[Federal Register Volume 87, Number 168 (Wednesday, August 31, 2022)]
[Rules and Regulations]
[Pages 53381-53404]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2022-18669]


=======================================================================
-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 52

[EPA-R03-OAR-2022-0347; FRL-9333-02-R3]


Federal Implementation Plan Addressing Reasonably Available 
Control Technology Requirements for Certain Sources in Pennsylvania

AGENCY: Environmental Protection Agency (EPA).

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: The Environmental Protection Agency (EPA) is promulgating a 
Federal implementation plan (FIP) for the Commonwealth of Pennsylvania 
(Pennsylvania or the Commonwealth). This FIP sets emission limits for 
nitrogen oxides (NOX) emitted from coal-fired electric 
generating units (EGUs) equipped with selective catalytic reduction 
(SCR) in Pennsylvania in order to meet the reasonably available control 
technology (RACT) requirements for the 1997 and 2008 ozone national 
ambient air quality standards (NAAQS). This action is being taken in 
accordance with the requirements of the Clean Air Act (CAA).

DATES: This final rule is effective on September 30, 2022.

ADDRESSES: EPA has established a docket for this action under Docket ID 
Number EPA-R03-OAR-2022-0347. All documents in the docket are listed on 
the www.regulations.gov website. Although listed in the index, some 
information is not publicly available, e.g., confidential business 
information (CBI) or other information whose disclosure is restricted 
by statute. Certain other material, such as copyrighted material, is 
not placed on the internet and will be publicly available only in hard 
copy form. Publicly available docket materials are available through 
www.regulations.gov, or please contact the person identified in the FOR 
FURTHER INFORMATION CONTACT section for additional availability 
information.

FOR FURTHER INFORMATION CONTACT: David Talley, Permits Branch (3AD10), 
Air & Radiation Division, U.S. Environmental Protection Agency, Region 
III, Four Penn Center, 1600 John F. Kennedy Boulevard, Philadelphia, 
Pennsylvania 19103. The telephone number is (215) 814-2117. Mr. Talley 
can also be reached via electronic mail at [email protected].

SUPPLEMENTARY INFORMATION:

I. Background

    On May 25, 2022 (87 FR 31798), EPA published a notice of proposed 
rulemaking (NPRM) addressing NOX emissions from coal-fired 
power plants in the Commonwealth of Pennsylvania. In the NPRM, EPA 
proposed a FIP in order to address the CAA's RACT requirements under 
the 1997 and 2008 ozone NAAQS for large, coal-fired EGUs equipped with 
SCR in Pennsylvania. As discussed in the NPRM, the FIP was proposed as 
an outgrowth of a decision by the United States Court of Appeals for 
the Third Circuit (``the Court''), which vacated and remanded to EPA a 
portion of our prior approval of Pennsylvania's ``RACT II'' rule which 
applied to the same universe of sources. See 87 FR 31798; 31799-39802.
    The Court directed that ``[o]n remand, the agency must either 
approve a revised, compliant SIP within two years or formulate a new 
Federal implementation plan.'' Sierra Club v. EPA, 972 F.3d 290, 309 
(3rd Circuit 2020) (``Sierra Club''). On September 15, 2021, EPA 
proposed disapproval of those portions of the prior approval which were 
vacated by the Court. See 86 FR 51315. EPA took final action to 
disapprove the vacated portions of our prior approval. 87 FR 50257, 
August 16, 2022. EPA is now finalizing a FIP to fulfill the Court's 
order.
    The collection of sources addressed by the RACT analysis in this 
FIP has been determined by the scope of the Third Circuit's order in 
the Sierra Club case and EPA's subsequent disapproval action. Herein, 
EPA is finalizing RACT control requirements for the four facilities 
that remain open and active that were subject to the SIP provision that 
the Court vacated EPA's approval of and that EPA thereafter 
disapproved: Conemaugh, Homer City, Keystone, and Montour. EPA's prior 
approval action and the Court's decision related to source-specific 
RACT determinations for the Cheswick, Conemaugh, Homer City, Keystone, 
and Montour generating stations. The Bruce Mansfield and Cheswick 
facilities ceased operation, so there is no longer a need to address 
RACT requirements for those facilities, so are not at included in this 
final action. Accordingly, there are a total of nine affected EGUs/
units at four facilities in this action: three at Homer City and two 
each at Conemaugh, Keystone and Montour.
    The Pennsylvania Department of Environmental Protection (PADEP) 
undertook efforts to develop a SIP revision addressing the deficiencies 
identified by the Third Circuit in the Sierra Club decision. PADEP 
proceeded to develop source specific (``case-by-case'') RACT 
determinations for the generating stations at issue. By April 1, 2021, 
each of the facilities had submitted permit applications to PADEP with 
alternative RACT proposals in accordance with 25 Pa. Code 129.99. 
Subsequently, PADEP issued technical deficiency notices to obtain more 
information needed to support the facilities' proposed RACT 
determinations. Although additional information was provided in 
response to these notices, PADEP determined the proposals to be 
insufficient and began developing its own RACT determination for each 
facility. The outcome of this process was PADEP's issuance of draft 
permits for each facility, which were

[[Page 53382]]

developed with the intention of submitting each case-by-case RACT 
permit to be incorporated as a federally enforceable revision to the 
Pennsylvania SIP. Each draft permit underwent a 30-day public comment 
period,\1\ during which EPA provided source-specific comments to PADEP 
for each permit. On May 26, 2022, PADEP submitted case-by-case RACT 
determinations for Keystone, Conemaugh, and Homer City as a revision to 
the Pennsylvania SIP. On June 9, 2022, PADEP submitted a case-by-case 
RACT determination for Montour as a revision to the Pennsylvania SIP. 
EPA has not yet fully evaluated those submittals and they are outside 
of the scope of this action. Any action on those proposed SIP revisions 
will be at a later date and under a separate action.
---------------------------------------------------------------------------

    \1\ See 51 Pa.B. 5834, September 11, 2021 (Keystone); 51 Pa.B. 
6259, October 2, 2021 (Conemaugh); 51 Pa.B. 6558, October 16, 2021 
(Homer City); 51 Pa.B. 6930, November 6, 2021 (Montour); Allegheny 
County Health Department Public Notices, December 2, 2021 
(Cheswick).
---------------------------------------------------------------------------

II. Summary of FIP and EPA Analysis

A. Overall Basis for Final Rule

    This section presents a summary of the basis for the final FIP. The 
overall basis for the proposal was explained in detail in the NPRM. The 
overall basis is largely unchanged from proposal, though as explained 
in the responses to comments and section IV of this document on the 
final limits, some adjustments were made to the resulting limits. For 
more detail on what was proposed, please refer to the May 25, 2022 
proposal publication (87 FR 31798).
    The basis for the final rule begins with the RACT definition. As 
discussed in the NPRM, RACT is not defined in the CAA. However, EPA's 
longstanding definition of RACT is ``the lowest emission limit that a 
particular source is capable of meeting by the application of control 
technology that is reasonably available considering technological and 
economic feasibility.'' \2\ The Third Circuit decision ``assume[d] 
without deciding'' that EPA's definition of RACT is correct. Sierra 
Club at 294. EPA is using its longstanding definition of RACT to 
establish the limits in this FIP.
---------------------------------------------------------------------------

    \2\ See Memo, dated December 9, 1976, from Roger Strelow, 
Assistant Administrator for Air and Waste Management, to Regional 
Administrators, ``Guidance for Determining Acceptability of SIP 
Regulations in Non-Attainment Areas,'' p. 2, available at https://www3.epa.gov/ttn/naaqs/aqmguide/collection/cp2/19761209_strelow_ract.pdf (Strelow Memo), and 44 FR 53761, at 53762, 
footnote 2 (September 17, 1979).
---------------------------------------------------------------------------

    The EPA proposed that RACT limits in this FIP will apply throughout 
the year. As discussed further in Section III of this preamble in 
response to comments on this issue, the EPA is retaining year-round 
limits because the limits herein are technologically and economically 
feasible during the entire year. While other regulatory controls for 
ozone, such as the Cross State Air Pollution Rule (CSAPR) and its 
updates, may apply during a defined ozone season, the RACT limits 
finalized herein do not authorize seasonal exemptions based on 
atmospheric conditions or other factors. As explained, this action is 
being finalized to meet the statutory requirement to implement RACT in 
accordance with sections 182 and 184 of the Clean Air Act. 
Implementation of RACT, and the definition of what is RACT, is not 
constrained by the ozone season or atmospheric consideration. 
Therefore, the limits finalized here apply throughout the year since 
the RACT emissions rates are technologically and economically feasible 
year-round. To the degree that the EPA analyses underlying the RACT 
emissions limits here rely on past performance data, those calculations 
typically use ozone season data. This is because ozone season data 
generally represent the time period over which the NOX 
emissions rate performance of these units is the best. Put another way, 
the ozone season data for the facilities subject to these limits are a 
reliable indicator of what is technologically and economically feasible 
for these facilities, and EPA has no reason to believe that achieving 
the same performance outside the ozone season would be technologically 
or economically infeasible. As explained further in the next section, 
no commenters presented compelling evidence to change EPA's conclusion 
on this point.
    The EPA proposed to develop the FIP limits using a weighted rate 
approach, and is retaining that overall approach here. EPA received 
significant comments both for and against such an approach, which are 
discussed in detail in the next section. Overall, upon consideration of 
these comments, the EPA's judgment is that this approach is still the 
best approach for addressing the Court decision and addressing SCR 
operation during EGU cycling (the operation of EGUs turning on and off 
or operating at varying loads levels based on electric demand). As we 
discussed extensively at proposal, the cycling of units, combined with 
the role of flue gas temperature in SCR performance, prompted EPA to 
consider how best to establish RACT limits that address the Third 
Circuit's concerns about allowing less stringent limits when flue gas 
temperatures went below what it considered to be an arbitrary 
temperature threshold. This is a challenging factor to consider in 
cases when the operating temperature varies, and when the units spend 
some time at temperatures where SCR is very effective, and some time at 
temperatures where it is not.
    At proposal, EPA provided an assessment of whether the units in 
this FIP exhibit a pattern of cycling between temperatures where SCR is 
effective and where it is not. EPA evaluated years of data submitted by 
these sources to EPA to characterize their variability in hours of 
operation or level of operation.\3\ In particular, EPA used this 
information to identify whether, or to what degree, the EGUs have 
shifted from being ``baseload'' units (i.e., a steady-state heat input 
rate generally within SCR optimal temperature range) to ``cycling'' 
units (i.e., variable heat input rates, possibly including periods 
below the SCR optimal temperature range). All of these EGUs were 
designed and built as baseload units, meaning the boilers were designed 
to be operated at levels of heat input near their design capacity 24 
hours per day, seven days per week, for much of the year. As a result, 
the SCRs installed in the early 2000s were designed and built to work 
in tandem with a baseload boiler.\4\ In particular, the SCR catalyst 
and the reagent injection controls were designed for the consistently 
higher flue gas temperatures created by baseload boiler operation. In 
more recent years, for multiple reasons, these old, coal-fired baseload 
units have struggled to remain competitive when bidding into the PJM 
Interconnection (PJM) electricity market.\5\ Nationally, total electric 
generation has generally remained consistent, but between 2010 and 
2020, generation at coal-fired utilities has declined by 68%.\6\ As a 
result, many of these units more recently have tended to cycle between 
high heat inputs, when electricity demand is high, and lower heat 
inputs or complete shutdowns,

[[Page 53383]]

when demand is low, sometimes on a daily basis. This cycling behavior 
can affect the ability of the EGUs to operate their SCRs because at 
lower heat inputs the temperature of the flue gas can drop below the 
operating temperature for which the SCR was designed.\7\ Nothing in the 
comments undermined EPA's basic conclusion that this cycling pattern is 
occurring. Accordingly, the final rule establishes limits that account 
for the technical limits on SCR operation that can result from this 
cycling behavior.
---------------------------------------------------------------------------

    \3\ See the Excel spreadsheet entitled ``PA-MD-DE SCR unit data 
2002-2020.xlsx'' in the docket for this action.
    \4\ This point is not applicable to the Conemaugh facility where 
SCR was installed much later than other facilities at issue in this 
rule. According to Key-Con's comment letter, ``KEY-CON Management 
understood that compliance with the near-future MATS Rule and PADEP 
RACT II Rule would preclude unit operations that bypassed the SCRs 
at both stations.'' See Key-Con comments at 10.
    \5\ PJM is a regional transmission organization (RTO) or grid 
operator which provides wholesale electricity throughout 13 states 
and the District of Columbia.
    \6\ U.S. Energy Information Administration, ``Electric Power 
Annual 2020,'' Table 3.1.A. Net Generation by Energy Source, https://www.eia.gov/electricity/annual/.
    \7\ U.S. EPA, ``EPA Alternative Control Techniques Document for 
NOX Emissions from Utility Boilers'' EPA-453/R-94-023, 
March 1994, p. 5-119, https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=2000INPN.txt.
---------------------------------------------------------------------------

    In the proposal, we also noted that in RACT II, PADEP attempted to 
address this cycling behavior by creating tiered emissions limits for 
different modes of operation based on the flue gas temperature, which 
its RACT II rule expressed as a transition from the 0.12 pounds of 
NOX per million British thermal units (lb/MMBtu) rate to 
much less stringent rates (between 0.35 and 0.4 lb/MMBtu, depending on 
the type of boiler) based on a temperature cutoff of 600 degrees, with 
the less stringent rate essentially representing a ``SCR-off'' mode 
(i.e., an emission limit applicable at times when the SCR has been 
idled or bypassed and is not actively removing NOX). The 
Third Circuit rejected this approach because the selection of the 
cutoff temperature was not sufficiently supported by the record. The 
Third Circuit decision also questioned the need for the less stringent 
rates, noting that nearby states do not have different emission rates 
based on inlet temperatures. EPA considered the Court's concerns as 
well as input received during the public comment period expressing both 
support for, and opposition to, a tiered limit. We also considered the 
practical and policy implications in structuring a tiered limit for 
these cycling EGUs based on operating temperature. EPA has decided to 
retain the proposed weighted approach instead of trying to develop a 
tiered limit. As noted at proposal, the effectiveness of SCR does not 
drop to zero at a single temperature point and defining the minimum 
reasonable temperature range to begin reducing SCR operation for the 
purposes of creating an enforceable RACT limit is a highly technical, 
unit-specific determination that depends on several varying factors.\8\ 
We noted the complexity and detailed information necessary to produce a 
justified and enforceable tiered limit that represents RACT and 
addresses the Court's concerns about the basis and enforceability of 
the tiers, and as explained further in the next section, none of the 
comments, including those supporting the tiered limit, provided 
sufficient basis for EPA to change its approach.
---------------------------------------------------------------------------

    \8\ See Chapter 2, subsection 2.2.2 of the SCR Cost Manual, 7th 
Edition, available at https://www.epa.gov/sites/default/files/2017-12/documents/scrcostmanualchapter7thedition_2016revisions2017.pdf.
---------------------------------------------------------------------------

