




Denial of Petitions for Reconsideration of Certain Startup/Shutdown Issues: MATS 
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                     U.S. Environmental Protection Agency
                 Office of Air Quality Planning and Standards
                     Sector Policies and Programs Division
                 Research Triangle Park, North Carolina 27711
                                       
                                   July 2016
                      This page intentionally left blank.

                                  Background

On February 16, 2012, pursuant to sections 111 and 112 of the Clean Air Act (CAA), the U.S. Environmental Protection Agency (EPA) published the final rules titled "National Emission Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units" (77 FR 9304). The National Emission Standards for Hazardous Air Pollutants (NESHAP) rule issued pursuant to CAA section 112 is referred to as the Mercury and Air Toxics Standards (MATS), and the New Source Performance Standards (NSPS) rule issued pursuant to CAA section 111 is referred to as the Utility NSPS. Following promulgation of the final rules, the Administrator received petitions for reconsideration of numerous provisions of both MATS and the Utility NSPS pursuant to CAA section 307(d)(7)(B). The Administrator granted reconsideration of the new source standards in MATS, the requirements applicable to periods of startup and shutdown in MATS and the Utility NSPS, and some definitional and monitoring provisions in the Utility NSPS on November 30, 2012.

The EPA took final action on reconsideration in two separate notices. On April 24, 2013, the EPA issued the final action on reconsideration of the new source MATS standards, the definitional and monitoring provisions in the Utility NSPS, and the technical corrections in both rules. 78 FR 24073. The EPA issued the final action on reconsideration of the startup and shutdown provisions in MATS and the Utility NSPS on November 19, 2014. 79 FR 68777.

Following promulgation of the final action on reconsideration of the startup and shutdown provisions in the MATS and Utility NSPS, the Administrator received two petitions for reconsideration of portions of that November 2014 final rule pursuant to CAA section 307(d)(7)(B). The EPA carefully reviewed the petitions and evaluated each issue raised in the petitions for reconsideration to determine if they meet the CAA section 307(d)(7)(B) criteria for reconsideration. In this action, the EPA is denying all the issues in the petitions for reconsideration because they do not meet the criteria for reconsideration. 


                     CAA Section 307(d)(7)(B) Requirements

Section 307(d)(7)(B) of the CAA states that "[o]nly an objection to a rule or procedure which was raised with reasonable specificity during the period for public comment (including any public hearing) may be raised during judicial review. If the person raising an objection can demonstrate to the Administrator that it was impracticable to raise such objection within such time or if the grounds for such objection arose after the period for public comment (but within the time specified for judicial review) and if such objection is of central relevance to the outcome of the rule, the Administrator shall convene a proceeding for reconsideration of the rule and provide the same procedural rights as would have been afforded had the information been available at the time the rule was proposed."

As to the first procedural criterion for reconsideration, a petitioner must show why the issue could not have been presented during the comment period, either because it was impractical to raise the issue during that time or because the grounds for the issue arose after the period for public comment (but within 60 days of publication of the final action). It is well established that an agency may refine its proposed approach and respond to comments received without providing an additional opportunity for public comment. See Community Nutrition Institute v. Block, 749 F.2d at 58 and International Fabricare Institute v. EPA, 972 F.2d 384, 399 (D.C. Cir. 1992) (notice and comment is not intended to result in "interminable back-and-forth[,]" nor is an agency required to provide additional opportunity to comment on its response to comments) and Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d 506, 547 (D.C. Cir. 1983) ("notice requirement should not force an agency endlessly to repropose a rule because of minor changes"). In addition, inherent in any proposed revision is the possibility that the agency will decline to make the change. Environmental Integrity Project v. EPA, 425 F.3d 992, 997 (D.C. Cir. 2005) ("Of course, there is nothing objectionable in the Agency's refusal to adopt its proposed amendments to Part 70's text. See Am. Iron & Steel Inst. v. EPA, 280 U.S. App. D.C. 373, 886 F.2d 390, 400 (D.C. Cir. 1989) (`One logical outgrowth of a proposal is surely, as EPA says, to refrain from taking the proposed step.')"). 

In the EPA's view, an objection is of central relevance to the outcome of the rule only if it provides substantial support for the argument that the promulgated regulation should be revised. See, e.g., the EPA's "Denial of the Petitions to Reconsider the Endangerment and Cause or Contribute Findings for Greenhouse Gases under Section 202 of the Clean Air Act," 75 FR 49556, 49561 (August 13, 2010). See also, 75 FR 49556, 49560-49563 (August 13, 2010) and 76 FR 4780, 4786-4788 (January 26, 2011) for additional discussion of standard for reconsideration under CAA section 307(d)(7)(B).




                 MATS STARTUP/SHUTDOWN RECONSIDERATION ISSUES

I. Legal Issues

Issue 1: Environmental Petitioners maintain that the EPA must reconsider the determination that a section 112(h) work practice standard for periods of startup and shutdown is authorized in lieu of a numeric standards for coal- and oil-fired EGUs. The petitioners maintain that they did not have the opportunity to comment on the EPA final determination that the end of startup (i.e., the time when the final numeric standards apply) begins 4 hours after initial generation, and the petitioners maintain that emissions can be accurately measured for the first 4 hours after generation. The petitioners argue that because the EPA did not indicate in the proposed reconsideration rule that the agency is unable to accurately measure emissions in the first hours after generation, that the EPA must reconsider the determination that a section 112(h) work practice is authorized for periods of startup and shutdown for EGUs. 

Response 1: The EPA is denying reconsideration on this issue. The public, including the petitioners, had ample opportunity to comment on the EPA's legal authority to establish a section 112(h) work practice standard in lieu of numeric emission standards for EGU startup and shutdown periods. The EPA provided an opportunity to comment on the issue during the comment period on the November 30, 2012, notice of proposed rulemaking. Petitioners, in fact, did comment that the EPA did not have the authority to establish a work practice standard for EGU startup and shutdown periods. However, they raised this issue late in the process and only after the EPA issued the startup and shutdown technical support document (TSD) and reopened the comment period on the November proposal. The EPA responded in full to these comments. See National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units Reconsideration of Certain Startup/Shutdown Issues; Summary of Public Comments and Responses (SS RTC), Pages 79-82; see also Reconsideration of Certain Startup/Shutdown Issues: National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units, 79 FR 68777, footnote 1 (2014 Startup/Shutdown Final Rule). Because the petition fails to demonstrate it was impracticable to comment on the issue during the public comment period, we are denying the petition for reconsideration with respect to this issue.  

As stated above, the public had ample opportunity to comment on the EPA's authority to establish a work practice for EGU startup and shutdown periods. On May 3, 2011, the EPA published the MATS Proposal. The proposed rule required compliance with the numerical standards at all times, including periods of startup and shutdown. The EPA explained that there was sufficient variability in the numerical standards to account for startup and shutdown periods and that the use of clean fuels during those periods would limit hazardous air pollutants (HAP) emissions. See National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units; Proposed Rule, 76 FR 23768, 25028 (2011 MATS Proposal).
The EPA received comments on the proposed rule suggesting that the EPA establish a work practice during startup and shutdown periods pursuant to section 112(h). The commenters cited section 112(h) and noted that HAP emissions cannot be measured during startup and shutdown periods and that it is not feasible to prescribe or enforce an emission standard for such periods. Other commenters noted that stack testing and monitoring for HAP emissions during startup/shutdown are not feasible or demonstrated due in part to flue gas conditions and the testing time (e.g., EPA Reference Method 5 requires 6 to 8 hours to complete a single test, but the conditions would have changed so significantly it would be impossible to understand what the data meant or extrapolate the data).

On February 16, 2012, the final MATS was published in the Federal Register. See National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units, 77 FR 9304 (2012 MATS Final Rule). The EPA noted the concerns with the feasibility of collecting HAP data during startup and shutdown and the concerns regarding the reliability of measurements obtained from EGUs during such periods. See 2012 MATS Final Rule, 77 FR at 9381. The agency also noted that despite the considerable data received in response to the EGU Information Collection Request, the EPA received almost no HAP data for startup and shutdown periods and almost all of the data failed to meet our data quality requirements. The EPA concluded that it did not have sufficient data on HAP emissions that occur during startup and shutdown on which to set emission standards; therefore, the agency established work practice standards rather than numeric emissions standards for periods of startup and shutdown in the final rule. Id. 

The EPA received multiple petitions for reconsideration of the final work practice for startup and shutdown periods from industry. The industry petitioners did not question the EPA's authority to establish a work practice standard during startup and shutdown periods, but they asserted reconsideration was appropriate because they did not have an opportunity to comment on the final definitions of startup and shutdown and the work practice standards applicable during such periods. See EPA-HQ-OAR-2009-0234-20179 and EPA-HQ-OAR-2009-0234-20183. Several environmental organizations submitted a petition for reconsideration of the 2012 MATS Final Rule, but they did not seek reconsideration of the final work practice or question the agency's authority to establish a work practice standard for EGUs during periods of startup and shutdown. See Petition for Reconsideration filed by Chesapeake Climate Action Network, Conservation Law Foundation, Environmental Integrity Project and Sierra Club, EPA-HQ-OAR-2009-0234-20187 (April 16, 2012). The EPA granted reconsideration to provide an opportunity for public comment on the final work practice standards for EGU startup and shutdown periods.  

On November 30, 2012, the EPA issued a notice of proposed rulemaking in which the EPA announced it was reconsidering and providing an opportunity for comment on the new source standards and the startup and shutdown work practice standards and associated definitions that were established in the 2012 MATS Final Rule. See Reconsideration of Certain New Source and Startup/Shutdown Issues: National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units, 77 FR 71323 (2012 MATS Startup/Shutdown Proposal). The EPA received comments on this reconsideration proposal, including comments from the petitioners, but no commenters questioned the EPA's authority to establish a section 112(h) work practice standard for periods of startup and shutdown at EGUs. See EPA-HQ-OAR-2009-0234-20305, 20427, and 20438. See also EPA-HQ-OAR-2009-0234-20422 and 20436. 

