, Appendix D with a certified fuel flow meter and fuel sampling and analysis, respectively. 40 C.F.R. § 75.13(b), pt. 75, App. G, Equation G-4, and App. F § 5. According to EPA, in 2006, fourteen percent (14%) of total CO2 mass emissions reported under Part 75 were reported using this method. EPA GHG MRR TSD at 6, Tbl. 4. That constitutes fifty-nine percent (59%) of the total number of units reporting under Part 75. This Appendix G method is the most frequently used method for Part 75 gas- and oil-fired EGUs. Part 75 Plain English Guide at 13. Although the CO2 mass emission values reported under Part 75 are not used to determine compliance with any requirement under the Acid Rain Program (“ARP”), all the data used to calculate CO2 mass emissions (except measured carbon and daily fuel use) also are used to calculate compliance with one or more emission limits under the ARP, and under other NSPS applicable to EGUs. Accordingly, EPA already has determined that the data are of sufficient quality for compliance calculations like those proposed in this rulemaking. States participating in the RGGI 57 trading program also rely on these Part 75 methods to determine compliance with that program’s CO2 mass emissions allowance holding requirement (although some states have 56 Gas-fired and oil-fired EGUs do not include EGUs that combust coal, or solid or liquid coal-derived fuel, but do include EGUs that combust coal-derived gaseous fuel. 40 C.F.R. § 72.2. 57 RGGI is a mandatory state CO2 cap and trade program covering nine northeastern and mid-Atlantic States. 269 opted not to allow use of the Part 75 option for calculating daily CO2 mass emissions under Appendix G § 2.1). Part 75 Plain English Guide at 12. EPA’s Proposed Guidelines generally recognize the adequacy of Part 75 data by proposing that states allow use of two of the four options allowed under Part 75, and by requiring compliance with Part 75 monitoring plan and quality assurance and quality control (“QA/QC”) requirements. Based in part on this similarity, EPA’s proposed information collection request (“ICR”) assumes “no new information collection burden” beyond what already is required under Part 75 and the GHG Reporting Rule at 40 C.F.R. Part 98, at least for the first three years following final rule promulgation. EPA, Supporting Statement, Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units (40 CFR Part 60, Subpart UUUU) at 3 (May 2014), Docket ID No. EPA-HQ-OAR-2013-0602-0438 (“Proposed Existing Source ICR Supporting Statement”). EPA appropriately proposes not to allow use of Part 75 bias-adjusted data. Proposed Subpart UUUU § 60.5805(a)(2)(iv). However, in some instances, EPA proposes to require that states impose new requirements, some of which could increase burden and stringency beyond existing requirements once state plans are implemented. 79 Fed. Reg. at 34,913-14. UARG supports EPA’s proposal that states rely on Part 75 monitoring procedures, without bias adjustment, 58 but objects to the proposed elimination of one of those options and imposition of additional requirements for existing units. EPA has provided no justification for requiring states to impose on EGUs monitoring that is different from, let alone more stringent 58 The bias adjustment factor is a one-way (positive) adjustment that was justified based on the ARP’s status as a market-based program. In recognition of that fact, Part 75 does not require bias-adjustment of diluent or flow data used to calculate CO2 mass emissions, and EPA has never required use of Part 75 bias-adjusted data to determine compliance under an NSPS. See, e.g., 40 C.F.R. § 60.49Da(c)(2), (d), (m). This rule should be no different. 270 than, Part 75. UARG also objects to the Agency’s failure to make clear that use of Part 75 missing data substitution procedures is not appropriate in compliance calculations. 59 A. Proposed CO2 Monitoring Methods 1. The CEMS Option EPA proposes that states require existing EGUs that combust solid fuel (and to allow other existing EGUs) to monitor CO2 mass emissions using a CO2 concentration monitor, flow monitor, and (if applicable) a moisture monitoring system in accordance with Part 75. Proposed Subpart UUUU § 60.5805(a)(2)(i). As an alternative, EPA proposes to allow EGUs that combust only liquid or gaseous fuel to determine CO2 mass emissions under the provision in Part 75, Appendix G § 2.3 that uses heat input determined under Part 75, Appendix D and a carbonbased F-factor to determine CO2 mass emissions. Id. at § 60.5805(a)(3). UARG supports these options, but believes that EPA has unreasonably eliminated several other Part 75 options. UARG does not oppose requiring solid fuel-fired EGUs, or allowing other EGUs, to use CEMS. Part 75 and Subpart Da already require solid fuel-fired units to use CEMS to determine compliance with other requirements. However, UARG does not understand why EPA has specifically required use of a CO2 CEMS (and not allowed use of an O2 CEMS to obtain CO2) at an EGU that does not use carbon separation. 