    In the proposal, EPA expressed an additional concern about 
addressing cycling operation through a tiered RACT limit based on 
operating temperature, which is that it would create an incentive for a 
source to cycle to temperatures where SCR is not required, in order to 
avoid SCR operating costs and potentially gain a competitive advantage. 
In the case of the Pennsylvania limits addressed by the Third Circuit's 
decision, there was no limit on how much time the units could spend in 
SCR-off mode. In section C of the TSD for the proposed action,\9\ EPA 
shows that over the last decade, some affected sources have varied the 
gross load level to which they cycle down, hovering either just above 
or just below the threshold at which the SCR can likely operate 
effectively. Depending on the unit, this slight change in electricity 
output could significantly affect SCR operation and the resulting 
emissions output. Though instances of cycling below SCR thresholds 
occurred in some cases prior to the implementation of Pennsylvania's 
tiered RACT limit and thus the limit may not be the sole driver of the 
behavior following its implementation, the tiered limit certainly 
allows this behavior to occur. While EPA acknowledges the need for EGUs 
to operate at times in modes where SCR cannot operate, EPA believes its 
RACT limit should minimize incentives to do that, and a tiered rate 
structure that effectively has no limit on SCR-off operation tends to 
do the opposite. We received significant comments on this concern, 
which are addressed in the response to comments section. EPA remains 
concerned about essentially unlimited SCR-off operation, and continues 
to believe that this is a key reason to retain the weighted rate 
approach over a tiered approach.
---------------------------------------------------------------------------

    \9\ EPA is not revising the TSD. Any new technical analysis will 
be discussed directly in section III (EPA's Response to Comments) of 
this preamble.
---------------------------------------------------------------------------

    On the other hand, EPA also expressed concerns in the proposal 
about a RACT limit that treats these EGUs as always operating as 
baseload units by imposing a NOX emission rate that applies 
at all times but can technically be achieved only if the boiler is 
operating at high loads. Recent data indicate that these units are not 
operating as baseload units and are not likely to do so in the 
future.\10\ Selecting the best baseload rate (the rate reflecting SCR 
operation in the optimal temperature range) and applying that rate at 
all times does not account for, and could essentially prohibit, some 
cycling operation of these units. Cycling has become more common at 
coal-fired EGUs because they are increasingly outcompeted for baseload 
power. In the past, these units were among the cheapest sources of 
electricity and would often run close to maximum capacity. Other EGUs 
can now generate electricity at lower costs than the coal-fired 
units.\11\ Thus, the coal-fired units now cycle to lower loads during 
hours with relatively low system demand (often overnight and especially 
during the spring and fall ``shoulder'' seasons when space heating and 
cooling demand is minimized) when their power is more expensive than 
the marginal supply to meet lower load levels. Hence, they cycle up and 
down as load- and demand-driven power prices rise and fall, and they 
operate when the price meets or exceeds their cost to supply power. EPA 
acknowledges that cycling down to a SCR-off mode may sometimes happen, 
for example, when electricity demand drops unexpectedly, and other 
units provide the power at a lower cost. The consideration of the 
technical and economic feasibility of a given RACT limit should 
reflect, to the extent possible, consideration of the past, current, 
and future expected operating environment of a given unit. In electing 
to finalize its weighted rate approach, EPA considered these 
feasibility issues to establish a rate for each unit that reflects a 
reasonable level of load-following (cycling) (e.g., a level consistent 
with similar SCR-equipped units) but that also accounts for the lower 
historic NOX rates that these units have achieved. While the 
comments generally affirmed that a weighted rate could be structured to 
address cycling, we did receive comments on the appropriate 
considerations in choosing

[[Page 53384]]

the final rates, which are responded to later in this notice.
---------------------------------------------------------------------------

    \10\ See section C of the TSD for the proposed action.
    \11\ The decreasing competitiveness of Pennsylvania's coal units 
is illustrated by the fact that their share of the state's total 
generation has declined from about 60% in 2001 to roughly 10% in 
2021. See Energy Information Administration. Form EIA-923, Power 
Plant Operations Report (2001-2021).
---------------------------------------------------------------------------

B. Weighted Rates

    As discussed in the NPRM, in order to address the concerns 
discussed previously in this section about how to determine RACT for 
EGUs that cycle, EPA proposed to express the RACT NOX limits 
for these units using a weighted rate limit. The weighted rate 
incorporates both a lower ``SCR-on'' limit and a higher ``SCR-off'' 
limit. Through assignment of weights to these two limits based on the 
proportion of operation in SCR-on and SCR-off modes during a historical 
period that encompasses the range of recent operation, the SCR-on and 
SCR-off limits are combined into a single RACT limit that applies at 
all times. The weight given to the proposed SCR-off limit (established 
as described later in this section) has the effect of limiting the 
portion of time a cycling source can operate in SCR-off mode and 
incentivizes a source to shift to SCR-on mode to preserve headroom 
under the limit. While driving SCR operation, the weighted limit 
accommodates the need for an EGU to occasionally cycle down to loads 
below which the SCR can operate effectively and does not prohibit SCR-
off operation or dictate specific times when it must not occur. In this 
way, this approach avoids the difficulty of precisely establishing the 
minimum temperature point at which the SCR-off mode is triggered, 
effectively acknowledging the more gradual nature of the transition 
between modes where SCR is or is not effective. Finally, it is readily 
enforceable through existing Continuous Emission Monitoring Systems 
(CEMS), without the need for development of recordkeeping for 
additional parameters that define the SCR-off mode. The approach is 
described in more detail below.
    As a starting point for developing the proposed weighted rates for 
each unit, EPA examined data related to the threshold at which these 
facilities can effectively operate their SCR. Then, EPA calculated both 
SCR-on and SCR-off rates using historic ozone season operating data for 
the unit to determine when the SCR was likely running and when it was 
likely not running, and then established rates based again on historic 
operating data that represent the lowest emission limit that the source 
is capable of meeting when the SCR is running and when it is not. EPA 
did this by using the estimated minimum SCR operation threshold as 
described in the proposed action, and then calculating average SCR-on 
and SCR-off rates for each unit based on historic ozone season 
operating data for that unit, when available, from 2003 to 2021. For 
more detail on the development of the proposed rates, see section D of 
the TSD for the proposed action. In particular, section D.1 addresses 
the proposed threshold analysis. The SCR-on rate is an average of all 
hours in which the SCR was likely running (operating above the 
threshold at which it can run the SCR with an hourly NOX 
emission rate below 0.2 lb/MMBtu) during each unit's third-best ozone 
season from the period 2003 to 2021. The third-best ozone season was 
identified based on the unit's overall average NOX emission 
rate during each ozone season from 2003 to 2021. This time period 
captures all years of SCR operation for each facility, though Conemaugh 
only installed SCR in late 2014. EPA included all these years of data 
in developing the proposed as well as the final limits because the 
Agency did not identify, and commenters did not provide, a compelling 
reason to exclude any of the years. This is in line with the Third 
Circuit's decision, which questioned EPA's review of only certain years 
of emissions data for these sources in determining whether to approve 
Pennsylvania's RACT II NOX emission rate for these EGUs. The 
use of the third-best year accounts for degradation of control 
equipment over time, and it avoids biasing the limit with 
uncharacteristically low emitting days, or under uncharacteristically 
optimal operating conditions. EPA similarly used a third-best ozone 
season approach for the Revised CSAPR Update (86 FR 23054, April 30, 
2021) (RCU) and the proposed Good Neighbor Plan for the 2015 Ozone 
NAAQS (87 FR 20036, April 6, 2022) (Good Neighbor Plan). The ``SCR-
off'' rate used to develop the proposal is an average of all hours in 
which the unit's SCR was likely not running (operating below the 
threshold at which it can run the SCR with an hourly NOX 
rate above 0.2 lb/MMBtu) during all ozone seasons from 2003-2021 
(except for Conemaugh). All ozone seasons in the time period were used 
in order to increase the sample size of this subset of the data, as an 
individual ozone season likely contains significantly fewer data points 
of non-SCR operation.
    EPA then calculated the SCR-on and SCR-off ``weights,'' which 
represent the amount of heat input spent above (SCR-on) or below (SCR-
off) the SCR threshold, for each EGU. For the weights used at the 
proposal stage, EPA evaluated data from the 2011 to 2021 ozone seasons 
and selected the year in which the EGU had its third highest proportion 
of heat input spent above the SCR threshold during this time period, 
using that year's weight (the ``third-best weight'') together with the 
SCR-on/SCR-off rates described previously to calculate the weighted 
rate. The years 2011-2021 were analyzed for purposes of the proposal 
because they likely are representative of the time period that 
encompasses the years when the units began to exhibit a greater cycling 
pattern, and it is reasonable to expect that this pattern will continue 
for the foreseeable future.
    Using these data, EPA proposed emissions limitations based on the 
following equation:

(SCR-on weight * SCR-on mean rate) + (SCR off weight * SCR off mean 
rate) = emissions limit in lb/MMBtu.

    Using this equation, EPA proposed the NOX emission 
limits listed in Table 1, based on a 30-day rolling average:

                                 Table 1--Proposed NOX Emission Rate Limits \12\
----------------------------------------------------------------------------------------------------------------
                                                                                                    Proposed
                                                Low range rate    High range     Weighted rate  facility-wide 30-
            Facility name                Unit     (lb/MMBtu)       rate (lb/      (lb/MMBtu)    day average rate
                                                                    MMBtu)                      limit (lb/MMBtu)
----------------------------------------------------------------------------------------------------------------
Cheswick.............................        1           0.085           0.195           0.099             0.099
Conemaugh............................        1           0.071           0.132           0.091             0.091
Conemaugh............................        2           0.070           0.132           0.094  ................
Homer City...........................        1           0.102           0.190           0.102             0.088
Homer City...........................        2           0.088           0.126           0.088  ................
Homer City...........................        3           0.096           0.136           0.097  ................
Keystone.............................        1           0.046           0.170           0.076             0.074

[[Page 53385]]

 
Keystone.............................        2           0.045           0.172           0.074  ................
Montour..............................        1           0.047           0.131           0.069             0.069
Montour..............................        2           0.048           0.145           0.070  ................
----------------------------------------------------------------------------------------------------------------

    EPA solicited comment on the proposed facility-wide average rate 
limits, as well as the low and high range of potential limits. The 
limits are calculated as a 30-day rolling average, and apply at all 
times, including during operations when exhaust gas temperatures at the 
SCR inlet are too low for the SCR to operate, or operate optimally. For 
facilities with more than one unit, EPA proposed to allow facility-wide 
averaging for compliance, but proposed that the average limit be based 
on the weighted rate achieved by the best performing unit. A 30-day 
average ``smooths'' operational variability by averaging the current 
value with the prior values over a rolling 30-day period to determine 
compliance. While some period of lb/MMBtu values over the compliance 
rate can occur without triggering a violation, they must be offset by 
corresponding periods where the lb/MMBtu rate is lower than the 
compliance rate (i.e., the 30-day rolling average rate). EPA is 
retaining its proposed overall approach to developing these limits, but 
for reasons discussed in Section III of this preamble, EPA is changing 
the way the rate calculation is done for facilities with more than one 
unit, and is making additional adjustments to the rate calculation in 
response to technical information received. These changes result in 
some changes to the final rates, which are discussed in section IV of 
this preamble.
---------------------------------------------------------------------------

    \12\ See 87 FR 31806 (May 25, 2022).
---------------------------------------------------------------------------

C. Daily NOX Mass Emission Rates

    EPA also proposed a unit-specific daily NOX mass 
emission limit (i.e., lb/day) to complement the weighted facility-wide 
30-day NOX emission rate limit and further ensure RACT is 
applied continuously. High emissions days are a concern, given the 8-
hour averaging time of the underlying 1997 and 2008 ozone NAAQS. The 
proposed daily NOX mass emission limit was calculated by 
multiplying the proposed facility-wide 30-day rolling average 
NOX emission limit (in lb/MMBtu) by each unit's heat input 
maximum permitted rate capacity (in MMBtu/hr) by 24 hours. While the 
30-day average rate limit ensures that SCR is operated where feasible 
while reasonably accounting for cycling, EPA is concerned that units 
meeting this limit might still occasionally have higher daily mass 
emissions on one or more days where no or limited SCR operation occurs, 
which could trigger exceedances of the ozone NAAQS if these high mass 
emissions occur on days conducive to ozone formation, such as 
especially hot summer days. EPA proposed a daily mass limit that would 
govern over a full 24-hr, calendar day basis as an additional 
constraint on SCR-off operation within a single day. The proposed limit 
was designed to provide for some boiler operation without using the 
SCR, which may be unavoidable during part of any given day, but also to 
constrain such operation because the mass limit will necessitate SCR 
operation (for example by raising heat input to a level where the SCR 
can operate) if the unit is to continue to operate while remaining 
below this limit. This provides greater consistency with the RACT 
definition. Table 2 shows the unit-specific daily NOX mass 
limits that were proposed in the NPRM.

                                  Table 2--Proposed Daily NOX Mass Limits \13\
----------------------------------------------------------------------------------------------------------------
                                                                                Permitted max
                                                                                 hourly heat     Proposed unit-
                           Facility name                               Unit      input rate       specific mass
                                                                               (MMBtu/hr) \14\   limit (lb/day)
----------------------------------------------------------------------------------------------------------------
Cheswick...........................................................        1             6,000            14,256
Conemaugh..........................................................        1             8,280            18,084
Conemaugh..........................................................        2             8,280            18,084
Homer City.........................................................        1             6,792            14,345
Homer City.........................................................        2             6,792            14,345
Homer City.........................................................        3             7,260            15,333
Keystone...........................................................        1             8,717            15,481
Keystone...........................................................        2             8,717            15,481
Montour............................................................        1             7,317            12,117
Montour............................................................        2             7,239            11,988
----------------------------------------------------------------------------------------------------------------

    EPA solicited comment on the proposed daily mass limits. As 
discussed in more detail in section III of this preamble, EPA 
considered the comments received and made some changes to the final 
limits. The final limits are discussed in section IV of this preamble.
---------------------------------------------------------------------------

    \13\ See 87 FR 31807 (May 25, 2022).
    \14\ Title V Permit maximum heat input rates.
---------------------------------------------------------------------------

III. EPA's Response to Comments Received

    EPA received 10 sets of comments on our May 25, 2022 proposed FIP. 
A summary of the comments and EPA's response is provided herein. All 
comments received are included in the docket for this action.
    Comment: Allegheny County Health Department (ACHD) submitted a