After reviewing the comments on the reconsideration proposal, the EPA conducted additional analysis to determine the appropriate work practice for EGU startup and shutdown periods. The EPA provided notice of the new analysis and reopened the public comment period on June 25, 2013. See Reconsideration of Certain Startup/Shutdown Issues: National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-Fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units,78 FR 38001 (2013  Reopening Notice) ; see also EPA-HQ-OAR-2009-0234-20378 (Startup and Shutdown Technical Support Document). In response to the reopening, environmental commenters asserted for the first time that a work practice for EGU startup and shutdown periods is unlawful because the commenters argued that the EPA could not conclude numeric emissions standards are infeasible during startup. See National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-Fired Electric Utility Steam Generating Units  -  Reconsideration of Certain Startup/Shutdown Issues; Summary of Public Comments and Responses (SS RTC), EPA-HQ-OAR-2009-0234-20447, Comment 82 on page 80. Commenters argued among other things that CAA section 112(h) allows the EPA to substitute a work practice standard for a numeric emissions standard only where it is not feasible to prescribe or enforce an emission standard. Commenters further asserted that air pollution control devices (APCD) are readily available to capture each pollutant and there are technically and economically feasible measurement methodologies available to measure those pollutants during startup and shutdown. The commenters concluded that the EPA needs to explain how the specific work practice standards satisfy the requirements of CAA section 112(h). On November 19, 2014, the EPA published the final startup and shutdown requirements and responded to the comments raised by the Environmental Petitioners in the SS RTC. See 2014 Startup/Shutdown Final Rule, 79 FR 68777; see also SS RTC at EPA-HQ-OAR-2009-0234-20447. 

As the above background demonstrates, the petitioners had multiple opportunities to challenge the EPA's authority to establish a section 112(h) work practice and, when the petitioners eventually submitted comments on the issue, the EPA responded to the comments. Had the petitioners raised their concern with the agency's authority in a petition for reconsideration of the 2012 MATS Final Rule or in comments on the November 2012 reconsideration proposal, the EPA would have responded earlier. 

Moreover, in the petition for reconsideration at issue here, the petitioners do not assert that it is feasible to measure HAP emissions from EGUs during the entire startup and shutdown period such that a work practice is completely barred; instead, the petitioners assert only that HAP emissions can be measured feasibly before the end of the startup period as defined in paragraph 2 of the 2014 final Startup/Shutdown rule. Thus, the petitioners appear to be challenging only the technical and factual conclusions in support of the final standards. As explained below, the public had ample opportunity to comment on the analysis in support of the startup and shutdown work practices and comments were in fact submitted on that analysis. In addition, the petitioners provided no information to support the argument that HAP emissions can be accurately measured during startup and shutdown periods. The petitioners thus have not presented any information that supports a change in the EPA's legal or factual conclusions. As such, the information is also not of central relevance to the rule. 

For all of these reasons, the EPA is denying reconsideration on this issue.


Issue 2:  Environmental Petitioners assert that they satisfy the requirements for reconsideration under CAA section 307(d)(7)(B) because they assert they could not have raised certain specific issues during the public comment period. The petitioners also assert that they did raise the same issues generally in response to the EPA notices, specifically commented on the legal basis for establishing a section 112(h) standard, that a work practice must be consistent with section 112(d), and that section 112(d) emission standards must be met at all times. Based on the comments submitted during the comment period, the commenters assert that they have made their arguments with reasonable specificity such that they may obtain legal review of the issues.

The petitioners maintain that notwithstanding their comments on the proposed reconsideration, they did not have an opportunity to comment on several aspects of the final rule, including: the EPA's conclusion that emissions cannot be measured during the first 4 hours of generation; the manner in which the EPA determined what the petitioners term the work practice "floor"; the conclusion that the work practice should be based on the performance of the best performers in the category; and the alleged vagueness of the final standards. The petitioners state the objections are all of central relevance because they go to what the petitioners' term the EPA's exemption for startup and shutdown periods. The petitioners also state that the objections are of central relevance because they maintain EGUs emit considerable particulate matter (PM) emissions during startup and shutdown periods and the final rule does not sufficiently address those emissions. The petitioners also argue that the final startup and shutdown requirements are vague and that the final rule will not sufficiently limit PM emissions during those periods. As support, petitioners cite to existing permit limitations on PM emissions and concerns with those permits.

Response 2: Petitioners have not demonstrated that it was impracticable to raise the identified issues during the public comment period. In fact, the petitioners acknowledge that they raised several of these issues in comments on the proposed startup and shutdown reconsideration action. The petitioners' own assertions that they raised issues sufficiently to preserve judicial review, undermine their arguments that they were not provided sufficient opportunity to comment. The EPA acknowledged and responded to the comments received on this issue. See 2014 Startup/Shutdown Final Rule, 79 FR 68777, section IV. Summary of Final Action and Changes Since Proposal  -  MATS Startup/Shutdown Issue; see also Startup/Shutdown RTC. Because petitioners have not shown that it was impracticable to raise the issues during the public comment period, we are denying reconsideration.

The EPA directs the petitioners to the response to Issue 1 above for a response to the argument that the EPA must reconsider its determination that a work practice standard for EGU startup and shutdown periods is authorized pursuant to section 112(h). 

In addition, petitioners' objections are not of central relevance because they are based on irrelevant data. Petitioners' objection appears to be based on assertions regarding PM emissions. Specifically, they assert that EGUs current (non-section 112) permits do not ensure compliance with the applicable PM standards during SS periods. Even if correct, current noncompliance with non-section 112 PM emissions standards is irrelevant to whether the agency has the legal basis to establish a work practice during startup and shutdown periods or whether the MATS work practice will reduce HAP emissions (including PM emissions as a surrogate for non-mercury metal HAP) during SS periods. In addition, the final rule contains sufficient monitoring, recordkeeping and reporting requirements to ensure compliance with the final work practice and the sources title V permit will be revised to include the requirements of MATS. The petition thus does not provide any information that would support a conclusion that the MATS work practice is unenforceable. For this reason, the EPA finds that this issue is not of central relevance and does not warrant reconsideration. In addition, as the petitioners themselves note, section 112(h) standards are required to be based on the performance of the best controlled sources, and the final startup and shutdown work practice requirements are based on the performance of the best performing sources as discussed in the record. See e.g., 2014 Startup/Shutdown Final Rule, 79 FR at 68779-86. The petition does not demonstrate a flaw in the EPA approach to determining the best performers or provide an alternative approach. As such, the petition does not provide information that would support a change to the rule and, for this reason as well, the objection is not of central relevance to the rule.

For these reasons, the Environmental Petitioners have not demonstrated that reconsideration of the final startup and shutdown requirements is warranted under CAA section 307(d)(7)(B) and the agency is denying reconsideration. 


II. Measuring HAP Emissions During Startup Periods Issues

Issue 3:   Environmental Petitioners argue that the EPA must reconsider the final work practice because the EPA has not demonstrated that it is impracticable to measure HAP emissions during startup and shutdown periods. The petitioners argue that the EPA did not follow D.C. Circuit case law requiring agencies to base decisions on substantial evidence when establishing the final work practice. The petitioners maintain that the final MATS rule suggests that the EPA believes HAP emissions can be measured accurately after generation. The petitioners state that the EPA did not make the alleged change in its position until the agency reconsidered the work practice standards. The petitioners also state that the EPA's reliance on what the petitioners characterize as a conclusory statement that HAP emissions cannot be accurately measured until the controls are engaged is not sufficient to support the imposition of a work practice standard in lieu of numerical standards. The petitioners then acknowledge that they commented on the ability to measure HAP emissions during periods of startup and shutdown and argues that the EPA's conclusory response is not sufficient. The petitioners conclude that there is substantial evidence that contradicts the EPA's determination that emissions cannot be practicably measured until 4 hours after first generation and that without consideration of that evidence the EPA cannot make a rational decision. The petitioners point to the Acid Rain program (ARP) as an example in support of their position because in that rule the agency requires EGUs to continuously measure sulfur dioxide (SO2) emissions, even during startup and shutdown periods.

Response 3:  The EPA is denying reconsideration on these issues because the petitioners have not demonstrated that it was impracticable to comment on these issues or provided new information that was not available during the public comment period (but within the time specified for judicial review) to support reconsideration. We direct the petitioners to the response to Issue 1 for a response to the argument concerning the authority to establish a work practice standard. In addition, the petitioners acknowledge that they commented on the ability to accurately measure HAP emissions during the comment period and the EPA responded to the comment. See SS RTC at 78-79, 82. Thus, on these issues, it is clear that the petitioners failed to demonstrate that it was impracticable to comment during the comment period and reconsideration is denied for that reason. In addition, there is no basis for the petitioners' assertion that the response to those prior comments on HAP measurement issue was conclusory. As noted above, the EPA responded in full to comments asserting that HAP could be accurately measured during startup and shutdown periods. The comments on the proposed rule did not provide any specific information concerning approved HAP measurement methodologies to support their argument that HAP can be measured during those periods, and the EPA cannot respond to comments that were not submitted. The EPA's response to the commenters' general assertions was more than sufficient to address the comment that was submitted on the proposed rule.

As an initial matter, the petitioners' assertion that the EPA made a change in interpretation at final that mandates reconsideration is also incorrect. The EPA did not conclude in the 2012 MATS rule (or imply through finalization of a work practice for startup and shutdown periods) that HAP emissions could be accurately measured from the time of first generation. The measurement methodologies were not and are not certified to accurately measure HAP emissions during startup and shutdown periods as defined in the 2014 Startup/Shutdown Final Rule. In addition, the EPA established the final work practice in 2012 with limited input from the public and reconsidered the final approach because the petitions for reconsideration demonstrated that certain provisions would benefit from additional consideration. 

More relevant to this reconsideration petition is the fact that the petitioners commented on an aspect of this issue after the EPA issued the startup and shutdown TSD (SS TSD) and reopened the comment period. See EPA-HQ-OAR-2009-0234-20427, pages 14-15. In the SS TSD and reopening notice, the EPA proposed to establish an alternative definition of startup that would last beyond first generation of electricity. See 2013 MATS Reopening Notice, 78 FR 38001, 38005 and SS TSD. Because the 2013 notice reopening the comment period provided ample notice that the EPA was considering a different approach to the work practice, the petitioners had ample opportunity to comment on the issue and the EPA is denying reconsideration.  