60 Both existing NSPS applicable to EGUs— 59 In the proposed GHG NSPS for new, modified and reconstructed sources under CAA § 111(b) EPA made clear that Part 75 missing data substitution procedures cannot be used. See, e.g., 79 Fed. Reg. 1430 (Jan. 8, 2014), Proposed Subpart TTTT § 5540(a)(1), Subpart Da § 60.46Da(g)(1)(i), Subpart KKKK § 60.4374(a)(1); Memorandum from EPA OAQPS to EGU NSPS Docket (EPA-HQ-OAR-2013-0603), “Amended Regulatory Text (Proposed Applicability)” (June 2014), Docket ID No. EPA-HQ-OAR-2013-0603-0044. UARG assumes that the failure to prohibit use of substitute data was an oversight. 60 An O2 CEMS cannot be used to obtain accurate CO2 measurement at the stack of an EGU that employs carbon separation because the calculation to convert O2 measured at the stack to CO2 relies on the use of F-factors. See, e.g., 40 C.F.R. pt. 75, App. F § 4.4. Any equation that 271 Subparts Da and KKKK—allow use of either a CO2 concentration monitor or an O2 concentration monitor as the diluent monitor to calculate emissions in lbs per MMBtu. And, Part 75 explicitly allows use of an O2 concentration monitor to measure CO2 concentration to obtain CO2 mass emissions. There is no valid reason not to allow both types of monitoring systems under this rule as well. Although most coal-fired EGUs opt to use CO2 concentration monitors, 61 there are valid reasons an EGU (particularly a gas- or oil- fired EGU) might prefer an O2 monitor. Oxygen analyzers generally cost less, are more stable, and are less affected by interferences (e.g., H2O and CO) that can impact CO2 measurements. EPA could easily add an option for use of an O2 monitor simply by adding a reference to § 75.10(a)(3)(iii) in Proposed Subpart UUUU § 60.5805(a)(2). UARG also is concerned that the proposed language requiring EGUs that measure CO2 concentration on a dry basis to “install, certify, operate, maintain, and calibrate a continuous moisture monitoring system” under § 75.11(b), Proposed Subpart UUUU § 60.5805(a)(2)(i), could be interpreted as eliminating the option in § 75.11(b)(1) to use appropriate default moisture values rather than installing and calibrating a monitoring device. The Part 75 default moisture values for coal and wood were derived from the Agency’s own evaluation of data in a 1999 rulemaking in which the Agency determined that they were sufficiently conservative to ensure no under-reporting. 64 Fed. Reg. 28,564, 28,567, 28,568, 28,590 (May 26, 1999). EPA uses an F-factor will not provide accurate measurements downstream of any device that removes CO2 from the gas stream. However, UARG does not expect any existing EGU to employ carbon sequestration. 61 Dilution-based CO2 concentration monitors are the monitoring system of choice for coal-fired EGUs because they measure on a wet-basis. Dilution systems are often the only reasonable choice for EGUs with wet FGD because they eliminate plugging and other issues associated with corrosive flue gas conditions. Dilution systems cannot measure O2 because the dilution air contains O2. 272 approved the default moisture value for natural gas (boilers only) in 2008 based on data submitted to EPA. 73 Fed. Reg. 4312, 4315, 4342 (Jan. 24, 2008). UARG can think of no reason why EPA would eliminate this option for existing units, and hopes that the reference to installing and calibrating systems was not intended to eliminate use of the default values. If the Agency has a reason for eliminating the O2 monitoring and moisture default value options for either coal-fired EGUs or liquid and gas-fired EGUs, EPA must issue a rulemaking proposal identifying that reason so that UARG can comment on the Agency’s rationale. Because existing EGUs still will need to comply with the ARP, eliminating these options in state plans effectively eliminates them in Part 75. For EGUs using CEMS, EPA proposes a new requirement to measure the dimensions of each stack or duct at the flow monitor and reference method sampling location using a laser device at “three or more distinct locations and average the results.” Proposed Subpart UUUU § 60.5805(a)(2)(iii). UARG objects to this provision. A requirement for a one-time measurement to establish stack diameter and area at one location using a laser device would not be unreasonable, if EPA had established that such devices are more accurate than other measurement methods and are not overly burdensome. EPA’s rulemaking record, however, contains no evidence that existing methods are not sufficiently accurate, and no evidence that laser devices would provide significantly more accurate results. As a result, it appears that the requirement imposes additional cost and burden with no corresponding benefit. UARG also has concerns that measurements taken with a laser device while a unit is operating may be adversely affected by stack flow. If EPA has evidence to support this proposed requirement, EPA should provide it. In addition, the proposal to require a measurement at three “distinct” locations is not reasonable and may not be possible on some stacks. Measurement at three “distinct” locations 273 presumably would require three ports none of which are directly across from each other. Common configurations depending on stack diameter are to have two ports 90 degrees apart or four ports 90 degrees degree apart—neither of which provides three distinct locations. In other words, most plants will have at most two distinct locations. EPA must remove the requirement. 