[[Page 53386]]

comment clarifying the operating status of the Cheswick Generating 
Station.
    Response: EPA acknowledges the comment provided by ACHD. In our 
NPRM, EPA described Cheswick as being in the process of closing, 
despite ACHD having issued a title V permit modification that included 
a provision requiring Boiler #1 to cease operations on April 1, 2022. 
While that deadline had come and gone by the time the NPRM was 
published, it was not entirely clear at the time of drafting the notice 
that the closure was permanent and enforceable. ACHD's comment 
addressed EPA's characterization of Cheswick's status in the NPRM and 
affirmed that ACHD has verified that Cheswick's main boiler and 
associated equipment have been permanently shut down. In the 
intervening months since the NPRM, EPA has confirmed, with assistance 
from ACHD, that the boiler has in fact ceased operating, and that 
Cheswick's title V operating permit has been terminated. Therefore, EPA 
finds that the closure is permanent and enforceable, and as such, is 
not finalizing any RACT limits for Cheswick as proposed in our NPRM.
    Comment: Commenters assert that EPA must take action on PADEP's May 
26, 2022 and June 9, 2022 SIP submittals, which included Pennsylvania's 
own source specific RACT determinations, and which were intended to 
address the deficiencies identified by the Third Circuit, prior to (or 
concurrently with) promulgating a FIP.
    Response: Although EPA generally pursues a ``state first'' approach 
to air quality management, giving deference to states to determine the 
best strategy for addressing air quality concerns within their 
boundaries in the first instance, EPA does not agree with the 
commenters' assertion that EPA must act on PADEP's RACT SIP submittals 
prior to or concurrently with finalizing a FIP. On September 15, 2021, 
EPA proposed to disapprove those portions of Pennsylvania's May 16, 
2016 SIP upon which EPA's prior approval had been vacated and remanded 
by the Third Circuit, and that are encompassed in this FIP action. 86 
FR 51315. EPA recently finalized that disapproval. 87 FR 50257. CAA 
section 110(c)(1)(B) requires the Administrator to ``promulgate a 
Federal implementation plan at any time within 2 years after the 
Administrator disapproves a State implementation plan submission in 
whole or in part, unless the State corrects the deficiency and the 
Administrator approves the plan or plan revision, before the 
Administrator promulgates such Federal implementation plan'' (emphasis 
added). Following EPA's August 16, 2022 (87 FR 50257) final 
disapproval, EPA has authority to promulgate a FIP under CAA section 
110(c) at any time because EPA has not approved a plan or plan revision 
from Pennsylvania correcting the deficiency. Nothing in the Clean Air 
Act requires EPA to act upon a SIP submitted by a state to address a 
deficiency identified in EPA's final disapproval prior to promulgating 
a FIP, and the commenters have not provided any statutory basis for 
such a position.
    As explained in the NPRM for this action, EPA may promulgate a FIP 
contemporaneously with or immediately following the predicate final 
disapproval action on a SIP (or finding that no SIP was submitted). EPA 
v. EME Homer City Generation, L.P., 572 U.S. 489, 509 (2014) (``EPA is 
not obliged to wait two years or postpone its action even a single day: 
The Act empowers the Agency to promulgate a FIP `at any time' within 
the two-year limit'') (internal citations omitted). In order to provide 
for this, it cannot be true that EPA must take further action on SIP 
submittals from the state prior to undertaking rulemaking for a FIP. 
The practical effect of applying the procedure commenters allege, that 
EPA must consider a new SIP submittal from the state prior to 
promulgating a FIP, would be that EPA would either approve the state's 
new SIP revision (thereby nullifying the need for a FIP) or EPA would 
disapprove the state's new SIP revision, which would essentially 
require a double disapproval from EPA in such circumstances. This 
cannot be understood to be Congress's intent. When considering a 
similar question, the Federal Court of Appeals for the Tenth Circuit 
agreed with the interpretation EPA here states. Specifically, the Tenth 
Circuit stated: ``The statute itself makes clear that the mere filing 
of a SIP by Oklahoma does not relieve the EPA of its duty. And the 
petitioners do not point to any language that requires the EPA to delay 
its promulgation of a FIP until it rules on a proposed SIP. As the EPA 
points out, such a rule would essentially nullify any time limits the 
EPA placed on states. States could forestall the promulgation of a FIP 
by submitting one inadequate SIP after another.'' Oklahoma v. EPA, 723 
F.3d 1201, 1223 (10th Cir. 2013) (emphasis in original).
    EPA has not fully evaluated Pennsylvania's May 26 and June 9, 2022 
submittals and has not yet proposed action on the SIP submittals. As 
explained, this does not alter EPA's authority to finalize this action 
promulgating a FIP. EPA intends to evaluate and take action on 
Pennsylvania's submittal in accordance with the timelines established 
in CAA section 110(k)(2). However, as noted in the NPRM, EPA submitted 
extensive comments on the draft permits. In those comments, EPA raised 
several concerns that remain unresolved, including whether 
Pennsylvania's continued use of tiered limits (i.e., separate limits 
for SCR-on and SCR-off operation) could be squared with the Court's 
clear objection to our approval of such an approach in the past, and 
whether Pennsylvania's record was adequate to support the limits 
selected, the need for separate limits, and how to determine when each 
limit applied.
    Comment: Several commenters asserted that EPA erred in the 
selection of SCR as RACT. PADEP asserts that EPA's proposal does not 
provide a source specific analysis of technological feasibility for 
each unit, and that it does not identify any specific control 
technology or technique as being technically feasible. They claim that 
EPA's approach fails to comport with previous RACT approaches. 
Keystone/Conemaugh (Key-Con) suggests that EPA overlooked the technical 
and economic circumstances of the individual sources in determining 
RACT. Additionally, one commenter, Talen Energy, alleged that EPA 
should have selected feasible controls that ``represent RACT for each 
mode of operation of the units, such as startup and shutdown.''
    Response: EPA disagrees with those comments suggesting that EPA's 
FIP proposal did not follow the long-standing definition of RACT. 
Courts have repeatedly concluded that the term ``reasonably available'' 
is ambiguous and therefore the statute does not specify which emission 
controls must be considered ``reasonably available.'' See, e.g., 
Natural Resources Defense Council v. EPA, 571 F.3d 1245, 1252 (D.C. 
Cir. 2009) (stating ``the term `reasonably available' within RACT is 
also ambiguous'' and ``[g]iven this ambiguity, the EPA has discretion 
reasonably to define the controls that will demonstrate compliance''). 
See also, Sierra Club v. EPA, 294 F.3d 155, 162-63 (D.C. Cir. 2002) 
(finding that the term ``reasonably available'' in the analogous 
``reasonably available control measure'' is ambiguous and ``clearly 
bespeaks [the Congress's] intention that the EPA exercise discretion in 
determining which control measures must be implemented''). As stated in 
the proposal, EPA's longstanding interpretation is that RACT is defined 
as ``the lowest emission limitation that a particular source is capable 
of meeting

[[Page 53387]]

by the application of control technology that is reasonably available 
considering technological and economic feasibility.'' \15\ Commenters 
correctly note that EPA has further explained that ``RACT for a 
particular source is determined on a case-by-case basis, considering 
the technological and economic circumstances of the individual 
source.'' \16\
---------------------------------------------------------------------------

    \15\ Memo, dated December 9, 1976, from Roger Strelow, Assistant 
Administrator for Air and Waste Management, to Regional 
Administrators, ``Guidance for Determining Acceptability of SIP 
Regulations in Non-Attainment Areas,'' p. 2, available at https://www3.epa.gov/ttn/naaqs/aqmguide/collection/cp2/19761209_strelow_ract.pdf and 44 FR 53762, footnote 2 (September 17, 
1979) (Strelow Memo). See also Sierra Club v. EPA, 972 F.3d 290.
    \16\ Id.
---------------------------------------------------------------------------

    EPA's action is in line with this longstanding guidance and other 
Agency actions concerning RACT under section 182 of the Clean Air Act. 
For each source, the EPA first selected a control technology that is 
reasonably available, considering technical and economic feasibility, 
and then identified the lowest emissions limitation that, in EPA's 
judgment, the particular source is capable of meeting by application of 
the technology (i.e., that a plant operator applying the selected 
technology is capable of achieving economically and technologically). 
With respect to the first step, for this set of sources EPA selected 
SCR as the control technology that is reasonably available. For each of 
the sources addressed in this final rule, SCR has already been 
installed and each SCR has a clearly demonstrated operating history. 
Most of the sources installed these SCRs in the early 2000s, with the 
exception being Conemaugh, which only installed SCR in 2014. These 
facts alone prove that SCR is a control technology that is reasonably 
available for these sources. In the prior EPA-approved PADEP SIP 
revision, SCR was selected as the control technology and that selection 
was not disputed in comments on the action or in the subsequent 
litigation, to which this FIP is a response. Additionally, no one 
raised concerns about whether SCR was the appropriate control 
technology when EPA initially proposed approval of PADEP's RACT 
regulations, nor did anyone raise such concerns at the State level when 
PADEP undertook notice and comment rulemaking in order to adopt the 
regulation in the first place. To the extent that the commenters are 
challenging EPA's judgment in choosing the emission limit that each 
source is ``capable of meeting,'' those comments are addressed later in 
this section. However, if the commenters are asserting that EPA has 
selected a technology that is not ``reasonably available considering 
technological and economic feasibility,'' the EPA disagrees based on 
the fact that SCRs are present and operating at each of these sources.
    Regarding the comment that EPA should select RACT limits for each 
mode of operation of the SCR, including startup and shutdown, the 
proposed FIP accounts for this. Given that these sources already have 
installed and operational SCRs, EPA determined it was appropriate to 
consider modes of operation, as applicable, during the selection of the 
emission limitation, rather than during the control technology 
selection. Indeed, EPA's proposed statistical approach to develop the 
rates is intended to select emissions limits that reasonably account 
for different modes, including consideration of modes where the 
selected RACT cannot be operated. As discussed in a comment response 
later in this document, EPA considered whether it was appropriate to 
create a tiered limit approach that also accounted for different modes 
in the different tiers, but as explained here and in the proposal, were 
EPA to define a mode where the chosen RACT technology need not operate 
but also fail to provide constraints on the use of that mode, that 
would essentially create an exemption from operating RACT when the 
source is clearly capable of meeting a lower rate, and would thereby 
create a regulatory incentive to operate at loads where the SCR is not 
in operation.
    Comment: PADEP claims that it is inconsistent with RACT to use a 
statistical approach for the selection of emissions limits. Key-Con 
similarly claims that routine data are insufficient for a RACT 
analysis.
    Response: As an initial matter, EPA affirms that a statistical 
approach is a valid way to select the lowest emissions limit that the 
source is capable of meeting through application of SCR. As explained 
in the response to the prior comment, once a technology is selected 
that is ``reasonably available considering technological and economic 
feasibility,'' the second step is selection of the emission limit that 
a plant operator applying the selected technology is economically and 
technologically capable of achieving. In order to select the emission 
limitation, EPA did an extensive statistical analysis of emissions data 
from the affected facilities. The rationale underlying that approach is 
outlined in significant detail in our proposal.
    EPA does not always have the benefit of a robust historic data set 
that reflects actual operation of the selected control technology to 
consider in selecting emission limits for purposes of establishing 
RACT. When, as is the case here, we do have such data, it is reasonable 
to use them. The proposal acknowledged several factors that affect the 
degree to which the historic data set represents the lowest rate that 
the source is capable of meeting and explains the adjustments EPA made 
to its proposed emissions limits to account for those factors. There 
are specific comments that take issue with certain choices EPA made in 
applying the statistical approach, which EPA addresses later in this 
notice, but nothing in the CAA or EPA rules or guidance precludes EPA 
from using a statistical approach as it has done here.
    Comment: PADEP takes issue with EPA's decision to not do a 
technical and economic feasibility analysis for other potential 
NOX control technologies at these sources, such as 
installation of newer low-NOX burners that achieve greater 
NOX reductions during the combustion process. Key-Con 
provided similar comments, asserting that our failure to analyze each 
of these other potential NOX control technologies for their 
economic and technological feasibility was not in keeping with RACT. 
These commenters took issue with EPA's presumption ``that the 
facilities have the flexibility to change their operations to emit less 
NOX per unit of heat input.''
    Response: The statements discussing other potential NOX 
control technologies that could be adopted, but that EPA was not 
requiring, were provided as additional information, and as noted in the 
proposal, ``EPA did not evaluate these technologies in the context of 
our RACT analysis.'' Commenters appear to assume that EPA expressly 
accounted for installation or increased use of these technologies when 
determining limits that each source is capable of meeting. To the 
contrary, this discussion was intended to clarify that these other 
control techniques were not accounted for in EPA's development of each 
source's limits; neither the rates nor the weights were adjusted to 
require more use of these other control technologies. To the degree 
that a source was using such other control technologies during the 
period used in selecting the RACT limits, EPA's approach for developing 
the limits assumed that the sources continued to operate these other 
technologies without any change.
    Also, although PADEP did an analysis of other NOX 
control technologies available to each source when setting the limits 
in the permits, PADEP rejected all of these other control technologies 
except boiler tuning, either

[[Page 53388]]

for technical feasibility or cost reasons, in setting the limits. This 
rejection of most of the other control technologies as RACT by PADEP 
essentially aligns with our own selection of SCR as RACT.
    Comment: Homer City objects to applying the RACT limit from the 
lowest emitting of the three sources at the facility as a facility-wide 
RACT NOX limit. Homer City asserts that the definition of 
RACT, i.e. ``. . .the lowest emission limit that a particular source 
[emphasis added] is capable of meeting. . .'' requires that EPA 
establish FIP limits on a unit by unit basis, rather than by a facility 
wide average.
    Response: Longstanding EPA policies have allowed for averaging to 
meet RACT limits, including averaging across multiple emissions units. 
The 1992 NOX supplement to the general preamble \17\ states 
that it is appropriate for RACT to allow emissions averaging across 
facilities within a nonattainment area (or Ozone Transport Region (OTR) 
state, as is the case here). In practice EPA has allowed averaging 
across units on a facility-wide basis, and even across facilities in 
the same system under common control of the same owner/operator, 
including its approval of PADEP's prior EGU RACT rules.\18\ EPA's 
implementation rule for the 2008 ozone NAAQS allows nonattainment areas 
to satisfy the NOX RACT requirement by using averaged area-
wide emissions reductions.\19\ EPA reasonably allows averaging for 
compliance, so long as the underlying rates used as the basis for the 
average meet the definition of RACT. The comments do not provide a 
basis for EPA to reject its longstanding emissions averaging policies. 
To the contrary, these policies provide additional flexibility for 
sources to manage their SCR operation across units to ensure compliance 
with the limits.
---------------------------------------------------------------------------

    \17\ 57 FR 55620, November 25, 1992
    \18\ See 25 Pa Code Sec. Sec.  129.94 and 129.98, which allow 
sources which cannot meet a presumptive RACT limit to average with 
lower emitting sources, provided that aggregate emissions do not 
exceed what would have been allowed under the presumptive limits.
    \19\ 80 FR at 12278-79 (``states may demonstrate as part of 
their NOX RACT SIP submittal that the weighted average 
NOX emission rate from all sources in the nonattainment 
area subject to RACT meets NOX RACT requirements''). This 
portion of the 2008 ozone SIP requirements rule was challenged, with 
petitioners arguing that such a rule violated the Clean Air Act 
because the statute at Sec.  182(b)(2) requires each individual 
source to meet the NOX RACT requirement. The D.C. Circuit 
rejected this argument, finding that the Clean Air Act ``does not 
specify that `each one of' the individual sources within the 
category of `all' `major sources' must implement RACT.'' South Coast 
Air Quality Mgmt Dist. v. EPA, 882 F.3d 1138, 1154 (D.C. Cir. 2018).
---------------------------------------------------------------------------