Petitioners also commented on the alleged inconsistency between the conclusion that HAP emissions from EGUs cannot be measured during the defined startup and shutdown periods and the fact that the ARP required EGUs to measure SO2 emissions. See EPA-HQ-OAR-2009-0234-20427, page 15. The Petitioners do not acknowledge that the distinct measurement techniques necessary for an emissions trading program, such as ARP, established under title IV of the Clean Air Act Amendments (CAAA), which allows source owners or operators to purchase credits for emissions in excess of an annual threshold differ from those found in continuous emissions compliance programs, which contain never-to-be exceeded emission limits, such as those established under section 112 of title I of the CAA. Moreover, the SS TSD analyzed the ARP data to determine the end of startup based on when controls are engaged, not to assess the accuracy of emission measurement at the beginning of startup. Petitioners thus cannot demonstrate that they lacked an opportunity to comment on the alleged inconsistency. In addition, the petitioner's comment is misleading because it is not reasonable to compare the quality and accuracy of ARP data with the HAP emission measurements that are necessary under MATS. The ARP is an annual mass accounting program whereas MATS requires compliance with heat input- or electrical output-based emission rate limits (i.e., lb/mmBtu or lb/MWh) on a 30-boiler operating day (or 90 day) rolling average basis. In the case of the ARP data, units must measure SO2 concentration and stack gas flow rates to calculate total mass emissions. However, for MATS the units that elect to comply with the heat input-based standard must use the EPA Method 19 equation to calculate the emission rates. The equation is only applicable and, therefore, accurate, under conditions of complete combustion and non-startup periods. The emission rates using this type of equation will be grossly overstated at startup conditions because the measured carbon dioxide (CO2) or oxygen (O2) levels, which are an essential term in the EPA Method 19 equation, will be close to those of ambient air.  

For these reasons, the EPA is denying reconsideration on this issue.   


Issue 4: Environmental Petitioners argue that, at least for SO2, the EPA's determination that emissions are not measurable during the 4 hours after generation is directly contradicted by the monitoring provisions of the ARP and the EPA's position on the accuracy of startup emissions measured as part of the ARP. The petitioners comment that in determining the appropriate time after the beginning of generation when the MATS Rule's numeric limits would apply, the EPA evaluated continuous SO2 data from startup events submitted by power plants under the ARP. In discussing these data from the ARP, the EPA specifically noted that "these data are not reliable for quantifying emissions for this analysis but, rather, the data allow us to evaluate when controls are turned on for the purpose of determining when startup ends." Id. at 68,780 f.n.6 (emphasis added). 

The petitioners argue that despite its statements in the MATS rulemaking, the EPA took the opposite position in the ARP regarding the accuracy of measuring SO2 emissions during startup and shutdown. Further, the petitioners argue that under the ARP, any coal-fired power plant must use continuous emission monitoring systems (CEMS) and are required to report emissions data for every hour that a power plant is operating, including periods of startup, shutdown and malfunction. The petitioners state that the ARP counts these emissions during these periods in determining whether a plant has stayed within its annual limit for tons of SO2 emitted. The petitioners cite numerous publications available on the EPA's website where the EPA has affirmed that the SO2 emissions data reported by power plants under the ARP, which includes data covering emissions during startup and shutdown, are complete and accurate. 

The petitioners argue that the EPA has taken two contradictory positions regarding the accuracy of SO2 monitoring during startup and that the EPA has made no attempt to explain its conflicting regulations or positions regarding the accuracy of SO2 monitoring during startup.

Response 4: The EPA is denying reconsideration on the issue raised because the petitioners have not demonstrated that it was impracticable to comment on this issue or provided new information that was not available during the public comment period (but within the time specified for judicial review) to support reconsideration. We direct the petitioners to the response to Issue 3 for a response to the argument concerning the alleged inconsistency between the conclusion that HAP emissions from EGUs cannot be measured during the defined startup and shutdown periods and the fact that the ARP requires EGUs to measure SO2 emissions. In addition, the petitioners acknowledge that they commented on the ability to accurately measure HAP emissions during the comment period and the EPA responded to the comment. See SS RTC at 78-79, 82. Thus, on this issue, the petitioners failed to demonstrate that it was impracticable to comment during the comment period and reconsideration is denied for that reason. 

Moreover, the EPA finds that the petitioners also could reasonably have commented on this issue after the EPA issued the SS TSD and reopened the comment period. In the SS TSD, the EPA proposed to establish an alternative definition of startup based on initiation of control devices for SO2 and nitrogen oxide (NOX) based on data from the EPA's Clean Air Markets Database; these data are contained in the monitoring data reported by EGUs as required under the ARP. The 2013 Reopening Notice provided ample notice that the EPA was planning to use SO2 and NOX monitoring data obtained under the ARP. The notice explicitly requested comment on the technical analysis conducted in response to the already received comments and explained that the analysis was in the docket for the rulemaking. The petitioners thus had ample opportunity to comment on the issue and the EPA is denying reconsideration for that reason as well.  

The petitioners' argument that the EPA has affirmed, in publications on its website that SO2 emission data reported by EGUs under the ARP, including periods of startup and shutdown, are complete and accurate is misleading. As the petitioners stated, under the ARP, EGUs must monitor SO2 emissions during all operating periods to determine whether an EGU has stayed within its annual limit for tons of SO2 emitted. However, the vast majority of hourly readings reported by EGUs are from operation outside of startup and shutdown periods and the vast majority of emissions measured (over 99 percent) are measured at times other than startup and shutdown. Any inaccuracies in SO2 hourly readings from periods of startup and shutdown will thus have a negligible effect on an EGU's annual SO2 emissions and moreover, sources covered by the ARP may purchase allowances from other sources to cover any emissions in excess of their annual allotment. Conversely, MATS requires continuous compliance with the numeric standards and sources do not have the option of purchasing allowances to cover emissions. Thus, the potential consequences associated with, for example, monitoring equipment inaccurately measuring SO2 levels in excess of the 30-boiler operating day rolling average alternative equivalent surrogate SO2 limit in MATS, are more serious. Furthermore, a statement cited by the petitioners concerning the Transport Rule relying on SO2 data from the ARP, which the petitioners state professes the EPA's belief that the SO2 data are accurate, actually states that "this approach uses the most accurate data currently available to the EPA." This implies that no better data may be available, not that the existing data are similarly accurate throughout all startup or shutdown periods. This data description also implies there may be some uncertainty in the data, and the EPA acknowledges there is greater uncertainty in emissions measurements taken during the startup period as defined in the final MATS SS rule. In addition, the SO2 limit in MATS is an alternative equivalent surrogate standard for acid gas HAP, and the petitioners have not provided any support for a conclusion that the acid gas or other HAP can be measured accurately during the defined startup and shutdown periods. For this reason, even if the EPA determined the SO2 emissions measurements were sufficiently reliable for MATS compliance purposes, the EPA would decline to establish a numeric standard for that one alternative standard until such time as the actual HAP emissions could also be reliably measured so that numeric standards could be established. It is in part because of a desire to collect additional information and data concerning startup and shutdown HAP emissions that the agency established the alternative work practice in this final rule. See 2014 Startup/Shutdown Rule, 79 FR 68779, footnote 3. For these reasons, the objection is not of central relevance to the outcome of the rule. 

For these reasons, the EPA is denying reconsideration on this issue.


Issue 5: Environmental Petitioners argue that the EPA's failure to explain alleged internal inconsistencies in the MATS rule that startup emissions cannot be accurately measured warrants reconsideration. The petitioners cite, for example, that the 4-hour period is optional and EGUs can choose to comply with the MATS numerical standards from the point of generation. Similarly, under MATS, units can choose to monitor mercury emissions through sorbent traps. The petitioners state that although the EPA indicates in the response to comments that accurately measuring mercury emissions with sorbent traps during startup and shutdown "may be problematic", one option for sorbent-trap monitoring allows EGUs to count startup and shutdown emissions in determining compliance with the MATS mercury limit. The petitioners argue that it is internally inconsistent to allow units to show compliance based on emissions from the first 4 hours after generation and simultaneously claim it is not feasible to reliably measure emissions during these 4 hours. The only explanation the EPA gives for having both definitions of startup they assert is that the EPA believes that both "meet the requirements of CAA section 112 to reduce HAP emission during this time period and will provide operators with flexibility, even though we question the ability to accurately measure HAP emissions at the start of electricity generation." The petitioners comment that either emissions can be measured accurately enough to determine compliance once generation begins, or they cannot. If not, then the EPA has no basis for allowing any EGU to elect to use this supposedly inaccurate data to determine compliance. 

The petitioners add that even though the EPA stated that emissions cannot be accurately measured during startup, the EPA used SO2 and NOX emissions from startup (reported as part of the ARP) in its floor analysis to purportedly determine when the best-performing sources engage the last of their pollution controls. The EPA claims that the data it used "are not reliable for quantifying emissions for this analysis but, rather, the data allow us to evaluate when controls are turned on . . . ."  The petitioners further comment that the EPA's reasoning is also internally inconsistent because the EPA takes the position that it cannot be sure that stack conditions are stable enough to allow for accurate measurement of emissions until controls are on, but the EPA's work practices do not actually require non-PM controls to be operated. Petitioners question why, if the EPA cannot be sure that emissions can be accurately measured until controls are operational, the EPA fails to require that these controls be operated. They assert that the EPA fails to explain this inconsistency.

Response 5: The EPA is denying reconsideration on these issues because the petitioners have not demonstrated that it was impracticable to comment on these issues during the comment period. The issue of the option for sorbent trap monitoring allowing EGUs to count startup and shutdown emissions in determining compliance with the MATS mercury limit was raised in public comments submitted in response to the proposed reconsideration rule. The EPA responses to the comments are in the SS RTC (see pages 180  -  182). The comments the EPA received on this issue demonstrate that the public had ample opportunity to comment on the issue, and the EPA responded to those comments. Because the petitioners have not demonstrated that it was impracticable to comment on the alleged inconsistency concerning sorbent trap monitoring during the comment period on the proposed rule, the EPA is denying reconsideration of the issue. 

The EPA also does not agree that the ability to measure HAP emissions accurately is as clear as the petitioners suggest or that the EPA acted unreasonably by allowing sources to rely on data collected during startup periods for purposes of determining compliance. Concerning the ability to measure emissions, the petitioners are correct that CEMS and sorbent trap monitoring systems will collect data during startup and shutdown periods: the rule acknowledges this and requires CEMS data to be collected (see 63.10007(a)(1) and 63.10011(g)(2)) but not used to report emissions (see 63.10020(c)). However, that does not mean the measurement methodologies are designed to measure HAP emissions during startup and shutdown periods with a sufficient degree of accuracy or that the stack conditions are conducive to measuring HAP emissions accurately with the current technology. In fact, stack conditions do not allow accurate measurement of HAP emissions during these periods, and the petitioners have provided no evidence to the contrary. The EPA provided the alternative work practice to address the concerns with measurement accuracy and maintained the option from the final MATS rule for the EGU owners or operators who found the reduced compliance assurance burdens of that option a sufficient motivation to comply with the more rigid timelines for demonstrating compliance with the numerical standards. The EPA continues to believe these options provide comparable HAP reductions and that the EGUs that choose to comply with the original option will be those that conclude based on the specific factors at the facility (e.g., coal testing, control efficiency, pre-combustion practices) that compliance with the 30-day average will be demonstrated even with an unexpected spike in emissions during startup or shutdown. Petitioners have not identified an unreasonable inconsistency in the final rule or provided information that would support a change in the final rule. The objection thus is not of central relevance to the rule and the EPA is denying reconsideration for this reason also. 