2. The Heat Input and Fuel Factor Option UARG agrees with EPA’s proposal to allow liquid and gaseous fuel-fired EGUs to use the Appendix G in lieu of CEMS. Requiring use of stack flow or diluent monitors on such EGUs would result in unnecessary additional construction, equipment, and implementation costs, and less accurate measurements of CO2 mass emissions. 62 UARG also agrees with the proposal to allow, but not require, the use of site-specific F-factors under equation G-4. Proposed Subpart UUUU § 60.5805(a)(3)(ii). Appendix F provides F-factors for most commonly used fuels and includes a procedure for annual determination of a site-specific F-factor for fuels that are not listed, or if a source wants to develop its own F-factor. However, the variability in commonly used fuels is not sufficient to warrant determination of a site-specific F-factor for all fuels. To the extent there is small variability, it also is unlikely that variability would be captured in an annual determination. EPA has relied on F-factors identical to those in Part 75, Appendix F in 62 Existing liquid and gas-fired EGUs use Appendix D to comply with the ARP requirements to monitor and report hourly heat input and SO2 mass, and the Subpart KKKK requirement to monitor NOx in lb/MWh. EPA previously considered and rejected the idea of requiring such units to install stack flow monitoring systems in lieu of using Appendix D. 66 Fed. Reg. 18,546 (Apr. 10, 2001). The Appendix D procedure, which relies on a calibrated fuel flow meter, provides a more direct, and more accurate, measurement for such EGUs, because fuel flow meters can meet tighter performance specifications than stack flow monitors, and they are not subject to the same interferences that can bias measurement. Requiring use of a diluent CEMS at gas- and oil-fired combustion turbines also would be unjustified. The combined potential errors associated with CEMS stack monitoring, including allowed variability in Protocol 1 calibration gases, calibration error, and relative accuracy, are greater than the error associated with the mass balance options available under Part 75. 274 NSPS compliance determinations since the mid-1970s. Requiring development of a site-specific F-factor for commonly used fuels would be a waste of time and resources. However, the option to develop such factors is necessary to account for the fact that EPA may not have developed a default factor for some fuels. 3. Monitoring of CO2 Mass Emissions From Integrated Equipment EPA proposes to define “steam generating unit” and “stationary combustion turbine” to include “any integrated equipment that provides electricity or useful thermal output” to the affected facility or auxiliary equipment. 63 Proposed Subpart UUUU § 60.5820. When EPA proposed to include “integrated combustion turbines and fuel cells” in the definition of a postMay 3, 2011 steam generating unit in Subpart Da, EPA recognized that emissions from that equipment likely would not be exhausted through the primary boiler stack and, because the emissions would be relatively small, suggested that they be estimated rather than monitored. 76 Fed. Reg. 24,976, 25,070 (May 3, 2011). The Agency also acknowledged that there likely were no such devices currently being used, but suggested that it hoped to promote their use. Id. UARG’s comments on that definition focused on its lack of clarity and usefulness, given the improbability that such equipment would be used in the near term. Although UARG continues to believe that there are few real world applications for such equipment, at least at fossil-fuel-fired utility boilers, UARG agrees that if such equipment were used, continuous monitoring likely would not be warranted. Such integrated equipment either 63 Such integrated equipment could be interpreted to include a fuel reformer that produces hydrogen to power a fuel cell or combustion turbine, a start-up boiler that enables cold start of a large power generation steam generator, a small industrial boiler that fires continuously to generate steam to warm the heat recovery steam generator (“HRSG”) of NGCC units, or a combustion turbine that preheats boiler feedwater for a HRSG or provides auxiliary steam for the stripping tower of a CCS. 275 has very few emissions, or operates very infrequently. UARG also believes that given