    Regarding the comments on EPA's proposal to base the facility-wide 
average rate on the best performing unit, the EPA is finalizing a minor 
change. In light of the unit-specific nature of EPA's weighted rate 
analysis, the EPA expects that the unit-specific rates already 
represent RACT for each unit, and that the most appropriate basis for a 
facility-wide average would be the weighted rates for each of the units 
at the facility. While some commenters felt that EPA should use the 
lowest single unit rate to drive facilities to use their best 
performing units most often, we expect that the stringent unit-specific 
rates, when averaged together, will still provide sufficient incentive 
to use the best performing units most often. See section IV of the 
notice for additional information.
    Comment: Key-Con notes that only one of the designated 
nonattainment areas in Pennsylvania is currently violating the 2015 
ozone NAAQS, and expresses concern that EPA appears to have 
inappropriately considered the potential for lower ozone levels in many 
areas in setting RACT, and states that the requirement for 
NOX RACT is simply tied to Pennsylvania's inclusion in the 
OTR. Key-Con also asserts that it is more appropriate to use interstate 
transport rules, not RACT, to address concerns about states' 
obligations to eliminate significant contribution to nonattainment, or 
interference with maintenance of NAAQS in other states.
    Response: The EPA agrees with the commenter's characterization that 
Pennsylvania must implement RACT level controls statewide due to the 
state's inclusion in the OTR, in accordance with CAA Sec.  184. The 
statutory direction to require ``implementation of reasonably available 
control technology'' in states included in an ozone transport region, 
CAA Sec. Sec.  182(f), 184(b), is the same in substance as the 
requirement for ozone nonattainment areas for ``implementation of 
reasonably available control technology,'' CAA Sec.  182(b)(2). 
Therefore, EPA's analytical method to determine what level of control 
technology is reasonably available does not differ based on whether 
RACT is being implemented in an ozone nonattainment area or the OTR.
    There are also areas of Pennsylvania that are still designated 
nonattainment for both prior and current ozone NAAQS. EPA notes that 
the implication of the commenter's statement, that an area's factual 
attainment of an ozone NAAQS, as perhaps shown by a Clean Data 
Determination, would have implications for whether that area needs to 
implement RACT, is incorrect. An area designated nonattainment must 
continue to meet the statutory requirement to implement RACT, if 
otherwise applicable, until the area is redesignated to attainment or 
unclassifiable under section 107(d)(3) of the CAA. While the EPA did 
identify improved air quality in many areas, including remaining ozone 
nonattainment areas, some of which are in other states, as a benefit of 
the FIP emissions limits, we did not determine RACT through the 
selection of control technology and identification of emission 
limitations that the sources are capable of meeting based on the air 
quality impact in any particular area(s). In other words, air quality 
improvement in nonattainment areas in Pennsylvania or other states was 
not a criterion in determining RACT in this action.
    Comment: Several commenters claim that EPA's economic feasibility 
analysis for SCR optimization was flawed. First, commenters assert that 
the economic analysis was flawed because it only considered the costs 
of additional reagent, and ignored considerable capital costs such as 
increased catalyst maintenance and replacement, and modifications to 
ancillary equipment. Second, commenters assert that the actual $/ton 
NOX costs far exceed what EPA's analysis claims, and are 
more likely in the $150,000-200,000/ton range. Additionally, commenters 
assert that EPA's analysis of reagent injection incorrectly assumes 
that reagent costs will return to historic, lower prices.
    Response: EPA disagrees. First, commenters are incorrect in the 
assertion that EPA did not consider capital costs, such as catalyst 
maintenance and replacement. As discussed in the NPRM and TSD, EPA 
relied on certain data from the recent evaluation of variable operating 
and maintenance (VOM) costs (which include increased catalyst 
maintenance and replacement costs), associated with increased use of 
SCRs at EGUs used in a number of national rulemaking actions related to 
the CAA's interstate transport requirements, including most recently 
the proposed Good Neighbor Plan for the 2015 ozone NAAQS. In the ``EGU 
NOX Mitigation Strategies Proposed Rule TSD'' (Good Neighbor 
Plan TSD) for the proposed Good Neighbor Plan (included in the docket 
for this action), EPA used the capital expenses and operation and 
maintenance costs for installing and fully operating emission controls 
based on the cost equations used within the Integrated Planning Model 
(IPM) that were researched by Sargent & Lundy, a nationally

[[Page 53389]]

recognized architect/engineering firm with EGU sector expertise. See 87 
FR 31808; TSD at 16-18. EPA's cost analysis for the proposed FIP only 
related to increased use, or optimization, of the SCRs, since each 
facility already had SCR installed. While that analysis was presented 
on a national, fleetwide basis, for this action EPA used site specific 
data in the ``Retrofit Cost Analyzer'' \20\ to perform a bounding 
analysis to demonstrate that the cost assumptions made in the RCU and 
Good Neighbor Plan were still accurate and reasonable for the current 
RACT analysis. Using that methodology, EPA estimated a cost per ton for 
these sources that ranged from $2,590 to $2,757, depending on the unit. 
As previously stated, these estimates did include capital costs 
associated with increased catalyst maintenance and replacement. Reagent 
costs have actually dropped since the May 25, 2022 NPRM,\21\ and the 
cost per ton of NOX removed is still well within a range 
that should be considered economically feasible.
---------------------------------------------------------------------------

    \20\ See TSD for proposed FIP at 16-18.
    \21\ Reagent prices have decreased since publication of the 
NPRM, from an average of $1515/ton anhydrous ammonia to slightly 
less than $1400/ton. See appendix 3 of the TSD for this action, and 
https://mymarketnews.ams.usda.gov/filerepo/sites/default/files/3195/2022-07-28/614317/ams_3195_00065.pdf.
---------------------------------------------------------------------------

    In Table 4 of the TSD for the proposed FIP, EPA calculated the 
potential change in NOX mass emissions, based on the 
proposed 30-day average NOX emission limits.\22\ Then, in 
Table 5 of the proposed TSD, EPA calculated the cost per ton of 
NOX removed based on the additional amount of reagent needed 
to meet to those limits.\23\ EPA has made slight adjustments in 
finalizing the emission limits after considering comments. Detailed 
discussion of the rationale for and of the limits themselves can be 
found elsewhere, but particularly in section IV of this preamble. Table 
3 of this preamble shows the reductions these limits will realize when 
compared to 2021 emissions data.
---------------------------------------------------------------------------

    \22\ See Id. at 15.
    \23\ See Id. at 19.

                       Table 3--2021 Annual NOX Emissions and Rates Compared to FIP Rates
----------------------------------------------------------------------------------------------------------------
                                   2021                                                     Potential
                               Average NOX  30-Day  NOX    30-Day NOX       2021 NOX     change in  NOX
           Facility             rate  (lb/   rate (lb/   rate  vs. 2021     emissions         Mass
                                  MMBtu)       MMBtu)      average (%)       (tons)         Emissions
                                                                                             (tons)
----------------------------------------------------------------------------------------------------------------
Conemaugh....................        0.149        0.072             -52           5,506          -2,837
Homer City...................        0.133        0.096             -28           3,144            -871
Keystone.....................        0.142        0.075             -47           5,481          -2,579
Montour......................        0.110        0.102              -7             649             -46
                              ----------------------------------------------------------------------------------
    Net......................  ...........  ...........  ..............          14,781          -6,333     -43%
----------------------------------------------------------------------------------------------------------------

    Based on the revised limits, and an updated cost of reagent, EPA 
calculated the cost per ton of NOX removed for the final 
limits:

                        Table 4--Cost per NOX ($/ton) Removed Based on Additional Reagent
----------------------------------------------------------------------------------------------------------------
                                       Predicted                                               Cost per ton  of
                                   reduction  (tons       Additional       Total annual cost    NOX removed for
            Facility              NOX per year  from  reagent  (tons per    for  additional   additional reagent
                                    2021  baseline)     year  from 2021     reagent [caret]         ($/ton) +
                                                          baseline) *
----------------------------------------------------------------------------------------------------------------
Conemaugh.......................               2,837               1,617          $2,263,800                $798
Homer City......................                 871                 496             694,400                 797
Keystone........................               2,579               1,470           2,058,000                 798
Montour.........................                  46                  26              36,400                 791
                                 -------------------------------------------------------------------------------
    Average cost/ton............  ..................  ..................  ..................                 796
----------------------------------------------------------------------------------------------------------------
* Additional reagent = predicted reduction (tons) x 0.57 tons reagent/ton NOX reduction.
[caret] Total cost = additional reagent x $1400/ton reagent.
+ Cost per ton = total cost/predicted reduction.

    With respect to the assertion by commenters that the $/ton value is 
actually in the $150,000-$200,000/ton of NOX removed range, 
commenters have not supplied adequate data or analysis to substantiate 
that assertion. Commenters (in this case, Montour) merely assert that 
in order to meet the proposed limits, the units will need to run for 
extended periods of time following a startup, even when electricity is 
not being sold to the grid, in order to achieve a certain number of 
hours of low hourly NOX emissions rates to offset the higher 
hourly NOX emission rates during startup, or else the source 
will not meet the proposed emission limits in the FIP. Montour claims 
that it has more frequent start-ups and shut-downs during which it 
cannot operate the SCRs. EPA notes that the comment did not provide any 
analysis of potential alternate methods of compliant operation, and 
merely submitted data relating to the extra cost of fuel oil during the 
period of time they assert they will be required to run. For example, 
it may be possible for the units to ramp up more quickly following 
startup so as to spend less time in SCR-off mode. Additionally, it may 
be possible for the units to spend more time ``hovering'' at a higher 
heat input (i.e. SCR-on) in anticipation of a need for quick dispatch. 
EPA acknowledges that the limits in the FIP may result in

[[Page 53390]]

the sources' needing to re-evaluate how they operate their EGUs in 
order to meet the new RACT limit, which may require adjusting the 
prices and certain operating parameters they specify to PJM when 
bidding into the market. However, EPA views these as free-market 
considerations, rather than an appropriate component of a RACT 
determination. EPA has long held that ``[e]conomic feasibility rests 
very little on the ability of a particular source to `afford' to reduce 
emissions to the level of similar sources. Less efficient sources would 
be rewarded by having to bear lower emission reduction costs if 
affordability were given high consideration. Rather, economic 
feasibility . . . is largely determined by evidence that other sources 
in a source category have in fact applied the control technology in 
question.'' \24\
---------------------------------------------------------------------------

    \24\ E.P.A., State Implementation Plans; General Preamble for 
the Implementation of Title I of the Clean Air Act Amendments of 
1990; Supplemental, 57 FR 18,070, 18,073 (proposed April 28, 1992) 
(first introducing RACT as a standard to regulate emissions from 
existing sources).
---------------------------------------------------------------------------

    EPA continues to believe that optimization of the SCRs to achieve 
the NOX emission limits in this FIP is economically 
feasible. Nothing submitted in the comments provided adequate 
justification or data to make a determination to the contrary. Indeed, 
evidence from the units' operating history supports EPA's view that 
when it is economically advantageous to do so, these units have no 
trouble meeting lower limits. Some of the lowest NOX 
emissions EPA observed coincided with high NOX allowance 
prices associated with the NOX SIP call which went into 
effect in 2003.\25\ Additionally, data for some of these units from May 
through June of the 2022 ozone season generally indicate SCR operating 
patterns (and, as a result, NOX emissions) that match or are 
among their best in the recent data record. EPA believes this is due, 
at least in part, to the market prices of NOX allowances 
needed for compliance with the RCU during this period, which were 
reported to range between $20,000 and $40,000 per ton.\26\
---------------------------------------------------------------------------

    \25\ Finding of Significant Contribution and Rulemaking for 
Certain States in the Ozone Transport Assessment Group Region for 
Purposes of Reducing Regional Transport of Ozone (NOX SIP 
Call), 63 FR 57356 (October 27, 1998) (codified in relevant part at 
40 CFR 51.121 and 51.122).
    \26\ See S&P Global Capital IQ, capitaliq.spglobal.com 
(subscription required).
---------------------------------------------------------------------------

    Comment: Commenters assert that EPA ignored equipment failure 
issues and failed to consider the deleterious effects on both control 
equipment and on the environment (ammonia slip, decreased mercury 
removal) of excess ammonia injection, particularly when operating below 
the catalysts' minimum effective temperature range. Commenters further 
assert that EPA failed to consider an engineering analysis submitted by 
Key-Con that PADEP relied upon in developing their case-by-case limit 
for Key-Con.
    Response: EPA disagrees. First, EPA did not presume that the 
proposed FIP limits would be met by simply injecting more reagent 
during sub-optimal SCR operating conditions, and the FIP does not 
require it. EPA continues to recognize that the NOX 
reduction capabilities of the SCRs are flue gas temperature dependent, 
and that the NOX removal efficiency curve decreases with 
flue gas temperature until a point is reached where the SCR offers 
little or no NOX control above what is achieved by the low 
NOX burners (LNB) and overfire air (OFA) that are also 
installed on all of the units subject to this FIP. We also recognize 
that catalyst fouling, catalyst poisoning, ammonia slip and damage to 
downstream equipment are all potential outcomes of excessive reagent 
injection or injection during low temperature conditions. We further 
recognize that there have been changes in the electricity market in 
more recent years that result in greater periods of time when the units 
are operating in SCR-off mode. EPA believes that because the 
calculation of the limits uses actual past performance data from the 
sources, which include times at low heat input and therefore time with 
the SCR off, sources can meet these limits without injecting excessive 
amounts of ammonia during unfavorable SCR operating conditions. 
Additionally, using the third-best weight means that the SCR-off weight 
is based on a recent year that is not the extreme SCR-on case in the 
last decade and thus provides additional buffer.
    The data show that during times when boilers are operating at high 
heat inputs and therefore SCRs are at optimum performance temperatures, 
sources have shown that they are capable of achieving limits in the 
0.05 to 0.07 lb/MMBtu range, so they could achieve additional 
reductions during times when the SCR can be optimized to offset higher 
emissions during times when the SCR may not be optimized, so as to meet 
their 30-day rolling average and daily mass limit.
    Also, EPA did review and consider the Key-Con engineering report 
referenced by the commenters. The information presented in that report 
appears to have been submitted to Pennsylvania to contest condition 
E.009 in PADEP's draft case-by-case RACT permit for Keystone, which 
would have required Keystone to set the SCR controllers at a target 
NOX emission rate of 0.06 lb/MMBtu.\27\ According to 
Attachment 3 of Key-Con's comment letter, they additionally evaluated 
operational data from 2019, which they claim is the last year of 
typical operations.\28\ The report evaluated ammonia injection rates, 
and purported to show that due to ammonia slip and fouling of 
downstream appurtenances, the SCR could not and should not operate at a 
set-point of 0.06 lb NOX/MMBtu. The report then determined 
that ``a NOX rate of 0.09 lb/MMBtu is tolerable and will not 
require air heater washes nearly as frequently as 0.08 lb/MMBtu \29\ or 
less would.'' See page 10 of Appendix 3 to Key-Con's July 11, 2022 
comment letter. The report also states that Key-Con conducted testing 
on Conemaugh unit 1 during 18 days in May 2017 to determine if 
continuous operation at a NOX setpoint of 0.04 lb/MMBtu was 
sustainable. The report claimed that it was not, because emissions of 
mercury spiked to a point where it appeared that Unit 1 would exceed 
its Mercury Air Toxics Standard (MATS) limit, and the NOX 
setpoint had to be increased to 0.07 lb/MMBtu to lower mercury 
emissions. A similar test was conducted on Conemaugh Unit 2 towards the 
end of the 2017 ozone season to determine if the 0.05 lb/MMBtu setpoint 
was sustainable, and the report claims that after 25 days at the 0.05 
setpoint, mercury emissions increased abruptly and nearly exceeded the 
MATS limit, so the NOX setpoint had to be ``relaxed'' an 
unspecified amount to decrease mercury emissions. P. 7 of Attachment 3.
---------------------------------------------------------------------------