The EPA also does not agree that it is unreasonable to allow an EGU owner or operator to choose to comply with a monitoring mechanism designed to measure emissions over a 30-day period when the mechanism has a greater degree of uncertainty for a short period of time. In addition, given the uncertainty associated with measurement accuracy during the first 4 hours after generation, it is reasonable to redefine the end of startup for the alternative option so that an EGU owner or operator is not required to rely on data with a greater degree of uncertainty than the data collected at times other than startup and shutdown periods (e.g., steady state operations). As noted above, however, both CEMS and sorbent traps will measure emissions during startup and shutdown periods. CEMS will measure emissions concentration as long as the values remain within the calibration range, and sorbent traps will capture total mercury mass, allowing for discrete, not continuous, measurement. Therefore, to the degree allowed by the stack conditions, emissions will be measured during the first 4 hours of generation even if the accuracy is of greater uncertainty, and EGU owners or operators who choose the option will be required to consider the data when determining compliance for any 30-day period including startup data. Thus, EGU owners or operators who choose the option that relies on the data from the first 4 hours may not use alleged uncertainty of the data from that time as an excuse for noncompliance. In addition, because the startup and shutdown time is limited to 4 hours after first generation and the record does not support a conclusion that EGUs have more than an average of 1 startup per month, the EPA believes the use of less certain data in this case is not statistically significant such that it undermines the reasonableness of including the data within the 30-day average at the sources that choose this compliance option. 

In regards to the use of SO2 and NOX emission data in the technical analysis conducted, we stated in the SS RTC that we established a work practice for periods of startup and shutdown because we did not have data on HAP emissions from those periods because accurate HAP data from those periods is unlikely to be available from either emissions testing (which is designed for periods of steady state operation) or CEMS (which are designed for measurements occurring during periods other than during startup or shutdown when emissions flow are stable and consistent) (see SS RTC pages 48-49). We further stated in the SS RTC that even though the CEMS data may not accurately reflect HAP emissions during periods of startup or shutdown, the data are useful in showing emission trends relative to emission control device operation for determining the end of startup. The EPA did not rely on the actual concentrations of SO2 and NOX when determining the end of startup in either the SS TSD or the revised SS TSD. As stated in the response to Issue 4, the EPA finds that the petitioners could reasonably have commented on this issue after the EPA issued the SS TSD and reopened the comment period. In the SS TSD, the EPA proposed to establish an alternative definition of startup based on initiation of control devices for SO2 and NOX, based on data from EPA's Clean Air Markets Database. Because the 2013 Reopening Notice provided ample notice that the EPA was planning to use SO2 and NOX monitoring data obtained under the ARP, the petitioners had ample opportunity to comment on the issue. Because the petitioners have not demonstrated that it was impracticable to comment on this issue of the use of SO2 and NOX emission data in the technical analysis conducted, the EPA is denying reconsideration on the issue. 

The EPA also does not agree that the requirement to turn on controls at the end of the work practice without requiring a specific control efficiency is inconsistent with the agency's general approach to regulation of HAP emissions, because the EPA generally leaves it to the sources to determine when and how controls must be operated to assure compliance with the numerical standard. In addition, as noted elsewhere, the EPA did not rely on the actual levels of SO2 and NOX for the analysis in the TSD, but instead used the ARP data to determine the time at which the EPA could be reasonably certain that emission controls had begun to operate. The EPA believes that stable stack conditions are the critical factor for measuring HAP with the available methodologies, and the EPA finds that stack flow has begun to stabilize 4 hours after generation, as evidenced by CO2 values above the default cap. Such stabilization is an indicator of stable flow rates and steady state conditions, and these conditions, coupled with in calibration range CEMS values, will be sufficient to measure HAP emissions accurately 4 hours after generation, regardless of whether all the air pollution control devices are engaged. As we noted above, emissions testing or CEMS are designed for periods of steady state or stable operation (SS RTC pages 48-49). The EPA finds that the petitioners could reasonably have commented on this issue of non-PM controls not being required during the startup period after the EPA reopened the comment period. In the 2013 MATS Reopening Notice, the EPA proposed to revise the work practice standard based on the engagement time of non-PM controls. The proposed revised standards do not, however, require non-PM controls to start except when necessary to comply with MATS or any other applicable requirement (e.g., the ARP). See 2013 MATS Reopening Notice, 78 FR 38003. Because the 2013 Reopening Notice provided ample notice concerning the EPA's consideration of non-PM controls, the petitioners had ample opportunity to comment on the issue. Because the petitioners have not demonstrated that it was impracticable to comment on the non-PM controls during the comment period, we are denying reconsideration. In addition, because the petitioners do not accurately characterize the EPA work practice or the justification, the objection is not of central relevance to the rule and reconsideration is denied for that reason as well.   

For these reason, the EPA is denying reconsideration on these issues.


III. Technical Analysis of Startup Periods Issues

Issue 6: Environmental Petitioners argue that the EPA failed to provide adequate rationale for picking the thresholds used to determine when controls were operable. The petitioners comment that the EPA apparently determined when EGUs initiate their controls based on two tests - the 2-hour averaging period when units were within 110 percent of their annual average, and the 2-hour averaging period when emission rates declined by 20 percent below the 2-hour average maximum emission rate. The petitioners assert that EPA seems to assume that emissions cannot be accurately measured until controls achieve a certain consistent level of efficiency in removing pollutants. Petitioners assert that the EPA picked the emission-rate thresholds that determined when controls were engaged without any explanation as to why those thresholds were appropriate or important. Specifically, they argue the EPA failed to explain why these thresholds, as opposed to any other threshold, represent the point in time when controls are operable and conditions are steady enough to support accurate measurement. They assert that the EPA simply states: "Using an emission decline threshold of 20 percent, the results indicate that the average of the best performing 12 percent of EGUs initiate SO2 and NOX control within 4 hours after the start of electricity generation." "Assessment of startup period at coal-fired EGUs  - Revised" at 21. They argue this is not a sufficient rationale to support EPA's determination that emissions cannot be accurately measured until 4 hours after generation.

Response 6: The EPA is denying reconsideration because the petitioners have not demonstrated it was impracticable to comment on these issues during the public comment period. The proposed SS TSD discussed when emissions decline demonstrates the engagement of controls, and petitioners could have commented on the information presented therein. The EPA also does not agree with the petitioners' assertion that the EPA picked the emission-rate thresholds that determined when controls were engaged without any explanation. The EPA stated in the revised SS TSD that at this threshold the EGUs that comprised the best performing 12 percent included EGUs covering the range of EGU types regulated under MATS. See EPA-HQ-OAR-2009-0234-20451, page 21. If a lower threshold were selected, the EGUs comprising the best performing 12 percent would not represent all the types of EGUs covered by MATS. This approach was taken in part based on comments from the petitioners after the EPA reopened the comment period and issued the SS TSD. Specifically, the petitioners commented that the EPA should determine the end of startup based on the performance of the best performing 12 percent of EGUs. The petitioners have also not provided information supporting an alternative, much less superior, approach to identifying the best sources based on the available information or otherwise demonstrated that the approach in the revised SS TSD is unreasonable. Thus, the petitioners have not presented information that would support a change to the rule and thus have not demonstrated that this objection is of central relevance to the final rule. 

For these reasons, the EPA is denying reconsideration on this issue.


Issue 7:  Environmental Petitioners assert that the EPA failed to establish that PM emissions cannot be measured or reasonably estimated during the first 4 hours after generation begins. The petitioners state that EPA's decision to exempt EGUs from MATS numerical standards during the first 4 hours after generation is based on EPA's finding that emissions cannot be accurately measured during these periods. The petitioners assert that the EPA's statements are conclusory and ignore evidence that particulate emissions can be measured or reasonably estimated once generation begins. 

The petitioners claim that the EPA has said that PM CEMS cannot reliably measure particulates during the first 4 hours of generation, but that the EPA offers no specific data to support that conclusion. The petitioners assert further that the fact that MATS does not require compliance with CEMS makes no difference.

The petitioners also claim that in the technical "floor" analysis to determine the length of startup, the EPA acknowledged that subcritical EGUs, which account for most coal-fired units, reach loads that average a third to one half of their maximum capacity between the second and third hour of generation. The petitioners maintain that the agency has not explained why PM CEMS are incapable of measuring emissions at these load levels, which can occur during normal operations (e.g., during load swings). The petitioners also state that if particulates or other HAP surrogates cannot be measured accurately when load drops below half of rated capacity, the EPA must explain the basis for allowing EGUs (particularly those that have chosen to rely upon PM CEMS) to rely upon such data during low load conditions that may occur during normal operations.

The petitioners maintain that the EPA's assumptions about the variability of PM controls during startup appear to be based on a subset of electrostatic precipitators (ESP) that cannot achieve reliable control of particulates (or HAPs) during the first few hours of generation. Petitioners state that in previous comments, as of August 2013, at least 166 EGUs (based on EPA's ARP data) relied on baghouses (or fabric filters) to control particulates. See 8/26/13 Comments (ID No. EPA-HQ-OAR-2009-0234-20422) at 11. The petitioners also assert that baghouses typically achieve very high rates of particulate control (sufficient to achieve the MATS limit) well before the fifth hour of generation. The petitioners claim that the EPA has not explained why baghouses should not be the basis for determining whether conditions are stable and PM emissions are measurable during the first few hours of generation, rather than ESPs that are characterized by erratic or poor performance during these periods. The petitioners maintain that the EPA has apparently designed what the petitioners term a startup exemption at least in part to accommodate a subset of PM controls that cannot meet the MATS numerical standards at all times  --  instead of basing the limits on what the best-performing 12 percent can achieve, as the law requires. 

The petitioners also assert that the EPA has not evaluated whether reasonable methods of estimating emissions during the first 4 hours of generation could substitute for the direct monitoring of emissions. As an example, the petitioners maintain that the EPA could estimate emissions based on the operating parameters of PM controls that can then be correlated with the relative efficiency of such equipment in removing the particulates. The petitioners then state that the operating parameters could be combined with coal-feed rates and the ash content of fuel to generate emissions estimates. The petitioners state that the EPA has not hesitated to rely upon estimates in lieu of direct measurement in the past or, indeed, in the present. As an example, petitioners point to the December 19, 2014, notice announcing the EPA's intent to clarify various aspects of the MATS rule, wherein the EPA proposed to allow EGUs to estimate emissions during emergency bypass events for up to 2 percent of total operating time. The petitioners note that such estimates can be included in compliance determinations and used to establish an EGU's eligibility as a Low Emitting EGU (LEE). The petitioners argue that the proposal does not define how these estimates will be made, or why they should be considered accurate, given the varying fuel use and operating conditions that will arise during emergency upsets. The petitioners argue that the EPA must explain the basis for determining that unqualified estimates of emissions during emergency events are sufficient to determine compliance with 30-day MATS limits and establish eligibility for LEE status, while simultaneously deciding that the lack of accurate monitoring data means that EGUs do not have to begin complying with MATS limits until 4 hours after generation begins.