the uncertainty about their use, the Agency should not spend resources specifying monitoring requirements in the rule. As a result, UARG suggests that EPA encourage states to include a provision allowing submission and approval of a petition for exemption of emissions that are de minimis from the compliance calculation, or for an alternative means of accounting for any resulting emissions (e.g., a mass balance calculation or engineering estimate) from any integrated equipment that are not otherwise required to be monitored under an NSPS or otherwise accounted for in the primary equipment’s monitoring (e.g., by venting the emissions into a stack with CEMS). B. Monitoring and Reporting of Electric Output Output-based standards require the monitoring of generation. To determine compliance with the proposed output-based standards, EPA proposes to require installation of a sufficient number of watt meters to measure “net electric output” from the facility, as well as specifying special provisions for CHP facilities and process steam applications. Proposed Subpart UUUU § 60.5805(a)(4). For the first time, EPA also proposes to require that these measurements be made “using 0.2 class electricity metering instrumentation and calibration procedures as specified under [American National Standards Institute (“ANSI”)] Standards No. C12.20.” Id. EGUs already are required to record and report hourly gross unit load from fuel combustion (in MWge) under Part 75. 40 C.F.R. §§ 72.2 (defining “unit load”), 75.57(b)(3) (requiring recording of hourly gross unit load), 75.64(a)(2) (requiring reporting of recorded values). 64 As a result, all existing EGUs already measure and record gross MWs for all 64 Although Part 75 does not specify a method for measuring generation, Subparts Da and KKKK require EGUs complying with an output-based standard to install, calibrate, operate, and record data from a watt meter. See 40 C.F.R. §§ 60.49Da(k)(1), 60.4335(b)(3). Combined cycle 276 generation. In order to facilitate reporting under state plans, EPA proposes that affected EGUs also “could report hourly net electric generation under [Part 75] via ECMPS.” Technical Support Document: Part 75 Monitoring and Reporting Considerations at 6, Docket ID No. EPAHQ-OAR-2013-0602-0461 (“Part 75 GHG Guideline TSD”). 65 It is not clear whether EPA is proposing that Part 75 be revised to simply provide a mechanism for such reporting in the event a state chooses to require or allow it, or whether EPA is suggesting that sources reporting under Part 75 be required to report net electric generation. UARG would oppose the latter requirement. EPA has no authority under section 111(d) to require reporting of net generation or output under Part 75. UARG also is concerned that EPA has overlooked important details regarding how generation already is being reported. EPA has made clear in its ECMPS Reporting Instructions for Emissions, 66 the value reported under Part 75 represents an hourly rate and not the total load for the hour. As a result, for hours of partial operation, the value already reported under Part 75 will not represent the actual output for that hour. See ECMPS Emissions Reporting Instructions § 2.4. Similarly, units on common stacks report the weighted sum of the hourly unit loads for all units that exhaust through the stack. In addition, reported steam load for an hour is reported at the measured temperature and pressure (i.e., is not corrected to ISO conditions). Id. EPA’s Proposed Guidelines, on the other hand, require calculation and use of actual electrical output for combustion turbines with HRSGs that do not have a duct burner are not required under Part 75 to report generation from the HRSG, but may do so voluntarily (and most do). See EPA, Part 75 Emissions Monitoring Policy Manual at Q. 17.2 n.1 (2013) (“Part 75 Policy Manual”) (advising EGUs to report total generation if required to monitor total MW under other applicable regulations), available at http://www.epa.gov/airmarkets/emissions/monitoring.html. 65 To implement such a scheme, CHP units also would need to report useful thermal output via ECMPS. 66 Available at http://ecmps.camdsupport.com/learn_docs.shtml (released June 12, 2013). 