    \27\ Per condition E.10 of the draft permit for Conemaugh, their 
target was 0.05 lb NOX/MMBtu
    \28\ Commenters assert that 2020 and 2021 were excluded due to 
low electricity demand and lack of coal supply, respectively.
    \29\ PADEP's proposed RACT limit.
---------------------------------------------------------------------------

    In response to the report, EPA notes that unlike Pennsylvania's 
proposed RACT permit terms, EPA is not requiring that the sources 
operate their SCRs at a certain set point below the 30-day rolling 
daily average NOX rate limit, so the validity and relevance 
of this testing to EPA's proposed limits is questionable. EPA is 
expecting that the operators of Keystone and Conemaugh will operate 
their SCRs in a way that balances concerns about catalyst and preheater 
fouling and mercury emissions with the emission rates set by EPA--rates 
which are based on operating data from these sources indicating 
achievement of these emission rates in the past, including the recent 
past. Also, we note that EPA's pounds of NOX per MMBtu of 
heat

[[Page 53391]]

input emission rate limit is a 30-day rolling daily average emission 
rate limit, whereas its daily limit is a mass limit. In contrast, 
Pennsylvania's RACT permit had a daily (24 hour) average NOX 
emissions rate, so EPA's 30-day rolling average emission rate limit 
gives the source operators more flexibility in how they operate the 
SCRs. That is, the operators do not need to keep the setpoint for the 
SCRs at a very low level each day for an extended period of time, as 
they would to meet Pennsylvania's daily average NOX rate. 
The ability to average NOX hourly emission rates over 30 
days allows the sources greater flexibility to vary NOX 
emission rates from their SCRs, raising NOX emission rates 
up or down in order to balance the various factors that must be taken 
into account, such as catalyst or preheater fouling and mercury 
emissions.
    Finally, EPA notes that the commenter did not perform a ``thorough 
review of EPA's NOX emissions analyses'' because of EPA's 
alleged technical failures and failure to understand current and 
expected unit utilizations.\30\ However, the commenter did not provide 
any information regarding expected unit utilization, and instead 
criticized EPA's proposed rates as unobtainable during startup events 
by providing 25 hours of minimal data regarding one cold-start of 
Keystone Unit 1 in January 2022. Given that this data covered only 25 
hours of startup, and was not then averaged with 29 other days of 
emission data to arrive at a 30-day average hourly emission rate, it is 
not proof that this one unit could not meet EPA's 30-day average rate. 
Absent more robust data to support commenter's claim, EPA declines to 
amend its proposed rates for the four units at Keystone and Conemaugh 
based on the thin data presented.
---------------------------------------------------------------------------

    \30\ P. 11 of Key-Con's July 11, 2022 comments.
---------------------------------------------------------------------------

    Comment: PADEP asserts that EPA's weighted rate approach is flawed 
because it relies on an analysis of past averages, which is contrary to 
the court's instruction that ``. . . an average of the current 
emissions being generated by existing systems will not usually be 
sufficient to satisfy the RACT standard.''
    Response: EPA disagrees with the commenter's contention that the 
analysis underlying EPA's RACT limits is flawed simply due to the fact 
that EPA uses the mathematical function of averaging as part of the 
Agency's overall calculation. As the commenter notes, the Sierra Club 
decision does include language noting that ``an average of the current 
emissions being generated by existing systems, will not usually be 
sufficient to satisfy the RACT standard.'' 972 F.3d at 300. However, in 
the preceding sentence, the court provides necessary context for its 
statement and a helpful summary of what Pennsylvania provided in its 
prior SIP, EPA's approval of which the Court was vacating. The Court 
notes that the chosen emission limitation ``was selected as it 
represents the average pollution output of the three plants that are 
already compliant over the past five years.'' Id. Therefore, the court 
did not take issue with the mathematical function of averaging; it took 
issue with the quantity being averaged, and its application in setting 
RACT. EPA does not believe that the court meant to forbid the use of 
any averaging in the determination of RACT, so long as it fit within 
the definition of RACT and the use of such averaging was adequately and 
reasonably explained in the record.
    As explained elsewhere in this action, EPA has used a statistical 
approach to establish the emission limitations contained in this FIP, 
which necessarily involves averaging. However, there are significant 
and meaningful differences between EPA's use of averaging and how PADEP 
previously used averaging to determine the RACT limits at issue in the 
Sierra Club decision. While Pennsylvania's limit was based on a five-
year ozone season average from three plants that were then averaged 
together again to calculate a single limit required at five different 
sources, EPA's approach uses a source-specific third-best ozone season 
rate from a larger range of data. EPA's approach is consistent with the 
RACT definition, including the interpretation of RACT contained in the 
Sierra Club decision, because it is aimed at representing the lowest 
rate the source is technologically and economically capable of 
achieving, not the average rate it has already achieved. (As explained 
elsewhere in this action, EPA used third-best to represent the source's 
current capability, but the approach is still aimed at defining the 
lowest rate, rather than a 5-year overall average).
    Comment: PADEP asserts that EPA's FIP is flawed because it relies 
on the third-best approach used in the RCU and Good Neighbor Plan, 
which is inappropriate because those rules evaluated more current data 
sets, and that EPA's data set selection is not driven by RACT 
regulations or guidance and does not set source specific limits 
considering technological and economic feasibility.
    Response: EPA proposed to use the third-best ozone season rate for 
each source based on the idea, which was also cited in both the RCU and 
the Good Neighbor Plan, that the performance of SCRs degrades over 
time, and that usually only one layer of catalyst is changed/
refurbished per year. Therefore, the SCRs may never be able to achieve 
the same emission reduction rate as when they started operating and all 
three catalyst layers were new. With the exception of the Conemaugh 
plant, which installed its SCRs in late 2014, the other sources 
installed their SCR by 2003.\31\ Thus, many other parts of the overall 
SCR system, such as the reagent injection system, may also have 
deteriorated in performance. The use of the third-best year for each 
source is consistent with EPA's past practices in other rulemakings, 
and also has a basis in the performance data of each source. The third-
best approach is a reasonable way of determining appropriate RACT 
limits. It avoids biasing the SCR-on limit with uncharacteristically 
low emitting ozone seasons, or under uncharacteristically optimal 
operating conditions. As stated in the April 6, 2022 proposed Good 
Neighbor Plan, the EPA found it prudent not to consider lowest or 
second lowest ozone season NOX emissions rates, which may 
reflect SCR systems that have all new components. Such data are 
potentially not representative of ongoing achievable NOX 
emission rates considering broken-in components and routine maintenance 
schedules. Additionally, the fact that CSAPR and the Good Neighbor Plan 
establish caps rather than limits does not preclude the use of the 
third-best approach for the purposes of the FIP. EPA is finalizing the 
use of the third-best year for all of the facilities except Conemaugh. 
As discussed elsewhere in this action, EPA has determined it is 
appropriate to use a different approach for establishing final RACT 
limits for Conemaugh due to the fact that Conemaugh has newer SCRs. As 
further discussed in section IV of this preamble, Conemaugh's final 
limit was calculated using the second-best rate and the second-best 
weight due to the more limited data set of years available for this 
facility based on the more recent installation of SCR.\32\
---------------------------------------------------------------------------

    \31\ As noted in the NPRM, the limits proposed for Conemaugh 
were based on the second-best ozone season, since Conemaugh's SCR 
was only installed in late 2014 and EPA therefore doesn't have the 
same volume of operating data as for the other sources.
    \32\ The proposed limit used the second best rate and the third 
best weight.
---------------------------------------------------------------------------

    Regarding the claim that the RCU and Good Neighbor Plan used more 
current data sets, this is because those rulemakings were undertaken 
under a completely different statutory provision with different 
requirements and purpose than this FIP. Both the RCU and Good

[[Page 53392]]

Neighbor Plan FIPs were addressing the requirement in section 
110(a)(2)(D)(i)(I) of the CAA to ensure that emissions from upwind 
sources, including EGUs, were not significantly contributing to 
nonattainment or interfering with maintenance in downwind areas. The 
RCU addressed upwind significant contributions to downwind areas for 
the 2008 ozone NAAQS, while the proposed Good Neighbor Plan addressed 
upwind emissions for the 2015 ozone NAAQS. As such, for both rules, EPA 
needed to use the most recently available and up-to-date data for both 
source emissions and ambient air monitoring results in order to 
identify upwind emissions currently affecting downwind monitors for the 
2008 and 2015 ozone NAAQS. Here, the purpose is to identify RACT, as 
required under subsections 182(b)(2), 182 (f)(1), and 184 of the CAA, 
which requires that major sources of NOX and/or VOCs in 
nonattainment areas, or in the OTR, meet RACT, which EPA defines as 
``the lowest emission limit that a particular source is capable of 
meeting by the application of control technology that is reasonably 
available considering technological and economic feasibility.'' Given 
this different purpose, the examination of historic operating data for 
the SCRs is relevant to the determination of the NOX 
emission rates each source attained while running their SCRs, and which 
the source was therefore capable of meeting. Also, EPA did consider 
ozone season emission rates from each source through 2021, which was 
the most recent data available at the time of the proposal, so PADEP's 
claim that EPA did not consider recent data is incorrect.
    Comment: PADEP further asserts that EPA's FIP is flawed because it 
only considers ozone season data, so fails to consider emissions for a 
major part of the year. Commenters claim the court acknowledged that 
their presumptive limit did account for seasonal variability. They cite 
to Motor Vehicles Mfrs. Ass'n of the U.S., Inc. v. State Farm Mut. 
Auto. Ins. Co., 463 U.S. 29, 43 (1983) (``State Farm'') (Providing that 
``the agency must examine the relevant data and articulate a 
satisfactory explanation for its action including a ``rational 
connection between the facts found and the choice made,'' and claim 
that because EPA failed to consider the majority of the operational 
emissions data (i.e., non-ozone season), EPA failed to adequately 
demonstrate that the proposed limits are technically and economically 
feasible year-round.
    Response: EPA disagrees with PADEP's claim that EPA should consider 
non-ozone season data for several reasons. Although these sources were 
subject to the CAIR annual NOX requirements starting in 2009 
and the CSAPR annual NOX requirements starting in 2015, 
these cap and trade programs initially set annual NOX 
emission budgets for states based on a NOX emission rate of 
0.15 lb/MMBtu starting in 2009, then based on a cost-effectiveness 
level starting in 2015, and allowed individual sources to exceed their 
allocated allowances by a certain percent by purchasing additional 
NOX allowances from other sources. As such, the non-ozone 
season emissions data beginning in 2009 does not necessarily reflect 
the NOX emission rates these SCRs are capable of achieving 
outside of the ozone season because the SCRs were not required to meet 
a specific NOX emission rate. Second, post-2017 (when 
Pennsylvania's RACT II limit of 0.12lb/MMBtu was effective), data show 
the sources generally did not operate the SCRs for significant time 
periods outside of ozone season. Hourly operating data submitted by 
Keystone and Conemaugh to PADEP show that in 2017, the SCRs did not 
consistently operate outside of ozone season, with the units at each 
source often cycling down to low heat inputs at night and therefore not 
operating their SCRs.\33\ Third, Pennsylvania also based the 0.12 lb/
MMBtu emission rate in its RACT II rule solely on ozone season 
emissions data. Finally, PADEP does not explain why EPA's determination 
of RACT for these sources would be altered by consideration of non-
ozone season data.
---------------------------------------------------------------------------

    \33\ For examples of this SCR-off operation, see the xl 
spreadsheet in the docket entitled ``KEY_Hourly emissions and 
operating data 2017-2020_06-24-21.'' For Keystone Unit 1, see 
February 5th to 28th, 2017, and for Unit 2 see October 1 through 
30th, 2017. For Conemaugh, see the spreadsheet in the docket 
entitled ``CON_Hourly emissions and operating data 2017-2020_6-24-
21.'' For Unit 1, see January 21 through 23rd, 2017 and for Unit 2 
see April 15th through 17th, 2017.
---------------------------------------------------------------------------

    Comment: Several commenters objected to EPA's methodology (and 
thus, results) in calculating the SCR-on/SCR-off thresholds. PADEP in 
particular asserts that by assigning an operating threshold for SCR 
operation at each facility, EPA has run afoul of the Court's objection 
to the 600-degree threshold in Pennsylvania's original RACT II 
regulation. Further, PADEP asserts that because EPA had only limited 
information from Key-Con and none from the other facilities, and 
because we failed to seek such information from the other facilities, 
the resulting emission limits are unsupported. Another commenter 
asserted that EPA's visual evaluation of scatterplot data to develop 
the thresholds was flawed, and that rather than accurately depicting 
the SCR-on/SCR-off thresholds, the diagrams actually depict the minimum 
sustainable load for the unit, which is ``. . . typically the level at 
which PJM places a unit at low load for spinning reserve during periods 
of low demand.'' See Homer City Comments at 2. Additionally, commenters 
assert that the use of 0.2 lb/MMBtu as an indicator of when the SCRs 
are or are not running is arbitrary, since there are times when an SCR 
is off, but the NOX emissions are below 0.2 lb/MMBtu, and 
conversely, there are times when an SCR is running, but the 
NOX emissions are greater than 0.2 lb/MMBtu.
    Response: First, EPA disagrees with Pennsylvania's assertion that 
the methodology for determining the SCR-on and SCR-off weights and 
rates using observed SCR thresholds in the data for purposes of 
developing an emissions limit that would restrict SCR-off operation is 
substantially similar to PADEP's use of the 600-degree threshold to 
justify essentially unlimited SCR-off operation. EPA further disagrees 
that the Sierra Club adverse decision concerning the 600-degree 
threshold has direct relevance to the permissibility of the approach 
used by EPA in utilizing SCR-on and SCR-off weights and rates. The 
Court found that Pennsylvania's blanket 600-degree temperature 
threshold, which Pennsylvania applied uniformly to all the sources 
regardless of the differences in SCRs at each source, was inadequately 
explained or supported by the record. 972 F.3d at 303 (``Regarding the 
threshold, neither the EPA nor DEP can explain why it is necessary at 
all. . . . [E]ven assuming such a temperature threshold were 
reasonable, the record does not support the conclusion that 600 degrees 
Fahrenheit is the proper limit.'') EPA's SCR-on and SCR-off thresholds 
were derived through careful unit-by-unit observation of actual 
operating data. Furthermore, rather than drawing a regulatory line 
below which less stringent emissions limits apply without any 
restriction on operating time, EPA used the 0.2 lb/MMBtu threshold to 
divide the operational data into SCR-on and SCR-off categories, then 
used those data to establish both average SCR-on and -off rates for 
each unit, and to identify the unit's past percentage of ozone season 
time with the SCR on or off to establish the weight applied to the 
respective rates. As such, the 0.2 lb/MMBtu is not an enforceable 
limit, but merely a data point that was one component of EPA's approach 
to use historical operating data to derive the lowest emission limit 
that these particular sources are capable