Response 7: The EPA is denying reconsideration on this issue. The public, including the petitioners, had ample opportunity to comment on the ability to accurately measure or estimate HAP and PM emissions during the first hours after generation based on the SS TSD that was provided for comment when the EPA reopened the comment period on the startup and shutdown work practice on June 25, 2013. As noted above, the EPA analyzed the ARP data in the SS TSD to determine how long after generation the agency could be reasonably certain HAP emissions could be measured with a reasonable degree of accuracy; thus, comments concerning the ability to accurately measure PM (with either stack tests or CEMS), the time to engage PM controls, and the ability to estimate emissions during generation all could have been raised during the comment period. In fact, some of the petitioners filed comments stating that the June 2013 proposal did not satisfy the requirements of section 112(h)(2), that the EPA should require all the controls to be engaged when EGUs fire any coal, and that baghouses work throughout the startup period; and the EPA responded to the comments. See EIP's August 26, 2013 comments, pages 9-11; see also SS RTC. For all these reasons, the petitioners have failed to demonstrate that it was impracticable to comment on these issues during the public comment and the EPA is thus denying reconsideration. 

In addition, the determination that HAP emissions cannot be reasonably accurately measured during the defined startup and shutdown period is supported by the record. During the comment period on the May 3, 2011, MATS proposal, the EPA received a number of comments stating that it is not feasible to obtain sufficient data for startup/shutdown periods and the commenters stated it is uncertain whether representative measurements could be obtained during startup/shutdown due to dynamic flue gas conditions and technical limitations with measurement technology, that  stack testing is impossible during startup due to the time required to conduct the test and the changing stack conditions during these periods, and that the stack conditions are also problematic for CEMS during startup. See EPA's Responses to Public Comments on EPA's National Emissions Standards for Hazardous Air Pollutants from Coal- and Oil-fired Electric Utility Steam Generating Units (RTC) at EPA-HQ-OAR-2009-0234-20126, pages 418-419). In addition, in the reconsideration action, the EPA noted that the stack conditions during startup and shutdown make the reliability of HAP measurements uncertain until stable stack conditions are achieved, that EGUs cannot reasonably conduct a stack test during the defined startup and shutdown periods, and that the testing methodologies are not designed to measure HAP emissions during startup and shutdown periods with a reasonable degree of accuracy even to the extent the HAP emissions can be measured at all (e.g., with CEMS). See 2014 Startup/Shutdown Final Rule, 79 FR 68781. The petitioners provided nothing more than their own conclusory statements that the EPA's conclusions are incorrect. They do not provide studies or information demonstrating the stability of stack conditions or point to the records for the development of CEMS and stack testing methodologies that indicate those monitoring approaches were designed for the defined startup and shutdown periods. In sum, the petitioners have not provided any new data or information demonstrating that the EPA's conclusions were in error.

The EPA also disagrees that there is any inconsistency with the statements about subcritical unit startup capacity and the ability to measure emissions when loads are reduced to the same level during periods other than startup and shutdown periods. The EPA has demonstrated that there is uncertainty as to the ability of EGUs to accurately measure HAP emissions during the defined startup period and the ability of some units to reach a certain capacity within the defined period does not somehow render the startup stack conditions sufficient to accurately measure HAP emissions from such units or render the stack conditions for other types of sources sufficient. The petitioners have neither shown otherwise nor explained why this comment could not have been made in response to the June 2013 reopening of the comment period. In addition, stack conditions are more stable during load swings than during startup because combustion chamber temperatures during load swings are more constant than combustion chamber temperatures during startup or shutdown periods. For these reasons, the petitioners have failed to demonstrate that it was impracticable to comment or that the objection is of central relevance to the rule and, therefore, the EPA is denying reconsideration.

The petitioners' claim that the EPA should establish the end of startup based on PM control engagement mischaracterizes the EPA's approach to establishing the end of startup. The EPA did not base the end of startup on the ability of PM controls to be engaged and operated because the EPA did not have the type of PM data that was available for SO2 and NOX. In addition, as the petitioners noted, baghouses passively remove PM as long as the emissions are routed through the filters no matter what the stack conditions. Thus, even if the EPA had the PM data similar to the SO2 and NOX data, the data from EGUs with baghouses would not be useful in determining the time it takes for stack conditions to be sufficient to accurately measure HAP (or PM) emissions. For these reasons, this objection is not of central relevance to the rule and the EPA is denying reconsideration on that basis as well.        

The EPA disagrees that the agency could readily estimate PM, much less HAP, emissions during startup and shutdown as the petitioners suggested. The agency does not have the significant data that the petitioners suggested the EPA would need to establish estimated standards or a mechanism for reliably estimating HAP emissions during startup and shutdown periods. The petitioners similarly do not attempt to suggest what those estimated emissions limits would be or, more importantly, how the agency would monitor compliance given the uncertainty of measurements during the defined startup period. The EPA also does not agree that allowing EGUs to estimate emissions during upsets for the purpose of demonstrating compliance with an established numeric standard is analogous to the establishment of numeric standards based on emissions estimates. EGUs are required to use HAP emissions measurements during periods other than startup and shutdown to demonstrate compliance with the numeric standards and have considerable data about operating conditions during testing. That information can be compared to the levels of HAP emissions to create a relationship that allows reasonable estimates during upset periods. The same is not true for startup and shutdown periods because the EPA does not have sufficient HAP emissions or operational conditions data, or the ability to reliably measure HAP mass emissions during the defined startup period with which to base a reasonable correlation.

The EPA also disagrees that anticipated emissions during LEE bypass are reasonably comparable to emissions during startup or shutdown periods. In order to qualify for LEE status, EGUs must demonstrate that their non-mercury (Hg) emissions are less than 50 percent of the MATS limits and that their Hg emissions are less than 10 percent of the MATS limits or 29 pounds or less, and the EPA included the bypass provision to allow EGUs to bypass if necessary and not lose their LEE status. The provision requires EGUs to burn clean fuels to the maximum extent possible and to estimate HAP emissions to demonstrate that LEE status is sustained notwithstanding a bypass event. See 40 C.F.R. 63.10000(c)(1)(i)(C)(2). Given the low emission levels necessary to qualify for LEE status, coupled with the mandatory clean fuel use during the limited bypass period, the agency believes this approach is reasonable. It is due to these conditions on LEE status and, more importantly, the fact that EGU stack conditions during these non-startup periods have normalized when the bypass is implemented such that HAP emissions can be reasonably measured with available methodologies that make the LEE bypass situation different from startup and shutdown periods. See response to Issue 5. The petitioners have thus failed to demonstrate that the LEE bypass provisions are relevant or comparable to the startup and shutdown work practice such that reconsideration is warranted.

Thus, the petitioners have failed to demonstrate that it was impracticable to comment during the comment period and they have not provided any new information that was unavailable during the comment period (with within the time specified for judicial review), and the petitioners have failed to demonstrate that the PM control engagement time and the LEE bypass provisions are of central relevance to the final rule. 

For these reasons, the EPA is denying reconsideration on this issue.       

Issue 8:  Environmental Petitioners assert that the EPA must reconsider the final rule because the final SS TSD does not represent the best performing 12 percent of sources and the EGUs identified as the best performers may include units that are not best performers. As support for the argument, the petitioners identify information in the record that they assert shows that certain units that the EPA identified as best performers for purposes of the SS TSD are not among the best performers for the numeric alternative equivalent SO2 emissions standard and that sources that do not meet that limit should not be included among the best performing 12 percent of sources.

The petitioners state that they evaluated the SO2 data and selected the units meeting the standard with the lowest average emissions rates during the first 4 hours after generation. The petitioners state that the average emissions from those 150 units was below the SO2 emission limit during the first 4 hours after generation. The petitioners state that the SO2 emissions from some of the EPA's identified best performers are above the standard. The petitioners also state that the 150 units they identified had relatively stable SO2 emissions. The petitioners assert that their analysis demonstrates that the SO2 controls at the facility are on directly after generation because the SO2 emissions are so low and stable and that the SO2 emissions can be measured during those time. The petitioners state their conclusion is supported by their analysis that shows the comparable SO2 emissions from the EGUs used in the SS TSD and the 150 EGUs that the petitioners identified. 

Response 8:  The EPA is denying reconsideration on this issue because the petitioners have not demonstrated that it was impracticable to comment on certain issues and because the objections raised are not of central relevance to the final rule.

As noted in the 2014 Startup/Shutdown Final Rule and above, the EPA revised the approach for determining the end of startup in the SS TSD between proposal and final in part in response to comments from some of the petitioners. Specifically, the June 17, 2013, draft of the SS TSD averaged the startup time of all EGUs included in the analysis to determine the end of startup; however, in response to comments from petitioners, the agency revised the SS TSD and averaged the startup time of the best performing 12 percent of EGUs in the analysis to be consistent with section 112(h)(2) in light of the unique data set available for EGUs. See Revised SS TSD. The EPA also responded to petitioners' other comments, but the petitioners did not comment on how the agency should revise the TSD. See EPA-HQ-OAR-2009-0234-20427. Thus, the EPA identified the best performers based on the critical metric for determining "best" in the SS TSD (i.e., the ability to most quickly engage air pollution control devices) and calculated the end of startup using the same analytical approaches set forth in the June version of the SS TSD. The EPA used the same analytical approach absent suggestions for alternative approaches in comments. For these reasons, the petitioners have not demonstrated that it was impracticable to comment on the EPA's approach to determining the end of startup during the public comment period and the EPA is therefore denying reconsideration on this issue.