277 each hour and useful thermal energy relative to standard ambient temperature and pressure conditions. If EPA intends sources to include partial unit operating hours in state compliance calculations, the values already being reported under Part 75 would need some adjustment for use in compliance calculations. EPA suggests as much with respect to output in its proposal to require that EGUs with common stacks or multiple stacks monitor “stack operating time.” Proposed Subpart UUUU § 60.5805(5), (6). As a result, EPA should clarify whether it intends that partial hours be included in state calculations. And, if EPA anticipates that states rely on data reported to ECMPS, EPA must address the discrepancies. With respect to the proposed requirement to comply with the ANSI 2010 C12.20 standard for electricity meters, EPA has not provided justification for that proposal in this rulemaking. In a TSD included in the docket, EPA says only that the requirement “would ensure a level playing field regarding the minimum acceptable accuracy of equipment … while minimizing any additional burden of upgrading equipment used to measure net generation.” Part 75 GHG Guideline TSD at 4. However, EPA has not provided any information to suggest that all EGUs are not already using meters of acceptable accuracy or that requiring compliance with the ANSI standard would not impose additional burdens or require the upgrading of equipment. EPA has never before found it necessary to impose such requirements because EGUs already have sufficient incentives to ensure that the electricity they generate and use is accurately measured. Compliance with that proposed requirement would be burdensome. Although some EGUs may have meters that comply with all or parts of the ANSI standard, some do not and the cost of replacing such meters is not insignificant. Moreover, the ANSI standard itself is extraordinarily detailed and not at all suited to a regulatory program under which penalties could 278 be sought for a failure to comply. UARG notes that EPA has not provided a copy of the standard in the docket, or even discussed the content of the standard in any meaningful way. In short, UARG believes the only entity that would significantly benefit from requiring compliance with the cited ANSI standard is ANSI itself, which charges a significant sum for access to the standard and restricts the networking and transferring of the purchased file to another person or computer. 67 With respect to the monitoring of net electric output, EPA’s proposal appears to be based on an assumption that net output would be determined using existing equipment and simple apportionment schemes. Part 75 GHG Guideline TSD at 3 (“EPA understands that the equipment needed to convert gross generation to net generation on an hourly basis exists at all EGUs.”). UARG agrees that if the monitoring of net electric output is required, EPA must allow EGUs to use existing equipment and methods for metering station service, and to apportion common station service to individual units using unit generation. Id. at 3-5. As EPA is well aware, measurement of unit-specific net output can be complicated by the difficulty in measuring auxiliary energy consumption of numerous plant-specific devices. Most plants do not have the internal electrical distribution system designed to distribute and measure internal usage on a unit basis. If EPA were to require replacement of existing equipment or to impose more stringent standards on existing equipment, EPA would need to revise its proposal to estimate and solicit input on the additional costs and burdens of such metering. 67 See ANSI eStandards Store, http://webstore.ansi.org/FindStandards.aspx?Search String=C12.20&SearchOption=0&PageNum=0&SearchTermsArray=null%7cC12.20%7cnull. 279 C. Monitoring and Reporting of Useful Thermal Output EPA proposes to define net energy output as net electric or mechanical output, plus 75 percent of any useful thermal output. Proposed Subpart UUUU § 60.5820. Steam is the most common type of useful thermal output and is commonly measured using a flow meter, a temperature sensor, and a pressure sensor. Part 75 GHG Guideline TSD at 6. For CHP units, EPA proposes to require the continuous monitoring of useful thermal output from CHP systems, Proposed Subpart UUUU § 60.5805(a)(4), and assumes that a CHP unit that sells its output already would be using such equipment, Part 75 GHG Guideline TSD at 6. However, EPA solicits comment on whether EPA should specify “best practices” for measurement of useful thermal energy and quality assurance protocols to ensure consistent and accurate reporting, “while minimizing additional burden.” Id. UARG does not believe that it would be appropriate for EPA to specify which technologies to use or to impose requirements for periodic QA. As an initial matter, UARG questions EPA’s focus on the accuracy of useful thermal output measurements given the Agency’s uncertainty regarding the amount of credit that should be provided for whatever is measured. As noted above, although EPA has proposed to allow credit for 75 percent of measured useful thermal output, it also has solicited comment on providing credit for as little as two-thirds and as much as 100 percent. 79 Fed. Reg. at 34,914. Wherever EPA comes out on that issue from a policy perspective, any accuracy gains that might be achieved from imposing additional requirements surely would be lost in the noise of the calculations and assumptions used to justify the arrived upon percent credit. See, e.g., Memorandum to Docket, Credit for Thermal Output at Combined Heat and Power (CHP) Facilities, Docket ID No. EPA-HQ-OAR2013-0495-0070. 