[[Page 53393]]

of meeting by the application of control technology that is reasonably 
available considering technological and economic feasibility.
    As for the assertion that the 0.2 lb/MMBtu cutpoint is arbitrary, 
EPA conducted a fleetwide analysis of EGUs with combustion and post-
combustion NOX controls and found that this rate indicates 
that the SCR is running to some extent.\34\ Nevertheless, in response 
to our May 25, 2022 (87 FR 31798) proposal, EPA did in fact receive 
additional information from certain sources (Montour and Homer City) 
regarding what they consider the proper megawatt (MW) threshold for 
operation of their SCRs. As described in section IV of this preamble, 
we have taken that information into account in developing the 
NOX emission limits finalized in this action.
---------------------------------------------------------------------------

    \34\ See ``Attachment 3-1 NOX Rate Development in EPA 
Platform v6'' for EPA's Power Sector Modeling Platform (IPM) at 
https://www.epa.gov/system/files/documents/2022-02/attachment-3-1-nox-rate-development-in-epa-platform-v6-summer-2021-reference-case.pdf.
---------------------------------------------------------------------------

    Comment: PADEP asserts that EPA's statistical approach to RACT in 
this case has led to absurd results, specifically a higher limit for 
Conemaugh than for Homer City and Keystone, despite the fact that 
Conemaugh's SCRs are newer and technically capable of achieving lower 
NOX emission rates.
    Response: EPA has developed the emissions limits for each source 
based on analysis of historical data for each source demonstrating what 
emissions the sources are capable of achieving through operation of 
their installed SCR equipment. The emission limits being established 
for Keystone are based on analysis of historical data extending back to 
2003, while the emissions limits being established for Conemaugh are 
based on historical data extending only back to 2015 due to the more 
recent SCR installations at Conemaugh. Because the shorter historical 
period of the Conemaugh data set does not contain periods with high 
NOX allowance prices that would necessarily have motivated 
Conemaugh to try to achieve the lowest possible emissions, it is 
possible that EPA's resulting emissions limits for Conemaugh are less 
stringent than would have been established with a more extensive data 
set. However, the limitations of the data available for Conemaugh in no 
way render the Keystone emission limits unreasonable. Nevertheless, the 
comment does illustrate that EPA should adjust its approach to account 
for the more limited Conemaugh data. As further discussed in section IV 
of this preamble, in response to comments received, EPA is finalizing 
limits that differ slightly from what was proposed, including an 
adjustment for Conemaugh that better accounts for the more limited set 
of ozone seasons from which to draw data for this source, while also 
addressing the circumstances that prompted the PADEP comment regarding 
absurd results. The Agency determined that for Conemaugh, it is 
reasonable to use the second-best weight instead of the third-best.
    Comment: PADEP asserts that EPA should have considered tiered 
limits as they did, and that such a limit structure would, in fact, 
result in optimized SCR operation.
    Response: EPA disagrees that we needed to establish a tiered limit 
structure like the one that was vacated by the Court, or the similar 
approach used by PADEP in their case-by-case permits. As explained in 
the proposal and the earlier section of this preamble, EPA did consider 
the appropriateness of tiered limits and opted to not propose such an 
approach for several reasons. First, while the Court did not explicitly 
preclude the threshold approach, they were clearly suspicious of its 
appropriateness: ``Regarding the threshold, neither the EPA nor DEP can 
explain why it is necessary at all. It is not a common exemption.'' 
Sierra at 20. Upon reconsideration, EPA believes that it is not 
necessary. EPA continues to believe that constraining SCR-off operation 
to the extent possible based on data reflecting the recent operations 
of each source is the appropriate means of implementing emission limits 
consistent with RACT. As EPA raised in the on-record comments we 
submitted to PADEP on draft permits,\35\ it is not clear to EPA how a 
tiered limit approach constrains SCR-off operation in any meaningful or 
enforceable way.\36\ Moreover, unconstrained SCR-off operation would be 
inconsistent with the Court's directive that the RACT limit must be 
technology-forcing.\37\ A set of limits that does not place limits on 
the source operating without its NOX control technology is 
not technology-forcing. Accordingly, EPA has chosen to forgo the tiered 
limit approach, and instead use a weighted rate approach, which we 
continue to believe provides the sources flexibility to address current 
operational realities (i.e., increased cycling), while at the same time 
providing meaningful constraint on SCR-off operation and objective 
enforceability.
---------------------------------------------------------------------------

    \35\ See document ID EPA-R03-OAR-2022-0347-0067 in the docket 
for this action at www.regulations.gov.
    \36\ EPA has not yet evaluated and is not pre-determining the 
approvability Pennsylvania's ultimate SIP revisions, which were 
submitted on May 26, 2020 and June 9, 2022.
    \37\ Sierra Club at 309.
---------------------------------------------------------------------------

    Comment: Talen Energy (Montour) asserts that EPA's limits are so 
restrictive that they extend the regulatory regime beyond the customary 
regulation of air pollutant emissions, and in effect dictate operation 
of units and may severely limit the ability of the units to run as 
directed by PJM and potentially compromise grid reliability.
    Response: EPA disagrees that these FIP limits are too restrictive 
or that they extend the regulatory regime beyond EPA's Clean Air Act 
authority or customary EPA action in a way that is inappropriate or 
inconsistent with past CAA implementation. Emission limitations are, by 
definition, a limitation on the amount of pollutants that may be 
emitted by a source and therefore all emission limits place 
restrictions on how sources operate in some fashion. For example, 
states or EPA may place enforceable requirements on sources for 
throughput limitations; federally enforceable requirements of this 
nature are a standard practice that substitutes for major source 
applicability of new source review (NSR) or national emission standards 
for hazardous air pollutants (NESHAPs). Some emission limitations may 
also take the form of work practice standards, which could place 
requirements on the type of fuel a source may use or limit the amount 
of time a source may operate under a certain status. These FIP limits 
do not prescribe when or how the affected units should operate in order 
to generate electricity. Rather, these limits ensure that when the 
units are operating, their already installed SCRs are also operated in 
a way that achieves the lowest emission rates that are technically and 
economically feasible.
    As discussed previously in this notice, EPA acknowledges that the 
weight given to the proposed SCR-off limit has the effect of limiting 
the portion of time a cycling source can operate in SCR-off mode and 
incentivizes a source to shift to SCR-on mode to preserve headroom 
under the limit. While driving SCR operation, the weighted limit 
accommodates the need for an EGU to occasionally cycle down to loads 
below which SCR can operate effectively. Nothing in the FIP being 
finalized in this document is intended to prohibit SCR-off operation, 
nor does it dictate specific times when SCR-off operation would not be 
permitted to occur.
    Comment: Montour commented that the compliance date should be 
extended and not be the same date as the effective

[[Page 53394]]

date of the regulation. Citing the need to identify and evaluate the 
updates/changes necessary, update programming for the CEMS and process 
control equipment, provide training to staff, and complete operational 
trials, Montour suggested extending the compliance date by six months. 
Other sources commented that EPA should not proceed at all with a final 
rule at this time and instead seek an extension from the Court to 
reconsider the proposed limits.
    Response: Before addressing the substance of this comment, EPA 
would like to correct an error in the NPRM regarding the effective date 
of the FIP. The effective date of the regulation was intended to be 
conveyed as an editorial note that the rule would be effective 30 days 
after publication of the final rule. Instead, the editorial note was 
converted into an actual date by the publisher, which was 30 days after 
the date the proposed rulemaking was published: June 24, 2022. This was 
a typographical error that produced an absurd result: the rule could 
not possibly be effective before a final approval, or indeed, even 
before the public comment period had ended (on July 11, 2022). The 
proposed compliance date was accurately described to ``commence 
immediately upon the effective date.'' \38\
---------------------------------------------------------------------------

    \38\ The proposal erroneously published the effective date of 
the rule as June 24, 2022 and not as an editorial note that the rule 
would be effective 30 days after the publication of the final rule. 
See 87 FR 31813.
---------------------------------------------------------------------------

    With regard to Montour's request to extend the compliance date, EPA 
agrees there will be a certain amount of time required for the 
facilities to adjust to the new requirements and make certain technical 
and administrative changes to ensure operations comply with the new 
RACT limits. After considering comments received on this rulemaking, 
EPA has determined that it is appropriate to extend the compliance date 
past the initial proposal of 30 days after the effective date of these 
regulations. The commenters have raised compelling concerns about being 
able to meet new, more stringent limits on the accelerated timeline. In 
light of the comment received from Montour, EPA is finalizing a 
compliance date of 180 days after the effective date of the FIP. EPA is 
under Court Order to ``. . . either approve a revised, compliant SIP 
within two years or formulate a new [FIP],'' which EPA interprets as 
requiring a final rule by August 27, 2022. Therefore, EPA will finalize 
the final rule in compliance with the Court.
    Comment: Homer City asserted that EPA's description of the 
methodology for determining SCR-on and SCR-off weighting is inadequate 
to allow for independent verification. Also, Homer City also commented 
that there is no explanation as to why the SCR-off weights (0.00 or 
0.01) are so small, which leave no margin for SCR-off operation.
    Response: The commenter did not provide adequate explanation as to 
why or where it had difficulty in understanding or replicating the 
calculations EPA outlined in the proposed notice. Homer City also did 
not submit its attempted calculations for EPA's consideration. All of 
the data EPA used to develop the proposed emission limits (including 
that which was used to establish the SCR-on and SCR-off weights) was 
either available in the docket, or, because of file type and size 
limitations of www.regulations.gov, was available upon request.\39\ 
Other commenters were able to replicate and/or modify EPA's 
methodology. Homer City's weights are representative of their ozone 
season operation over the time period analyzed for the weights (2011 to 
2021). Further discussion of their revised weights can be found in 
section IV of this preamble.
---------------------------------------------------------------------------

    \39\ See ``Memo to Docket--Availability of Additional 
Information,'' document number EPA-R03-OAR-2022-0347-0060.
---------------------------------------------------------------------------

    Comment: Sierra Club asserts that the requirement that the sources 
submit reports of their compliance every six months should be shortened 
to every three months (quarterly), because the information needed to 
demonstrate compliance with the FIP is already submitted to EPA for 
various purposes on a quarterly basis, and that it does not make sense 
for the FIP to require less frequent (biannual) reporting. In addition, 
if EPA elects to keep the FIP reporting data separate from reporting to 
the Clean Air Markets Division, Sierra Club requests that EPA put a 
mechanism into the FIP by which the public can readily access this data 
to ensure compliance, such as posting that data to the Clean Air 
Markets Program Data tool. Finally, the commenter requests that the FIP 
recordkeeping requirements be updated to include information about SCR 
runtime and/or bypass as well as reagent usage.
    Response: EPA selected the six-month reporting period in order to 
be consistent and streamlined with the sources' existing title V 
reporting requirements. These title V reports are submitted to EPA 
Region 3 and the state for review. The fact that certain data used to 
determine compliance with the FIP requirements are also reported 
quarterly to other EPA offices under various programs, such as the Acid 
Rain program and Cross State Air Pollution Rule, and then placed into 
EPA's Clean Air Markets Data Program online tool, does not provide a 
sufficient basis to increase the frequency of reporting compliance with 
the FIP requirements to match the reporting frequency for the 
underlying data. There is nothing about the FIP limits that would 
necessitate a reporting frequency greater than the reporting frequency 
required by title V. The FIP does require deviation reports to be 
submitted to EPA when NOX emission limits have been exceeded 
for three or more days in any 30-day period.
    With respect to the assertion that the reporting requirements 
should be updated to include SCR runtime and reagent injection data, 
EPA believes that reporting of CEMS data consistent with title V 
requirements is sufficient for compliance demonstration purposes. EPA 
has not tied the emission limits directly to SCR operating parameters 
in a way that would necessitate the submission of additional SCR data. 
Compliance with the emission limits is the ultimate regulatory 
requirement, and this is adequately demonstrated through submission of 
CEMS data. EPA does not believe it is appropriate at this time to 
include reporting requirements to this FIP that are not directly 
necessary to show compliance with the regulatory requirements finalized 
herein.
    Regarding the assertion that EPA should provide mechanism by which 
the public can readily access additional data beyond the regularly 
reported emissions data to ensure compliance, such as posting that 
additional data to the Clean Air Markets Program Data tool, EPA is not 
taking that step at this time. There is nothing about the 
NOX limits in this FIP which would require EPA to provide a 
novel approach to providing access to additional compliance data. 
Further, the tools EPA makes available for providing the public with 
access to reported emissions data are not at issue in this proceeding, 
and comments requesting changes to those tools are outside the scope of 
the rule.
    Comment: Sierra Club asserts that EPA should have used the best 
year, rather than the third-best, which is what EPA used in 
establishing the SCR-on rate. First, they assert that EPA has not 
established that control equipment degrades over time, and that by 
selecting the third-best ozone season, EPA is allowing sources to forgo 
maintenance and good operating practices that would allow them to 
otherwise meet limits that were established on a best ozone season 
basis. Further, pointing to the rates achieved during the period of 
2003-2010 when NOX allowance prices were high due to

[[Page 53395]]

the NOX SIP call, Sierra Club asserts that the decline in 
SCR performance is due not to equipment degradation, but to the lack of 
a regulatory requirement to achieve better emissions. Finally, Sierra 
Club asserts that an examination of the best performing years does not 
support the idea that equipment degradation due to the passage of time 
necessarily leads to an inability to meet lower limits, and again 
asserts that higher emissions rates are tied to less stringent 
regulatory requirements rather than equipment degradation.
    Response: EPA disagrees that we should have used the best ozone 
season instead of the third-best to establish the SCR-on rate. First, 
although equipment degradation is not the only consideration we 
evaluated when selecting the third-best approach, it is certainly a 
contributing factor. While degradation can be slowed or mitigated 
through proper operation, there is little question that it occurs and 
can impact the removal efficiency. EPA has explained this previously 
that ``[o]ver time, . . . the catalyst activity decreases, requiring 
replacement, washing/cleaning, rejuvenation, or regeneration of the 
catalyst.'' \40\ EPA acknowledges that catalyst management practices 
can be adapted to address catalyst degradation, but that does not mean 
that the degradation does not occur.
---------------------------------------------------------------------------

    \40\ See https://www.epa.gov/sites/default/files/2017-12/documents/scrcostmanualchapter7thedition_2016revisions2017.pdf at 
16.
---------------------------------------------------------------------------

    In addition, EPA's longstanding interpretation of RACT does not 
require RACT-level controls to be equivalent to the ``best.'' The Court 
agreed with this interpretation in the Sierra Club decision: ``we do 
not suggest that Pennsylvania must achieve the absolute lowest level of 
emissions that is technologically possible for the approved limit to 
satisfy RACT.'' \41\ As explained in the NPRM and in response to the 
previous comment, EPA believes that the third-best approach is a 
reasonable way of establishing appropriate RACT limits. Use of the 
third-best year avoids biasing the limit with uncharacteristically low 
emitting ozone seasons, or under uncharacteristically optimal operating 
conditions.
---------------------------------------------------------------------------

    \41\ 972 F.3d at 302.
---------------------------------------------------------------------------