This objection is also not of central relevance to the rule because it is based on a misinterpretation of the EPA's approach to determining the best performers in the SS TSD. The petitioners appear to assume that the actual level of SO2 (and NOX) emissions is critical to determining the end of startup, but in the SS TSD the critical factor is the amount of time (not emissions) that it takes to bring the SO2 and NOX controls on-line after the start of generation. See Final SS TSD at fn. 1 ("For EGU startup and shutdown, the EPA defines the best performing EGUs by determining the EGUs that are able to bring their pollution controls on line the most efficiently."). The EPA explained in the proposed and final SS TSD that the EPA was trying to determine the end of startup based on the time it takes to engage the SO2 and NOX controls, and the agency took this approach because those controls are generally the last controls to engage at an EGU and the agency is confident that stack conditions will be sufficient at that time to reasonably accurately measure HAP emissions from EGUs. See also 2014 Startup/Shutdown Final Rule, 79 FR at 68781-82. The actual removal efficiency of the SO2 and NOX controls is not relevant to determining the end of startup because after that time sources are required to comply with the MATS numerical limits by whatever means necessary. For example, it is possible that EGUs will continue to co-fire clean fuel for some period of time after the end of startup if there is a concern that the controls are not able to sufficiently remove HAP at the end of startup or to ensure there is a compliance margin. See EPA-HQ-OAR-2009-0234-20447, Response to Comment 54 on page 59.

The EPA also finds that the analysis conducted by the petitioners is not of central relevance to the rule. The analysis petitioners present focuses on an aspect of the ARP data that is not relevant to determining the end of startup in the SS TSD (i.e., the actual level of emissions of SO2 and NOX at the time the controls are engaged). The SS TSD only included EGUs that control SO2 and/or NOX emissions with add-on controls because sources that comply with the ARP by using low sulfur coal (or low NOX burners) and/or the purchase of credits will not show a marked decline in such emissions during the startup period. In fact, sources that use low sulfur coal will have lower and more stable SO2 emissions, but that does not make such units best performers under the metric used in the SS TSD. Instead, it makes such units irrelevant for purposes of that analysis because it is at the time controls are engaged that the agency is confident that HAP emissions can be measured with a reasonable degree of accuracy. Because the petitioners' argument relies on the measured SO2 and NOX emissions levels without regard to whether controls are in place, the analysis is not comparable to the analysis in the SS TSD. In addition, the EPA does not believe the petitioners' analysis demonstrates that HAP emissions can be reasonably accurately measured during the startup period. 

For these reasons, the EPA is denying reconsideration on these issues.


Issue 9:  Industry Petitioner claims the revised analysis the EPA used to identify the best performing EGUs for determining the end of startup must be reconsidered because the EPA proposed to base the end of startup on all EGUs with the relevant data and the petitioner argues that the revised SS TSD is impossible to replicate and appears flawed. Specifically, the petitioner asserts the spreadsheet accompanying the revised analysis is opaque and cryptic, rendering the revised analysis difficult to follow. The petitioner posits that the spreadsheet lacks formulae and that the revised analysis does not explain its selection criteria. The petitioner states that it is not possible to identify the EGUs the EPA determined are the best performers. The petitioner further argues that the Excel spreadsheets for the revised SS TSD do not include the formulas and calculations. The petitioner states that the EPA did not explain the criteria for determining when the controls would be determined to be engaged. The petitioner is also concerned because the EPA did not consider all EGUs in the analysis and that other EGUs may control SO2 and NOX by means other than add on controls. The petitioner also states that the EPA used less than 12 percent of the available data set when it determined the end of startup based on SO2 data from 44 EGUs and NOX data from 27 EGUs. The petitioner also argues that the EPA has not explained how the analysis accounts for the performance of the best performing EGUs under all conditions and that absent that information the commenter is unable to evaluate the EPA's conclusion that there is not a significant difference in the startup events at different EGUs. 

Response 9: The EPA is denying reconsideration on this issue because the petitioner has not demonstrated it was impracticable to comment on the approach to determining the end of startup and because the objections are not of central relevance.

The EPA provided the SS TSD for comment and in that document the EPA established an approach for evaluating when SO2 and NOX controls are engaged to determine the end of startup. Despite the petitioner's assertion to the contrary, the EPA used the same data (i.e., the ARP data) for determining the end of startup for the best performing EGUs as was used for determining the end of startup based on all the EGUs with the relevant data. The data provided for comment with the SS TSD included hourly generation and emission rates for each unit in the study population (i.e., units with the relevant controls). Thus, the petitioner could have commented on the data and the approach for determining the end of startup. Concerning the selection of the best performing 12 percent of EGUs, the EPA chose to base the revised analysis on the best performing 12 percent of EGUs in response to comments submitted by a number of commenters. The EPA selected the best performers based on the time it takes the EGUs with the relevant controls to engage those controls, being sure to include EGUs of all types. The revised TSD includes the unit-level hourly change in emission rates and the number of hours for each unit to achieve the threshold changes in emission rates. The data file also includes a `flag' indicating whether the unit is included in the best 12 percent of EGUs for SO2 based on the time to engage SO2 controls and/or in the best 12 percent of EGUs for NOX based on the time to engage NOX controls. If the petitioner wished to evaluate the analysis and conclusion, the simple averaging exercise used to determine the time it takes the best performers to engage the relevant controls does not require manipulation of the data as the petitioner suggested. Therefore, petitioner's assertion that it was unable to identify the best performers is not credible. Section 3 of the revised SS TSD presents information from startup events at the best performing 12 percent of coal-fired EGUs, which were determined using built-in easy-to-use functions  --  average() and percentile()  --  that are well-documented within the Excel application. Similarly, the EPA did explain in the initial SS TSD how it was determining the end of startup for individual EGUs and the same approach was used in the revised SS TSD. See EPA-HQ-OAR-2009-0234-20378 and EPA-HQ-OAR-2009-0234-20451. The petitioner has mischaracterized the record on this issue and, to the extent its objections go to these mischaracterizations, they are not of central relevance to the rule.

For these reasons, the EPA concludes that the petitioner has not demonstrated it was impracticable to comment on the approach for determining the end of startup and thus denies reconsideration.

As noted above in response to the petition from Environmental Petitioners, the actual concentrations of SO2 and NOX are not relevant to determining the end of startup in the SS TSD. Instead, the reductions in those emissions simply serve to indicate how quickly the controls come on-line. Thus, only EGUs with SO2 and NOX controls are relevant to the analysis and EGUs that control those emissions by means other than add on controls are not of central relevance to the SS TSD. See response to Issue 5. 

The petitioner also incorrectly stated that the EPA used less than 12 percent of the available data set when it determined the end of startup based on SO2 data from 44 EGUs and NOX data from 27 EGUs and that the EPA has not explained how the analysis accounts for the performance of the best performing EGUs under all conditions. Only EGUs with the relevant SO2 and NOX controls have valid data for the analysis in the SS TSD. Consistent with section 112(d)(3), the EPA identified the best performers based on 12 percent of the units for which the EPA had relevant data. The EPA believes this approach is reasonable because the agency had a complete set of startup and shutdown data from the relevant EGUs, which is unusual when establishing a work practice, and because section 112(h)(2) requires work practice standards consistent with section 112(d). Petitioner's argument that the EPA has failed to account for the performance of the best performing EGUs under all operating conditions is also incorrect. As noted, the EPA has data on all startup events at EGUs pursuant to the ARP database. Because the petitioner's objection is based on a mischaracterization of the final action and the applicable law, this objection is not of central relevance and the EPA is denying reconsideration. 

Finally, the adjustments the EPA made to its analysis in response to comments do not justify reconsideration and the petitioner's objections do not provide support for the argument that the promulgated regulation should be revised and thus also are not of central relevance to the rule. It is well established that an agency may refine its proposed approach and respond to comments received without providing an additional opportunity for public comment. See Community Nutrition Institute v. Block, 749 F.2d at 58 and International Fabricare Institute v. EPA, 972 F.2d 384, 399 (D.C. Cir. 1992) (notice and comment is not intended to result in "interminable back-and-forth[,]" nor is an agency required to provide additional opportunity to comment on its response to comments). The final TSD evaluated the same data to determine the end of startup as was used in the initial TSD. In the initial TSD, the EPA presented the data, including hourly electricity generation, heat input, mass emissions, and emission rates; explanation of the data analysis; and tables and figures illustrating the results of the analysis, and provided an opportunity for public comment. Petitioner and other stakeholders had ample opportunity to comment on the proposed methodological approach and the hourly data used in the analysis. In the final TSD, the agency used this same data set to determine how long it took the best performing units to complete start-up. The EPA focused on the best performing units, instead of evaluating the data of all the EGUs with the relevant controls. The EPA took this approach in the final TSD because we received comments on the proposal noting that the statute requires the EPA to establish section 112(h) work practices consistent with the requirements of section 112(d). As section 112(d) requires standards to be based on the best performing EGUs in a category or subcategory, the EPA concluded that it should consider only data from the best performing EGUs when determining the end of start-up. The refinement the EPA made to its approach for the final rule was a direct response to comments on the initial TSD, and the petitioner here does not directly challenge the EPA's legal conclusion or suggest the adjustment made in response to comments was in error. Information submitted during the comment period often helps the EPA refine its approach and the notice and comment requirements of the act do not prohibit the EPA from taking that information into account. Where, as here, there is comprehensive data for evaluating an aspect of the work practice, it was reasonable for the EPA to take that information into account and give effect to the express requirements of the statute. In addition, the petitioner does not provide data or information that demonstrates that it was in error for the EPA to conclude that stack conditions will be sufficient at the defined end of startup such that HAP emissions can be measured with a sufficient degree of accuracy. The petitioner also does not provide data or information supporting a conclusion that the EPA erred in its approach to selecting the best performer, nor does the petitioner provide information or analysis to support an alternative approach to identifying the best performing sources. The petitioner thus has not presented any information that would support a change in the EPA's legal or factual conclusions. As such, petitioner's objection would not support modification of the rule and thus the objection is not of central relevance to the rule and, for this reason also, the EPA is denying reconsideration.

For these reasons, the EPA is denying reconsideration on this issue.


IV. APCD Operational Issues

Issue 10: Environmental Petitioners raise the issue that EPA's determination that emissions cannot be measured until 4 hours after generation also failed to consider an important aspect of the problem. The EPA determined when conditions are stable enough for emissions to be accurately measured based on when EGUs choose to engage their pollution controls, which does not rule out that these controls could have been engaged earlier or that conditions were not earlier stable enough to measure emissions.

Response 10: The EPA is denying reconsideration on this issue because the petitioners have not demonstrated that it was impracticable to comment on this issue or provided new information that was not available during the public comment period (but within the time specified for judicial review) to support reconsideration. We direct the petitioners to the response to Issue 5 for a response to the comment concerning when air pollution controls can be safely and efficiently engaged. 