280 Regardless of the amount of credit provided for the measured thermal output, each EGU should be allowed to select the technologies that best suit its needs under the circumstances, and to determine the best mechanism to ensure an appropriate level of accuracy. Certain technologies (like ultrasonic flowmeters and coriolis meters) typically are not used for steam measurements due to the high temperature applications and costs, and it would not be appropriate for EPA to attempt to impose their use. Although some manufacturer’s may claim that their more complex and costly equipment provides greater accuracy than the widely used differential pressure type flow meters, those benefits often are not realized when the meters are used in real world plant applications. Different types of meters, instruments, and calculation methods also provide varying levels of accuracy depending on the steam conditions, which can vary from source to source. For example, some steam is supplied, and therefore metered, at conditions that are barely superheated or barely at the saturation temperature. Moreover, some equipment simply would not benefit from any periodic QA. For example, thermocouples and resistance temperature detectors either work or fail completely. Because removing such sensors for calibration may be difficult, and even dangerous in high pressure processes, there is no benefit to removing them until they fail and have to be replaced. Manufacturers’ recommendation also are not universally appropriate for ongoing maintenance or QA, and are not appropriate for use in rules because they vary from vendor to vendor and can change over time. UARG cannot comment on the reasonableness of requirements, procedures, or specifications EPA cannot reasonably identify. ASME Performance Test Codes also are not appropriate for ongoing QA because they were designed for performance testing, not long term performance. Although most flow meters probably would meet the ASME measurement of fluid flow in closed conduits (“MFC”) standards generally, 281 UARG does not believe EPA has justified imposition of such standards in this context, and any requirement for their use would have to refer simply to the standard series as a whole in order to ensure that the entire range of flow meter types and conditions used at existing plants would be covered. D. Monitoring Plan and QA/QC Testing EPA proposes to require preparation of a site-specific monitoring plan consistent with 40 C.F.R. § 75.53(g) and (h) and that “each monitoring system . . . meet the applicable certification and quality assurance procedures in § 75.20 … and Appendices B and D to part 75.” Proposed Subpart UUUU § 60.5805(a)(1), (2)(ii). EPA also says that it is considering requiring that the Part 75 monitoring plan include “the reporting of equipment used to measure net electric output (and net energy output for CHP units) in an EGU’s monitoring plan under [Part 75].” Part 75 GHG Guideline TSD at 7. EPA also is considering requiring the reporting of the results of any new QA/QC test on equipment used to measure net electric or energy output to ECMPS. Id. Again, it is not clear whether EPA is proposing that Part 75 be revised to simply provide a mechanism for such reporting in the event a state chooses to require or allow it, or whether EPA is suggesting that all source reporting under Part 75 be required to report this information. UARG does not oppose a requirement for preparation of a monitoring plan for systems to measure and record CO2 mass emissions, or submission of that plan to ECMPS. The Part 72 “designated representative” (“DR”) for modified and reconstructed EGUs already will have to develop and submit such plans under Part 75 that include that information. UARG also supports reliance on Part 75 QA/QC. However, UARG does not believe that there is any benefit to including equipment used to measure electric and energy output in the Part 75 monitoring plan or to requiring additional QA/QC for that equipment. EGUs already have sufficient incentives to ensure that their equipment for measuring output is accurate, and EPA has not provided any 282 information to suggest that additional quality assurance is warranted. UARG also believes that any increase in accuracy would be dwarfed by other sources of allowed error in the overall measurements that cannot be eliminated. On the other hand, the resources necessary to track such equipment, to purchase and install new equipment if necessary to meet new accuracy standards, and to perform such testing would be significant. UARG also does not understand EPA’s proposed reference to compliance with Part 75, Appendix D for units using CEMS. Proposed Subpart UUUU § 60.5805(a)(1), (2)(ii). Under Part 75, CEMS certification requirements are in Appendix A, and ongoing QA/QC requirements are in Appendix B. Appendix D is the procedure for calculating heat input and mass emissions of SO2 for units without