    EPA does agree with the commenter that there does appear to be a 
correlation between increased SCR operation (and correspondingly lower 
NOX emissions), and periods when new regulatory requirements 
such as CAIR, CSAPR, the CSAPR Update, and the RCU, have created 
meaningfully more stringent NOX emission budgets. More 
stringent emissions budgets can compel EGUs to operate their SCRs more 
often and at lower NOX emission rates to meet these new 
budgets. They accomplish this result by raising the cost of 
NOX allowances, creating an economic incentive for EGUs to 
operate their SCRs more often and at lower NOX emission 
rates to either avoid having to purchase costly allowances or to 
generate NOX allowances to sell. EPA continues to believe 
that our proposed weighted rate approach takes these factors into 
consideration and establishes appropriate limits that are consistent 
with the CAA's RACT requirements.
    Comment: Similar to comments relating to EPA's consideration of 
operating data from years when the units were operating in a base load 
capacity, commenters assert that ozone season operations are not 
consistent with year-round operations and therefore should not be the 
sole timeframe considered in development of the limits that apply all 
the time. Further, Key-Con in particular noted that the SCRs at 
Keystone were designed to only run during ozone season, and that in the 
past, they had considerable down time for cleaning and maintenance of 
the controls. Additionally, they assert that ammonium bisulfate salts 
(ABS) form more readily in colder ambient temperatures, leading to 
increased fouling.
    Response: EPA acknowledges some of the technical challenges 
associated with temperature and SCR activity. Because of this, among 
other reasons, we performed an analysis of actual operating and 
emissions data and developed reasonable limits to account for 
challenges such as seasonal ambient temperature changes and increased 
cycling operation rather than selecting the absolute lowest rates that 
these units have ever achieved. EPA primarily used ozone season data to 
develop these limits, which is appropriate, not only because the ozone 
season generally represents a period of increased electricity demand 
and operation at these sources, but also because it is indicative of 
what these units can achieve when there are additional regulatory 
constraints and economic disincentives against sub-optimal SCR 
operation in place.
    To the degree that the comment is suggesting that this RACT FIP 
should create seasonal limits that do not require SCR operations in 
non-ozone-season months, the EPA does not believe that this would be 
consistent with the CAA RACT requirement. As noted in the background of 
this preamble, NOX RACT for major sources is required to be 
applied year-round. There are numerous coal-fired EGUs operating in the 
OTR that operate SCR controls on an annual basis. Additionally, there 
are coal-fired EGUs operating outside the OTR subject to other 
regulations that mandate SCR controls be operated throughout the year 
as well. Like the four Pennsylvania facilities addressed in this 
notice, many of these other coal-fired EGUs were built in the same era 
(1960s and 1970s) and then later retrofitted with SCRs in response to 
the EPA interstate transport requirements for ozone season 
NOX emissions, which began in 2003. So, while EPA has 
applied RACT on a case-by-case, source-specific basis, EPA cannot 
ignore the fact that there are many coal-fired EGUs, outside of 
Pennsylvania, that can, and do, operate their SCR controls year-round 
with NOX emission limits similar to the final limits 
determined in this notice for the purposes of NOX RACT as 
well as for other regulatory requirements.\42\
---------------------------------------------------------------------------

    \42\ Delaware Administrative Code, Title 7 Natural Resources & 
Environmental Control, 1100 Air Quality Management Section, 1146 
``Electric Generating Unit (EGU) Multi-Pollutant Regulation''.
    Maryland--Code of Maryland Regulations (COMAR), Title 26 
Department of the Environment, Subtitle 11 Air Quality, Chapter 38, 
``Control of NOX Emissions from Coal-Fired Electric 
Generating Units''.
    New Jersey State Department of Environmental Protection, New 
Jersey Administrative Code, Title 7, Chapter 27, Subchapter 19, 
``Control and Prohibition of Air Pollution from Oxides of 
Nitrogen''.
    ``Coal-Fired Power Plant Enforcement'' US EPA, retrieved August 
2022. See https://www.epa.gov/enforcement/coal-fired-power-plant-enforcement.
---------------------------------------------------------------------------

    EPA also disagrees that the Keystone units cannot operate their 
SCRs effectively outside of the ozone season or that the rates must be 
further adjusted to account for seasonal effects. In response to 
Keystone's comment, EPA further reviewed non-ozone season emissions 
data reports for Keystone units and found that between 2009 and 2010, 
both Keystone units operated their SCRs in non-ozone season months for 
extended periods whereby their NOX emissions were generally 
below the final NOX emission limits determined in this 
notice.\43\ Therefore, EPA cannot justify exempting Keystone from 
operating its SCRs, with reasonable effectiveness, for NOX 
RACT during non-ozone season months.
---------------------------------------------------------------------------

    \43\ ``Custom Data Download'' US EPA Clean Air Markets Program 
Data, retrieved August 2022, see https://campd.epa.gov/data/custom-data-download.
---------------------------------------------------------------------------

    Comment: Key-Con asserts that EPA's limits severely and 
inappropriately limit the amount of time either facility can operate 
without ammonia injection, especially during start-up and low load

[[Page 53396]]

operation. They further assert that the duration of a cold start-up is 
18-24 hours, and that at loads between the minimum sustainable load 
(340 MW) and the unit load (which they do not identify) where the 
minimum continuous operating temperature (MCOT) of the SCR is reached, 
emissions can reach 0.35 lb/MMBtu for Keystone units, and 0.30 lb/MMBtu 
for Conemaugh. They assert that Keystone units 1 and 2 in particular 
would be unable to demonstrate compliance if there was one cold start-
up in a 30-day period, even if they spent the rest of the time 
operating at the proposed limit of 0.074 lb/MMBtu.
    Response: Key-Con's comment is not sufficient to demonstrate an 
inability to meet the proposed FIP limits. Key-Con presented no data to 
justify the amount of time spent in a cold start-up during which the 
unit load is above the sustainable limit, but below whatever threshold 
is necessary to bring flue gas up to the MCOT of the SCR and begin 
ammonia injection. As noted in a previous response, Key-Con did not 
provide any information regarding expected unit utilization, and 
instead criticized EPA's proposed rates as unobtainable during startup 
events by providing 25 hours of minimal data regarding one cold-start 
of Keystone Unit 1 in January 2022. Given that this data covered only 
25 hours of startup, and was not then averaged with 29 other days of 
emission data to arrive at a 30-day average hourly emission rate, it is 
not proof that this one unit could not meet EPA's 30-day average rate.
    In response to this comment, EPA further reviewed startup data for 
Keystone in non-ozone season months. On November 5, 2009, Keystone Unit 
1 started operations after having been inoperable since October 20, 
2009. During the first three days of operation, the daily 
NOX emission rates were 0.229, 0.160, and 0.058 lb/MMBtu 
respectively. During the subsequent days of operation, up until 
reaching 30 operating days, the daily NOX emissions varied 
from a low of 0.046 to a high of 0.116 lb/MMBtu. The resultant 30-day 
NOX emission rate after 30 days of operation was 0.064 lb/
MMBtu.\44\ This is well below the final NOX emission rate 
limit determined in this notice of 0.075 lb/MMBtu. This example 
illustrates that the unit is entirely capable of achieving the emission 
rate limits in this notice, with startup periods, provided the normal 
operating days are sufficiently controlled and the facility was able to 
achieve these results without a specific 30-day regulatory requirement 
to do so. Moreover, EPA has purposely granted an emission rate averaged 
over 30 days, which is the maximum averaging time EPA can grant for 
NOX RACT. EPA has also issued facility-wide emission rate 
limits to allow the facilities to further average the emission rates 
amongst their units. This amount of dual averaging, in terms of 
averaging days and then units, affords Key-Con, and the other 
facilities, additional flexibility to manage startup operations.
---------------------------------------------------------------------------

    \44\ See ``Keystone winter-time SCR use unit 1.xlsx'' in the 
docket for this action.
---------------------------------------------------------------------------

    Further, even if we are to accept this claim on its face, Key-Con's 
argument fails because they merely point out the obvious mathematical 
certainty that any appreciable amount of time spent operating above the 
average limit would lead to a violation if the entirety of the 
remaining averaging period was spent operating exactly at the limit. 
The entire purpose of establishing average limits (and in this case a 
30-day average) is to smooth out the peaks and valleys of shorter-term 
emissions and arrive at a limit that can be met by offsetting periods 
when the units emit above the limit (generally, SCR-off periods), with 
periods of optimal operation where the units emit below the limit 
(generally, SCR-on periods). This is one of the reasons that we did not 
select the lowest achievable SCR-on rate as RACT. EPA's limits provide 
for some level of SCR-off operation, while still representing the 
lowest rate the source is capable of meeting over such period through 
the application of control technology that is reasonably available 
considering technological and economic feasibility. To the degree that 
this limit acts as a constraint on low-load operation without the SCR, 
the commenter did not explain why such a constraint is inappropriate. 
In light of the high NOX emissions that can occur with such 
operation, the EPA believes this is a reasonable approach to define a 
limit that represents the application of RACT. Moreover, Key-Con's own 
analysis appears to support an ability to meet 0.075 lb/MMBtu, even 
based on cold start-ups taking place in January.\45\ As discussed in 
section IV of this preamble, EPA has re-evaluated our proposed limits, 
with the resulting limits being consistent with what Key-Con's comments 
appear to show is attainable.
---------------------------------------------------------------------------

    \45\ Id.
---------------------------------------------------------------------------

    Comment: Homer City asserts that because the proposed 24-hour mass 
limits are based on the 30-day average rate limits, the mass limits do 
not provide adequate margin for periods of start-up and shut down.
    Response: EPA disagrees. First, as previously discussed, the 30-day 
rate-based limits upon which the daily mass limits are based were 
derived in such a way as to incorporate several layers of flexibility, 
or margin, including emissions during periods of startup and shutdown. 
We used weighted averages considering years when the units were 
operating in more of a load-following mode rather than as baseload, we 
used a 30-day averaging period to ``smooth'' variability of shorter-
term emissions, and we used the ``third-best'' rather than the ``best'' 
approach in order to add additional buffer and still establish limits 
that represent RACT. Additionally, it is not clear what period of time 
the commenter is considering as ``startup,'' nor have they established 
that they could not begin operating the SCRs sooner. While emission 
rates during the startup process do tend to be higher before the 
control equipment is fully operational, mass emissions are typically 
lower for most startup hours, since startup generally happens at lower 
levels of fuel combustion. Finally, commenters have not presented any 
actual operating data to demonstrate that they cannot meet the proposed 
limits. Indeed, EPA's review of historical data, and in fact, some data 
from the 2022 ozone season reported so far, supports a determination 
that the sources can achieve EPA's final 30-day NOX emission 
rate limits, and that when the units operate in compliance with the 30-
day rate limit, they have generally operated below the final daily 
NOX mass emission limits.
    Comment: Homer City claims that EPA's proposed limits are not 
technically feasible because, they assert, from 2010-2021, only 
Keystone and Conemaugh Units 1 and 2 have been able to achieve EPA's 
proposed limits on a 30-day basis, and even then, it was only 7 
instances or 6.36% of the time.
    Response: First, if sources were not meeting the proposed limits in 
the selected years during which there was no regulatory requirement or 
economic incentive to do so, it is not necessarily proof that they 
could not have. Nor is it proof that they cannot in the future. EPA 
notes that in rejecting EPA's approval of PADEP's original 0.12 lb/
MMBtu limit as ``a mere acceptance of the status quo,'' 972 F.3d at 
302, the Court in Sierra Club affirmed that ``an average of the current 
emissions being generated by existing systems, will not usually be 
sufficient to satisfy the RACT standard,'' id. at 300. Homer City 
rejects EPA's limits, but presents no data or analysis that 
demonstrates what they are in fact capable of achieving, and what EPA 
should establish as RACT for these

[[Page 53397]]

units. EPA has demonstrated that the limits are achievable when the 
regulatory environment requires it, and that the limits in the FIP 
represent RACT for these sources.
    Comment: PADEP asserts that EPA's FIP is based on an incomplete 
record. First, PADEP asserts that EPA ignored information that the 
Department obtained from the sources and failed to obtain additional 
information that would be necessary to conduct a source specific RACT 
analysis. Additionally, PADEP claims that meetings between EPA staff 
and the Maryland Department of the Environment (MDE) prior to our 
proposal may be relevant to the development of the FIP, and that 
records from that meeting should have been in the docket.
    Response: EPA disagrees. First, to the extent it was relevant to 
our approach, we did consider the information that PADEP obtained and 
submitted, and in fact cited to it on numerous occasions, and included 
it in the record as appropriate. EPA had a sufficient technical basis, 
that is thoroughly documented in the rulemaking record, to support the 
RACT limits included in this FIP. To the extent that PADEP or the 
sources at issue in this rulemaking believe the Agency should have 
considered additional or alternative data, the 45-day comment period 
provided an opportunity for the sources to submit such information. EPA 
considered all of the additional information submitted prior to 
finalizing the FIP. With respect to the assertion that records from 
EPA's discussions with MDE prior to EPA proposing this action should 
have been contained in the record, EPA disagrees. All documentation and 
information that EPA relied upon in developing this rule action have 
been included in the record. The cited discussion with MDE did not 
contain information that was relied upon for development of the FIP 
approach and limits.
    Comment: Montour submitted a technical analysis which built upon 
EPA's methodology in the May 25, 2022 (87 FR 31798) NPRM in order to 
demonstrate what they felt are more achievable limits, based on a 
dataset that represents what Montour contends are more consistent with 
current operating parameters. Montour asserts that EPA should have only 
considered ozone season data from 2017-2021, that the correct SCR 
threshold is 440MW, and that as a result, Montour should have a 
facility-wide, 30-day NOX emission rate limit of 0.099 lb/
MMBtu, with daily mass-based limits of 17,385 and 17,200 lb 
NOX/day for Units 1 and 2, respectively.
    Response: As further discussed in section IV of this preamble, as a 
result of comments received and while largely retaining the methodology 
described in the NPRM, EPA has revised some of the limits from the 
proposal based on the submittal of additional data or the 
reconsideration of some of the weights in the case of Conemaugh. 
Specifically, in cases such as Montour where a facility submitted SCR 
threshold data to counter that which EPA used in the proposal, EPA 
recalculated the NOX rate limits using the facility's 
information, but EPA's original methodology. In the case of Montour, 
this recalculation resulted in limits that are very much in line with 
the alternate limits proposed by the facility in its technical 
analysis. Specifically, EPA's methodology resulted in a facility-wide, 
30-day NOX emission rate limit of 0.102 lb/MMBtu, and daily, 
mass-based limits of 17,912 and 17,732 lbs NOX/day for Units 
1 and 2, respectively. In the interest of consistency, EPA is 
finalizing the limits derived from our original methodology rather than 
the alternate limits proposed by Montour. Additionally, because EPA's 
limits are in line with, and in fact very slightly higher than what 
Montour proposed, EPA is not evaluating the remainder of Montour's 
technical analysis.
    Comment: Several commenters assert that because achieving 
compliance with MATS has a negative effect on NOX reduction 
efficiency, EPA should not have considered years prior to MATS 
requirements, and that the limits are therefore too stringent.
    Response: EPA recognizes the co-benefits of SCRs regarding the 
oxidation and ultimate removal of mercury from flue gas. Commenters 
suggest that there is a trade-off between NOX and mercury 
removal, resulting in higher NOX rates to ensure sufficient 
mercury capture. EPA has conducted analysis to evaluate this contention 
in a previous rulemaking. Specifically, to respond to comments received 
on the proposed CSAPR Update, EPA examined ozone-season NOX 
rates from 86 units subject to the MATS rule with SCR and rates below 
0.12 lbs NOX/MMBtu in 2015 (i.e., units that were removing 
the necessary mercury while operating their SCRs during the 2015 ozone 
season). EPA selected the rate cut-off of 0.12 lbs NOX/mmBtu 
to clearly identify units that were operating their SCR. EPA found that 
the average 2015 NOX rate at these 86 units was 0.072 lb/
MMBtu. The average rate for these same units in previous years was 
0.080 and 0.078 lb/MMBtu for 2014 and 2013, which was prior to the MATS 
compliance date when the units would have only needed to optimize 
operations for purposes of NOX removal rather than mercury 
removal. The 2014 and 2013 rates were each statistically significantly 
higher than the rate in 2015 when these units were complying with the 
MATS rule (Student's t-test probability (p) <0.03 and 0.03). Based on 
the CSAPR Update analysis, which is included in the docket for this 
rulemaking,\46\ EPA concludes that units are able to simultaneously 
comply with MATS (i.e., remove mercury from flue gas) while maintaining 
or even lowering their NOX rates, and that the comment 
therefore does not provide a sufficient basis for EPA to exclude data 
from years before MATS implementation from the analysis conducted for 
this rule.
---------------------------------------------------------------------------