In addition, the EPA is denying reconsideration because this objection is not of central relevance to the rule. The EPA used the considerable data available to it when determining the end of startup and that approach is consistent with section 112(d)(3). The agency would not be required to collect additional data even if we granted reconsideration. Moreover, the EPA has no data to support a conclusion that the best performers could improve their control engagement time and the record indicates that there are safety and control device and unit integrity considerations that would caution against directing facilities to attempt to engage their controls earlier for purposes of data collection. See 2014 Startup/Shutdown Final Rule, 79 FR 68783 and EPA-HQ-OAR-2009-0234-20447, page 46. In fact, the EPA believes the information that will be collected for the alternative option in the final SS reconsideration rule will be useful in further refining the work practice and potentially moving to a numeric standard in lieu of a work practice during subsequent reviews pursuant to section 112(d)(6). See 2014 Startup/Shutdown Final Rule, 79 FR 68779, footnote 3 and 79 FR 68786. 

For these reasons, the EPA is denying reconsideration on this issue. 


Issue 11:  Industry Petitioner initially argued that the EPA should reconsider its decision not to add limestone injection into a fluidized bed combustion boiler and selective non-catalytic reduction to the list of excepted control devices in the work practice standard applicable under paragraph 1 of the definition of startup for the final startup and shutdown work practice. In a supplement to their petition dated March 20, 2015, the Industry Petitioner withdrew its request with respect to these two issues. 

Response 11:  As petitioner's request for reconsideration of this issue has been withdrawn, the issue is now moot.  

 
V. Work Practice Issues

Issue 12:  Environmental Petitioners maintain the EPA must grant reconsideration because the final work practice is inconsistent with CAA sections 112(d) and 302(k). According to the petitioners the final definition of the end of startup as being 4 hours after generation means there is no continuous standard in place to limit HAP emissions. The petitioners argue that the statute requires continuous emission standards even for work practice standards and that the final rule does not comply with that requirement.

The petitioners characterize the definition of the end of startup as a 4 hour exemption from the numerical standards and argue it is not a work practice. The petitioners state that EGUs using the alternative definition are only required to reduce emissions by use of clean fuels to the maximum extent possible, by the engagement of PM controls within 1 hour of generation, and by the requirement to engage other controls as expeditiously as practicable (The petitioners also mention a proposed change to the requirement that the EPA did not finalize). The petitioners also acknowledge the expanded recordkeeping and reporting requirements that attend the alternative definition of startup and that such requirements could provide useful information about EGU startup and shutdown periods as the EPA explained, but the petitioners assert such requirements do not reduce emissions. 

The petitioners also assert that the requirement to engage and operate PM controls is vague and could result in minimal reductions in HAP emissions during startup. The petitioners state that ESPs for example have a wide range of effectiveness depending on how they are operated and the regulation does not require sources to operate the controls in any particular manner. The petitioners further argue that the clean fuels and non-PM controls are not a work practice because they are vague and unenforceable according to the petitioners. The petitioners assert that they interpret the regulation to allow EGUs to cease using clean fuels even before generation and also to allow EGUs to engage controls more than 4 hours after generation. The petitioners also argue that many coal-fired EGUs cannot co-fire once they begin to generate electricity and fire coal so the work practice will not require any clean fuel use after generation.

The petitioners further state that the work practice does not comply with section 112(h) because that provision requires the work practice to be as stringent as the limitation in HAP emissions achieved by the best performing 12 percent of EGUs. Specifically, the petitioners assert that the EPA erred because the agency did not determine exactly how long the best performing sources use clean fuels during startup. The petitioners also assert that the non-PM control requirement is inconsistent with the statute because the requirement does not reflect the time the best performers actually take to engage the controls. In sum, the petitioners maintain that the EPA made an insufficient effort to comply with the requirements of section 112(d)(2) and that the EPA is shirking its responsibility to establish valid requirements because the sources will have some flexibility as to the timing of the engagement of controls and the efficiency at which the controls will operate during startup and shutdown.

The petitioners further argue that the non-PM controls "requirement" is an equipment standard that fails for vagueness. Finally, the petitioners assert that the EPA failed to explain its change from the final rule and that agencies are allowed to make changes only if the changes are adequately explained. In this case, the petitioners assert that the EPA has not explained why the work practice standard in the final MATS rule was changed and the alternative approach promulgated.

Response 12:  The EPA is denying reconsideration on these issues because the petitioners have failed to demonstrate that they did not have an opportunity to comment on these issues during the comment period. In fact, the petitioners commented on these issues and the EPA responded to the comments. For example, several of the petitioners commented that the section 112 emissions limits must be met at all time (EIP comments at 2-4), that the work practice is not consistent with the statute (EIP comments at 9), operation of PM controls during startup (EIP comments at 10-11), and non-PM controls (SS RTC, page 122, commenter 20427 Sierra Club)). The EPA responded to the comments in the preamble to the final rule when it explained the final alternative work practice and in the RTC. See 2014 Startup/Shutdown Final Rule at 79 FR 68779-68783 and SS RTC, pages 41-46. 

The work practice standard in the November 30, 2012, reconsideration proposal and the alternative work practice standard in the 2014 Startup/Shutdown Final Rule are similar in regards to several of the requirements noted by the petitioners. Specifically, both work practices require that the EGU startup on clean fuel with the alternative requiring clean fuel to the maximum extent possible. Once coal firing starts, the proposed work practice requires all applicable controls, with a few exemptions (none being PM controls), be engaged, with the remaining controls started as expeditiously as possible and as necessary to comply with the applicable requirements. The alternative requires that PM control devices be engaged within 1 hour of coal firing, with all other applicable controls started as expeditiously as possible as necessary to comply with the numeric standards. As stated in the 2014 Startup/Shutdown Final Rule, we took this approach of redefining the end of startup because we determined that flue gas conditions will be adequate to accurately measure HAP emissions with CEMS 4 hours after the generation of electricity and to conduct a stack test. The approach evaluated the time for all APCDs to be functioning because we determined that stack conditions will be stable at this point (see 79 FR 68781). This determination was based on the results of the technical analysis conducted of startup events of the best performing coal-fired EGUs.

In addition, the requirements to "engage and operate" PM controls within 1 hour, to use clean fuels to the "maximum extent possible", and to engage non-PM controls "as expeditiously as possible" are similar to the requirements in the option contained in the final MATS rule and in the 2012 MATS Startup/Shutdown Proposal. See MATS Final Rule, 77 FR at 9381 (describing the work practice in part as requiring EGUs to "operate all control devices necessary to meet the normal operating standards under this final rule (with the exception of dry scrubbers and SCRs) when coal ... is fired in the boiler during startup or shutdown"); see also 2012 MATS Startup/Shutdown Proposal, 77 FR at 71331 (proposing for comment the startup/shutdown work practice in the final MATS with some revisions). Petitioners thus had ample opportunity to comment on these specific requirements during the comment period on the 2012 MATS Startup/Shutdown Proposal. They have not demonstrated that it was impracticable to raise this issue during the comment period for that rule and the EPA is therefore denying reconsideration. 

Furthermore, the EPA does not agree that the alternative approach is vague or unenforceable and, as the petitioners noted, EGUs that choose to comply with the alternative approach are subject to more recordkeeping and reporting requirements. These additional recordkeeping and reporting requirements assure compliance with the alternative standard and the information will be very useful for the EPA as it attempts to characterize HAP emissions during these periods. The EPA further believes that the additional recordkeeping and reporting requirements are part of the reason the two work practice approaches are equivalent. It is correct that the approach in the final MATS requires compliance with HAP standards at first generation, but the EPA does not believe the current HAP monitoring methodologies are capable of accurately measuring HAP mass emissions at those times and EGUs are not required to monitor emissions at those time unless they choose to comply with CEMS or sorbent trap monitoring systems. Thus, the EPA will likely obtain little useful information from sources that comply with the work practice in the final MATS. Conversely, the alternative approach established in the final reconsideration requires sources to keep records of date and time that clean fuel use starts and stops, quantity and heat input of clean fuel combusted, gross output, and PM control device parameter values and upstream and downstream flow, as well as temperatures and fan amps from EGUs without instrumental PM measurements or LEE status, and that information can be used to ensure EGUs are minimizing HAP emissions as required in the final rule and potentially be used to develop better approaches to monitoring and minimizing emissions during those periods. See e.g., 63.10020(e) (requiring EGUs to collect certain data during startup and shutdown periods).

The petitioners submitted comments on the proposed reconsideration arguing that the work practice standard must reflect the performance of the best performing 12 percent of sources in the category. The agency revised the SS TSD between proposal and final in response to the petitioners' comment on this issue. See 2014 Startup/Shutdown Final Rule, 79 FR at 68779. Petitioners thus had ample opportunity to comment on the issue and the EPA responded to those comments. The EPA is thus denying reconsideration on this issue as petitioners have not demonstrated it was impracticable to submit comments during the public comment period. 

In addition, the EPA notes that it agrees that the statute requires work practice standards "which in the Administrator's judgment is consistent with the provisions of subsection (d) or (f) of this section."  See CAA section 112(h)(1) (emphasis added). The EPA does not agree, however, that each individual element of a work practice standard must be based on the best 12 percent of sources as the petitioners imply. The EPA does not believe that such a literal reading of section 112(h)(1) is reasonable because it is the lack of available data and ability to obtain the necessary data that generally support the inclusion of a work practice standard under section 112 and there is generally no reasonable way to determine the best performing 12 percent. For example, the EPA based the clean fuel requirement on information indicating that EGUs generally use clean fuels to startup and that clean fuels were the only fuel used until and even after generation at some EGUs. See EPA-HQ-OAR-2009-0234-20447, Response 104 on page 113. The record in this case was unusual when it came to data available to determine the end of startup for the work practice because of the significant information available from the ARP. For this reason the agency revised the SS TSD to base the end of startup on the EGUs that were able to start their SO2 and NOX controls most quickly after generation instead of basing the time on the average time it took all EGUs to engage those controls. See 2014 Startup/Shutdown Final Rule, 79 FR 68781-68782.    

The EPA further notes that the agency agrees that the CAA generally requires emission standards to be continuous; however, we disagree with the petitioners' characterization of the final work practice as something other than a continuous requirement. Specifically, the petitioners focus only on the 4 hours after generation as being problematic, but, as explained in the preamble to the final rule, the work practice applies to the entire startup period and that over that entire period the agency believes that HAP emissions will be as low if not lower than emissions of HAP during normal operations. See 2014 Startup/Shutdown Final Rule, 79 FR at 68780. The petitioners have provided nothing more that bare assertions to the contrary as support for their argument that the work practice is not continuous and does not reduce HAP emissions, and the EPA does not agree that it is reasonable to segment the work practice as suggested by the petitioners.