CEMS. EPA should replace the reference to Appendices B and D with “Appendices A and B.” E. Use of Specific Methods for Flow RATAs and Baseline Adjustments Following a Change in Method EPA proposes to require that if an EGU chooses to use Method 2 to perform the required relative accuracy test audit (“RATA”) on a flow monitoring system, the EGU must use a calibrated Type-S pitot tube (rather that the default Type-S pitot). Proposed Subpart UUUU § 60.5805(a)(2)(v). UARG does not oppose such a requirement. However, EPA also solicits comment on requiring use of what it calls “the most accurate RATA reference method for specific stack configurations” when performing tests on stack gas flow monitors, and use of a “computation[al] adjustment” when an EGU changes RATA reference methods. 79 Fed. Reg. at 34,913-14. Approved methods for flow RATAs under Part 75 include Methods 2F, 2G, 2H, and conditional test method (“CTM”) 041. Method 2F uses a three dimensional (“3-D”) probe to determine yaw angle, pitch angle, axial velocity, and volumetric stack flow. Method 2F was 283 developed to eliminate high bias of stack flow measurements in stacks with cyclonic flow. Method 2H and CTM-041 provide procedures for adjusting stack flow measurements to correct for velocity decay near a stack or duct wall that is not present in the measurements taken elsewhere in the stack. 68 In the Part 75 GHG Guideline TDS at 7-8, EPA expresses concerns that allowing sources to use these alternative methods (and in particular allowing EGUs to change methods over time) could lead to inconsistencies between emissions or heat input values. 69 To avoid this result, EPA solicits comment on whether it should require use of the more accurate methods, and whether it should develop adjustment factors for normalizing data when an EGU opts to use a different reference method to calibrate its stack flow monitor during a RATA. Id.; 79 Fed. Reg. at 34,913-14. Although EPA must allow use of Methods 2F and 2H (or CTM-041), EPA should not require either one. First, not all EGUs will obtain more accurate measurements using these methods. EGUs with axial flow and smooth stack liners may not experience any improvement in flow measurements from those methods, which are designed to correct for non-axial flow and wall effects caused by friction. Although the methods can provide more accurate (and lower) stack flow measurements for other EGUs, they also impose additional burdens. Any EGU concerned about overestimation of measured flow due to cyclonic flow conditions, or wall effects, can opt to use the methods. There is no basis to require their use. EPA also should allow 68 Methods 2F and 2H, which address stacks and round ducts, are codified at 40 C.F.R. Part 60, Appendix A-1 and A-2, respectively. CTM-041 addresses wall effects in square ducts. Although EPA has proposed to revise Method 2H to reflect the CTM-041 procedure for square ducts, 74 Fed. Reg. 42,819 (Aug. 25, 2009), EPA has not yet finalized it. 69 In Section XIV.A, UARG discusses how changes in the manner of calibrating flow monitoring systems and other sources of variability in heat input measurements can result in changes in reported heat rate that EPA erroneously concluded were the result of purposeful efforts to improve unit efficiency. 284 use of Method 2G. Method 2G uses a 2-dimensional probe to measure yaw angle (but not pitch) and near-axial velocity. Method 2G also can provide more accurate (i.e., lower) flow measurements than Method 2 under some conditions. 40 C.F.R. pt. 60, App. A-2. Method 2G is often performed using an auto-probe, which can reduce testing time. Part 75 allows use of Methods 2, 2F, 2G, 2H, and CTM-041. Id. pt. 75, App. A § 6.5.10. 70 EGUs should be allowed to choose which version of the flow method is best for the particular application and should not be required to perform 3-D testing where it is not needed. Regarding EPA’s suggestion that it develop “adjustment factors” that would be applied if an EGU changed flow methods, EPA has not provided sufficient information for UARG to comment on the reasonableness of such a requirement. EPA’s example – that a unit that transitions from Method 2 to Method 2H when performing flow RATAs would apply a “percentage reduction of baseline data” – is unclear. Part 75 GHG Guideline TDS at 7-8. Specifically, it is not clear what data EPA is suggesting would be altered – the baseline data used in setting the state goal or data used to determine compliance under an approved state plan. EPA’s suggestion that data would be “reduced” by some percentage suggests that EPA’s reference is to baseline data. But would that require recalculation of the state goal every time an EGU changed flow RATA methods? The altering of data collected under Part 75 could have serious consequences and UARG certainly would object to use of such altered data for any other program. EPA also does not provide a timeframe for this requirement. If