Finally, the EPA does not agree with the petitioners' equipment standard characterization. Petitioners had ample opportunity to comment on the end of startup and when EGUs must engage controls. The EPA presented information on both issues in the SS TSD and petitioners had an opportunity to comment on that TSD and all the analyses contained therein. Thus, petitioners have not demonstrated that it was impracticable to comment on this issue during the comment period we are therefore denying reconsideration on this issue. In addition, the petitioners suggest the final rule is flawed because of uncertainty concerning when the work practice requires EGUs to operate non-PM controls. It appears that the petitioners misunderstand the rule because the alternative work practice does not in fact require EGUs to operate the non-PM controls at all if the sources are able to meet the HAP standards without such controls at the end of startup. Instead, the work practice simply requires EGUs to engage the controls as soon as possible as necessary to meet the HAP emission limits in MATS or other applicable requirements. Thus, the EPA does not agree that the work practice is an equipment standard as it relates to the non-PM controls and, for this reason, reconsideration is also denied because this issue is not of central relevance to the final work practice.

For all of these reasons, the EPA is denying reconsideration on this issue. 


Issue 13:  Environmental Petitioners assert that the EPA must reconsider all the final MATS numerical standards if the final alternative approach defining the end of startup as 4 hours after generation is lawful. The petitioners assert that the 30-day averaging period used to determine compliance with the MATS numerical HAP limits included variability to account for periods of startup and shutdown and that because the EPA has defined the end of startup in the final SS reconsideration as 4 hours after generation the EPA must explain how the numerical standards continue to satisfy the requirements of section 112(d)(3). The petitioners assert that such an explanation is particularly important because of what the petitioners characterize as the massive amounts of HAP emissions that EGUs emit during startup. The petitioners state that the EPA's assertion that HAP emissions under the final work practice standard will be as low if not lower during startup pursuant to the final work practice ignore the evidence to the contrary presented by Environmental Petitioners. The petitioners assert that the EPA has constructively reopened the MATS numerical limits, and the petitioners argue that those limits do not comply with section 112(d)(2).

Response 13:  The EPA is denying reconsideration because the petitioners have not demonstrated that it was impracticable to comment on this issue during the comment period. As noted above, the EPA finalized a work practice standard in lieu of a numerical emission standard for startup and shutdown periods. The petitioners could have and should have at that time raised their concern with the impact of such a work practice on the numerical limits in a petition for reconsideration. Similarly, the petitioners could have and should have raised this issue in response to the 2012 MATS Startup/Shutdown Proposal and the June 2013 reopening of the comment period. In fact, the EPA received comments on this issue from petitioners in response to the June 2013 reopening, and the EPA responded to those comments. See SS RTC at 74-79.

Thus, the petitioners' assertions that it was not practicable to comment on this issue during the comment period are without merit because the June 2013 SS TSD and reopening of the comment period clearly indicated that the EPA was considering redefining the end of startup. In addition, the petitioners' information that HAP emissions during startup are significant are not supported and fail to consider the impact of the final work practice on such emissions. In addition, the petitioners continue to characterize the startup period as only the 4 hours after generation, but the EPA instead considers the entire startup period (including the times when only clean fuels are being combusted) to support the conclusion that the HAP emissions during the startup period should be as low if not lower than during periods other than startup for sources complying with the final work practice. The petitioners provide no information or analysis to support a contrary conclusion.

For these reasons, the EPA is denying reconsideration on this issue.     


VI. Use of Diluent Cap Issue  

Issue 14: Industry Petitioner claims that the EPA unreasonably failed to allow the use of the diluent cap in a manner consistent with the agency's analysis. The petitioner states that the 2012 MATS allowed EGUs complying with a heat input based limit to cap diluent values for calculating mercury emissions during SS periods and the EPA solicited comment on the need for the diluent cap for other HAP emissions. The petitioner notes that it commented on the need for such caps during any hour where the diluent values are outside normal values. 

The petitioner notes that the 2014 SS rule allowed the use of diluent caps, but only during the defined startup and shutdown periods. The petitioner states that the EPA attempts to explain its decision by pointing to CO2 CEMS data showing that default CO2 values were exceeded within 2 hours of generation. The petitioner argues that the EPA did not provide the CO2 analysis for comment and the EPA must provide it for comment. In addition, the petitioner states the EPA's analysis shows that diluent values may be below default values for up to 2 hours after generation. The petitioner states that based on this information EGU owners or operators may have to use non-normal diluent values to calculate emissions for up to 2 hours because generation of electricity triggers the end of the startup period for EGUs complying with paragraph 1 of the definition of startup. The petitioner states that this result is contrary to the EPA's stated intent. The petitioner asserts the issue is of concern to units with both single and shared stacks. The petitioner also asserts that the EPA stated that the agency does not "expect" diluent cap values to be used when the startup period ends suggests such values might be used, but the final rule prohibits use of the default diluent cap values after startup ends. The petitioner claims that the EPA does not explain what analysis of diluent values it performed and that no such analysis was made available for public comment. 

Response 14: The EPA is denying a reconsideration on this issue because the petitioner has failed to demonstrate that it was impracticable to comment on this issue during the comment period. In fact, Industry Petitioner acknowledges that it did submit comments on the need for diluent caps during periods when diluent values are outside normal values. The EPA received comments on whether the use of diluent caps should be allowed during periods of startup or shutdown during the comment period for the 2011 proposed MATS, in petitions for reconsideration of the 2012 final MATS, and during the comment period for the November 2012/June 2013 startup and shutdown reconsideration notices. The EPA responded to the comments on this issue in the Federal Register notices for the 2011 MATS Proposal, the 2012 MATS Final Rule, and the 2014 Startup/Shutdown Final Rule and accompanying response to comment documents. See, e.g., 2011 MATS Proposal, 76 FR 25028; Volume 2, Response to Comments on the Proposed Rule, pp 28, 127, 423, 424, and 755; 2012 MATS Final Rule, 77 FR 9381; 2014 Startup/Shutdown Final Rule, 79 FR 68785; SS RTC pp 69, 85, 91, 92, 148, 161, 164-167, and 172. The EPA provided multiple opportunities for comment on this issue and petitioners used these opportunities to raise their concerns regarding the use of diluent values. In addition, the petitioner's assertion that they could not comment on the potential differences under paragraphs 1 and 2 of the final startup definition are without merit. The EPA provided an opportunity to comment on the SS TSD and that analysis showed that the EPA was considering providing an alternative definition of the end of startup; thus, concerns about potential inconsistencies in availability of a default diluent cap could have been raised at that time. Because the petitioner has not demonstrated that it was impracticable to comment on this issue during the comment period, the EPA is denying reconsideration. 

With regard to the analysis of the diluent values during startup periods, as explained in the preamble, a review of reported CO2 values for the EGUs contained in the Startup Technical Analysis (the analysis) was performed on the same categories as those used for SO2 and NOX emissions. For all categories, average CO2 values exceeded the diluent cap level (5 percent) within 2 hours of electrical generation. See 2014 Startup/Shutdown Final Rule, 79 FR 68785. The information contained in the rule, was not, as the petitioner suggests, an in depth analysis of CO2 values. Instead, it was simply a review of reported CO2 values for these EGUs. The values came from publically available Acid Rain data that clearly provides the necessary information (i.e., the hour after generation and the corresponding reported CO2 values). See TSD, footnote 1. The EPA did not conduct an analysis as the petitioner suggests and no such analysis is required to support the conclusion in light of the available data. Petitioners' objections thus have no bearing on the analysis done to support the final rule and thus are also not of central relevance to the rule. The EPA is denying reconsideration for this reason as well. 

In addition, the petition could be interpreted as asking the EPA to reconsider the availability of a default diluent cap during periods other than startup and shutdown periods. The diluent cap issue has been raised in comments on the proposed rule and on the startup and shutdown reconsideration proposal, and that issue has been addressed in the response to comment document for the proposed rule, in the final rule, in the response to comment document for the startup and shutdown reconsideration, and in the final startup and shutdown reconsideration rule. See, e.g., 76 FR 25028; 76 FR 25106; 77 FR 9380; 77 FR 9381; 77 FR 9606; 76 FR 25028; 78 FR 38002, footnote 1; 78 FR 38004; 78 FR 38005; RTC, Volume 2, comment and response numbers 94, 48, 49, 85, and 106 on pages 217, 423, 424, 754, and 760; SS RTC, comment and response 164; 79 FR 68785; and 79 FR 68789. In fact, one of the commenters (UARG) has already provided a series of comments on the diluent cap issue. See comment 164 in the SS RTC. The EPA need not address the petitioners' arguments that this issue is of central relevance because the petitioners have not demonstrated that it was impracticable to comment during the comment period, when relevant, as discussed above.  

In this particular circumstance, the availability of a default diluent cap during periods other than startup and shutdown periods as an issue is beyond the scope of the 2014 Final Startup/Shutdown Reconsideration Rule. Because this issue was not addressed in the final 2014 rule, there is no relevant final action with regard to that issue that could be reconsidered. Thus, to the extent that the petition is asking the EPA to reconsider the availability of a default diluent cap during periods other than startup and shutdown periods, we are denying reconsideration for that reason as well. 

Moreover, nothing in the final rule requires EGU owners or operators to comply with paragraph 1 of the definition of startup. EGUs may opt to comply with that paragraph (meaning that CO2 values will be used in calculations to determine compliance with emission limits when electrical generation starts) or with the definition of startup contained in paragraph 2 (meaning that CO2 values will be used in calculations to determine compliance with emission limits beginning 4 hours after electrical generation starts). Any EGU owner or operator concerned about the potential impact on emissions of using actual diluent values once the cap has been exceeded should choose to comply with paragraph 2 of the definition of startup. Furthermore, as the commenter noted, the standards are based on the 30 boiler operating day rolling average and the inability to use default values for about 2 hours cannot reasonably be expected to impact an EGU compliance with the final MATS standard for EGUs that choose to comply with paragraph 1 of the definition of startup. In addition, EGU owners and operators have considerable operational flexibility and could, for example, choose to combust a greater amount of clean fuels before electrical generation to raise the diluent values or continue to combust as much clean fuel as possible even after startup to limit HAP emissions until the diluent values exceed default levels. Given the significant flexibility given to EGU owners or operators with regard to the need to include CO2 values in compliance calculations and the number of potential compliance options, the commenter's concerns about the use of diluent value caps in one of the available options would not support modification of the rule. For all these reasons, the EPA finds petitioner's objection is not of central relevance to the rule and reconsideration is also denied for that reason.  

FOR THESE REASONS, THE EPA IS DENYING RECONSIDERATION On this issue. 