EPA is suggesting that the compliance data – not baseline data – be altered, there would be no adjustment needed for data collected prior to implementation of the state plan, or for changes in methods made before 70 Currently, to use CTM-041, EGUs must submit a one-time written request to EPA. See EPA, Rectangular Duct Wall Effects, http://www.epa.gov/airmarkets/emissions/rect-wallducts.html. 285 2012. Finally, EPA has not provided any information to inform how it might calculate an adjustment factor. Since the impact of a change in flow methods would be site specific, EPA presumably would have to require a comparison of flow data collected with the CEMS before and after the RATA. But that is hardly a concept that is sufficiently developed to allow for meaningful content. In short, the two sentences of vague conceptual language in the TSD is not sufficient to inform commenters of the scope or impact of EPA’s proposal. If EPA intends to pursue such a requirement, EPA must issue a rulemaking proposal that fully explains what kind of adjustment EPA is proposing and specifically how the percentage value would be determined. F. EGU Recordkeeping Requirement EPA proposes to require that EGUs maintain records for at least 10 years following the date of each occurrence, measurement, maintenance, corrective action, report, or record, and that those records be kept “on site” for at least 2 years. Proposed Subpart UUUU § 60.5805(b). UARG opposes the proposed 10 year record retention period for all of the supporting information, unless EPA can show a special need as required under the Paperwork Reduction Act regulations. 5 C.F.R. § 1320.5(d)(2)(iv). 71 To the extent EPA or a state intends to audit the accuracy of a certified report by reference to supporting information, EPA should do so within a reasonable period of time (e.g., 3 years). Absent an audit, the data have little or no usefulness. Part 75 has a 3-year record retention requirement, and EPA has not provided any information to suggest that has been inadequate. 40 C.F.R. § 75.57(a). Given EPA’s proposed reliance on Part 75 monitoring, compliance with Part 75 recordkeeping should be sufficient for approval of a state plan. 71 UARG could find no such justification in EPA’s Proposed ICR Supporting Statement. 286 UARG also opposes the proposed on-site requirement if it means that retention of electronic records that can be accessed “on site” (even if they are physically located offsite) would not be sufficient. Many of the records to be retained are recorded only in electronic form and, depending upon the facility’s management practices, may be retained at a central location in order to facilitate backups and other services. As long as the facility can access the information on site, there is no reason for it to be physically located there. XVII. Objection to Promulgation of Rules Not Proposed NSPS rulemakings are subject to section 307(d) of the CAA. CAA § 307(d)(1)(C). Among other requirements, the rulemaking proposal must “be accompanied by a statement of its basis and purpose,” that includes: “(A) the factual data on which the proposed rule is based; (B) the methodology used in obtaining the data and in analyzing the data; and (C) the major legal interpretations and policy considerations underlying the proposed rule.” Id. § 307(d)(3). Further, all of those required “data, information, and documents” must be “included in the docket on the date of publication of the proposed rule.” Id. Section 307(d) also establishes requirements for final rules. Among other things, the promulgated rule must be accompanied by (1) a statement of basis and purpose like that required for the proposed rule, (2) an explanation of the reasons for any major changes in the promulgated rule, and (3) a response to each of the significant comments submitted during the comment period. Id. § 307(d)(6)(A), (B). These requirements are fundamental to both reasoned decision making and judicial review. Small Refiner, 705 F.2d at 519 (the fundamental purpose of section 307(d)(3) is to require that “EPA…give a detailed explanation of its reasoning at the ‘proposed rule’ stage” of the rulemaking process); Home Box Office, Inc. v. FCC, 567 F.2d 9, 35 (D.C. Cir. 1977) (per curiam) (without a response to significant comment, the “opportunity to comment is meaningless”). 287 Consistent with these key principles, UARG objects to promulgation of any final rule the content of which UARG could not reasonably have anticipated based on the proposed rule, and to any failure by the Agency to adequately respond to UARG’s comments. In particular, UARG notes the absence from this rulemaking docket of 21 of the 25 IPM modeling runs that EPA performed. EPA relied on these modeling runs in establishing the Proposed Guidelines, and their absence from the rulemaking docket has prevented meaningful comment on the Proposed Guidelines, in violation of section 307(d) of the CAA. 288