    \46\ See MATS Compliance Impact on SCR Control Rates.xlsx.
---------------------------------------------------------------------------

    Comment: Several commenters note the role PJM plays in directing 
the units' dispatch and then assert various implications concerning the 
feasibility or cost of the proposed emissions limits. For example, 
Talen states that ``PJM retains complete and unilateral discretion for 
calling the units to run at certain load profiles. In addition to 
directing Montour SES when to start up the units, PJM's typical 
dispatch also includes the lowering of the unit output down to minimum 
load during off-peak periods daily.'' Talen further states that ``PJM 
dispatch information can dictate the ramp rate of the unit after a 
startup. It is not wholly in Montour SES's control to adjust unit 
operation to fit EPA's proposed model.'' Homer City states that 
``operations today are, in large part, determined by PJM and are beyond 
control of the source operators'' and that the proposed emissions 
limits would not accommodate emissions during ``startups, shutdowns, 
and low-load operations directed by PJM.'' Homer City also asserts that 
sometimes ``[PJM's] direction requires Homer City to operate at levels 
. . . which [do] not allow for operation of the SCR.'' Key-Con states 
that, ``in general'' dispatch of units in the PJM market ``is 
controlled by PJM, not the EGU owner or operator.'' Key-Con suggests 
EPA has assumed that unit owners can choose to ignore PJM's dispatch 
instructions. Key-Con also states that the proposed emission rates 
``will require Key-Con to forfeit most dispatch opportunities at lower 
electrical loads as directed by PJM and suffer resultant revenue 
impacts in order to maintain compliance with the limits.''
    Response: The fact that PJM generally directs the day-to-day and 
hour-to-hour dispatch of the units subject to this rule is not in 
dispute, and any comments

[[Page 53398]]

suggesting that EPA has assumed otherwise mischaracterize the 
proposal.\47\ However, in EPA's view, the consequences that commenters 
assert could result from requirements to follow PJM's dispatch 
instructions are unrealistic because the commenters largely fail to 
acknowledge sources' considerable ability to influence those 
instructions through the offer prices and operating parameters that the 
sources provide to PJM for use in PJM's decision-making process. In 
particular, EPA does not agree with commenters' suggestions that PJM's 
dispatch instructions would create a material obstacle to the sources' 
efforts to comply with the limits in an economic manner. Rather, EPA 
believes it is entirely reasonable to assume, first, that the source 
owners will have the opportunity to consider their emission limits when 
developing the information they supply to PJM for use in PJM's 
decision-making process and, second, that PJM's subsequent dispatch 
instructions will consider the information supplied by the owners when 
determining the dispatch instructions. In other words, contrary to the 
commenter's suggestions, EPA believes that the sources' role as 
suppliers of inputs to PJM's decision-making process means that the 
sources in fact are well positioned to prevent PJM's dispatch 
instructions from interfering with the sources' compliance strategies.
---------------------------------------------------------------------------

    \47\ For example, EPA views Key-Con's extended argument that 
sources do not have incentives to violate PJM's dispatch 
instructions not as an attempt to rebut anything EPA actually said 
in the proposal but rather as the creation and subsequent rebuttal 
of Key-Con's own strawman.
---------------------------------------------------------------------------

    A few examples of the information that sources can specify to PJM 
for use in PJM's decision-making illustrate how the sources covered by 
this rule could cause PJM to issue dispatch instructions that are 
generally compatible with what the source owners consider necessary to 
facilitate effective SCR operation. First, the operating parameters 
that a source can specify include ``Economic Min (MW),'' representing 
the owner's specification of ``the minimum energy available, in MW, 
from the unit for economic dispatch'' under non-emergency 
conditions.\48\ If a source is concerned about the possibility that PJM 
otherwise might direct the unit to run extensively--for example, during 
all or most overnight off-peak hours--at low load levels that would be 
insufficient to maintain SCR inlet temperatures high enough for 
effective SCR performance, the source can avoid that outcome by 
specifying higher values for Economic Min (MW). Second, the operating 
parameters include ``Ramp Rate (MW/Min),'' representing the default 
rate, in MW per minute, for increasing or decreasing a unit's 
output.\49\ If a source is concerned about the possibility that PJM 
would otherwise frequently direct the unit to increase or decrease its 
output at rates that would cause difficulty in sustaining consistent 
SCR performance, the source can avoid that outcome by specifying lower 
values for Ramp Rates. Third, sources can submit cost-based or price-
based values for a variety of parameters associated with unit start-
ups, such as ``Cold Startup Cost,'' ``Intermediate Startup Cost,'' and 
``Hot Startup Cost,'' representing the cost-based or price-based offers 
for the source's compensation for each start-up, differentiated 
according to the unit's temperature before the start-up.\50\ If a 
source believes that its compliance strategy should include efforts to 
reduce start-up emissions by substituting gas or oil for some of the 
coal that would otherwise be combusted during the start-up process, the 
source generally can revise its offered Startup Cost values to reflect 
any resulting changes in start-up fuel cost.
---------------------------------------------------------------------------

    \48\ See the PJM Markets Gateway User Guide (PJM Guide), 
available at https://pjm.com/~/media/etools/markets-gateway/markets-
gateway-user-guide.ashx, at 35.
    \49\ See PJM Guide at 35. Different Ramp Rate values can be 
specified for different portions of a unit's overall load output 
range, and different values can be specified for output increases 
and output decreases. Id. at 38-40.
    \50\ See PJM Guide at 51-53.
---------------------------------------------------------------------------

    EPA recognizes that under certain emergency system conditions, PJM 
may issue dispatch instructions that reflect various ``emergency'' 
parameters rather than the parameters discussed above that would be 
used for economic dispatch under more typical system conditions. EPA 
further recognizes that dispatch instructions issued by PJM in an 
emergency could theoretically require a unit to temporarily operate in 
a manner that precludes effective SCR operation until the emergency 
ends or until PJM can implement alternative measures to address the 
emergency. EPA is also aware that PJM's procedures include lead times 
that may affect how soon sources could change certain elements of the 
information they provide to PJM for use in PJM's decision-making. 
However, EPA believes these considerations are sufficiently addressed 
by the fact that the emission rate limits established in this rule are 
defined as 30-day rolling averages and the fact that EPA is not making 
the requirements established in this rule effective until 180 days 
after the rule's effective date.
    EPA found no information in the comments indicating that the 
sources could not improve their abilities to run their SCRs 
continuously or at improved overall emissions rates by taking advantage 
of opportunities to optimize the values they provide to PJM for offer 
prices and operating parameters, potentially including but not limited 
to Economic Min (MW), Ramp Rate (MW/Min), and Cold, Intermediate, and 
Hot Startup Cost.\51\ Rather, in suggesting that PJM's dispatch 
instructions could conflict with the proposed emission limits, 
commenters relied solely on the fact that the sources generally must 
comply with PJM's instructions once the instructions are issued, with 
no discussion of the process by which PJM determines what its 
instructions should be and no discussion of the sources' own 
opportunities to influence that process.\52\
---------------------------------------------------------------------------

    \51\ In addition to Economic Min (MW), sources can also specify 
``Economic Max (MW),'' representing the owner's specification of the 
maximum energy available from the unit for economic dispatch under 
non-emergency conditions. See PJM Guide at 35. PJM evaluates whether 
the ratio of the value submitted for Economic Max (MW) to the value 
submitted for Economic Min (MW)--known as the ``Turn Down Ratio,'' 
see PJM Guide at 103, falls below a default floor value established 
by PJM for that type of unit. If so, the source must obtain PJM's 
approval for the submitted Economic Min and Economic Max parameter 
values (i.e., an ``exception'' to the Turn Down Ratio default floor 
value) by providing additional information to justify the source's 
submitted values. In an attachment to its comments, Key-Con has 
indicated its awareness of the availability of such exceptions and 
its expectation that PJM would likely be willing to approve 
exceptions if needed to facilitate continuous SCR operation during 
overnight off-peak periods. See Key-Con comments, attachment 3 at 
20-22. Moreover, the operating data reported for Keystone to EPA for 
May and June of 2022 appear to show that Key-Con has in fact 
received approval of such an exception, because the Keystone units' 
ratios of daytime maximum load levels to overnight minimum load 
levels for much of this period fall below the ratio's default floor 
value that would apply to the units in the absence of an exception.
    \52\ The commenters generally chose not to discuss their 
opportunities to influence PJM's dispatch instructions. However, the 
comments do include some implicit recognition that those 
opportunities exist, most of which consist of qualifiers such as 
``in general,'' ``not wholly,'' or ``in large part'' to various 
statements. The clearest confirmation that those opportunities exist 
is found in a statement by Key-Con that the proposed emission rates 
``will require Key-Con to forfeit most dispatch opportunities at 
lower electrical loads as directed by PJM and suffer resultant 
revenue impacts in order to maintain compliance with the limits.'' 
EPA views this statement as an implicit admission that Key-Con has 
the ability to ``forfeit . . . dispatch opportunities'' when it 
believes such forfeiture is in its interest. Given PJM's undisputed 
role in directing units' dispatch, the only mechanism for a source 
to accomplish such a ``forfeiture'' would be for the source to 
provide information to PJM that causes PJM to issue dispatch 
instructions that do not require the units to dispatch at low load 
levels.
---------------------------------------------------------------------------

    Finally, EPA notes that changes in the emissions and operating data 
reported

[[Page 53399]]

by the Conemaugh and Keystone units for the first half of the 2022 
ozone season relative to the data reported by these units for the 2021 
ozone season appear to corroborate EPA's understanding that sources 
have the ability to influence PJM's dispatch decisions. During the 
periods of the 2021 ozone season when these units operated, a frequent 
operating pattern for each of the units was to cycle between a full 
load level of approximately 900 MW during daytime peak hours and a 
lower load level of approximately 440 MW during overnight off-peak 
hours, running their SCRs at the higher daytime loads and turning off 
their SCRs at the lower nighttime loads. During the periods of the 
first half of the 2022 ozone season when the units operated, while they 
continued to display the same general daytime-nighttime cycling 
pattern, the load levels to which they cycled down overnight were 
higher than in 2021, apparently producing flue gas temperatures 
sufficient to allow the units to run their SCRs overnight. 
Specifically, during May and June 2022 the Conemaugh units generally 
cycled down to a load level of approximately 545 MW, and the Keystone 
units generally cycled down to a load level of approximately 700 MW. 
EPA believes the reason for the change in overnight load levels is that 
the sources must have provided higher values of Economic Min (MW) to 
PJM for use in making dispatch decisions during the 2022 ozone season. 
Taking such a step would have increased the likelihood that the units 
would be given dispatch instructions that would allow them to run their 
SCRs continuously and would have been a rational response by the 
sources to the higher reported NOX allowance prices during 
the 2022 ozone season.\53\ In summary, EPA finds these comments 
unpersuasive when appropriately evaluated in the context of sources' 
extensive ability to influence PJM's decision-making, which is 
unchallenged in the comments.
---------------------------------------------------------------------------

    \53\ For the complete hourly data discussed in this paragraph, 
see PA SCR unit 2021-2022 hourly ozone season data.xlsx, available 
in the docket for this action. The spreadsheet contains graphs for 
each unit illustrating the changes in load levels and SCR operation 
described here. EPA notes that the 2022 data have not been used to 
set the emission limits being finalized in this rule but are being 
presented to support EPA's response to the sources' comments 
relating to PJM's control of dispatch decisions.
---------------------------------------------------------------------------

IV. EPA's Final RACT Analysis and Emission Limits

    After consideration of all public comments, the EPA is establishing 
the 30-day NOX Emission Rate Limits in Table 5 and Daily 
NOX Mass Emission Limits in Table 8 for the four facilities 
covered by this FIP to meet the statutory requirement to implement RACT 
for the 1997 and 2008 ozone NAAQS.

 Table 5--Facility-Wide 30-Day Rolling Average NOX Emission Rate Limits
------------------------------------------------------------------------
                                                    Facility-wide 30-day
                   Facility name                     average rate limit
                                                         (lb/MMBtu)
------------------------------------------------------------------------
Conemaugh.........................................                 0.072
Homer City........................................                 0.096
Keystone..........................................                 0.075
Montour...........................................                 0.102
------------------------------------------------------------------------

    The limits in Table 5 are based on a 30-day rolling average, and 
apply at all times, including during operations when exhaust gas 
temperatures at the SCR inlet are too low for the SCR to operate, or 
operate optimally. As discussed in the proposal and in response to 
comments, a 30-day average ``smooths'' operational variability by 
averaging the current value with the prior values over a rolling 30-day 
period to determine compliance. While some period of lb/MMBtu values 
over the target rate can occur without triggering a violation, they 
must be offset by corresponding periods where the lb/MMBtu rate is 
lower than the compliance rate (i.e., the 30-day rolling average rate).
    To calculate the final 30-day rates, EPA used the same weighted 
rate methodology from the proposal, with three key changes. The data 
underlying the weighted rates calculation for each unit is shown in 
Table 6 below.

                                                       Table 6--Unit-Specific Weighted Rates Data
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                                                                                           Facility-wide
                 Facility name                     Unit     SCR on rate    SCR on weight   SCR off rate   SCR off weight   Weighted rate      average
                                                                                (%)                             (%)                        weighted rate
--------------------------------------------------------------------------------------------------------------------------------------------------------
Conemaugh......................................        1           0.070            98.5           0.255             1.5           0.073           0.072
Conemaugh......................................        2           0.070            99.8           0.258             0.2           0.071
Homer City.....................................        1           0.103            99.8           0.341             0.2           0.103           0.096
Homer City.....................................        2           0.087            99.3           0.322             0.7           0.088
Homer City.....................................        3           0.096            99.6           0.292             0.4           0.097
Keystone.......................................        1           0.041            86.7           0.309            13.3           0.076           0.075
Keystone.......................................        2           0.043            88.4           0.312            11.6           0.074
Montour........................................        1           0.045            81.5           0.384            18.5           0.108           0.102
Montour........................................        2           0.047            85.7           0.396            14.3           0.096
--------------------------------------------------------------------------------------------------------------------------------------------------------

    First, using information from the comments, EPA revised the SCR 
thresholds for certain sources. As explained previously, these 
thresholds are applied to the historical data set for the purpose of 
calculating SCR-on and SCR-off rates and weights to calculate the final 
weighted rates. EPA revised the thresholds for Homer City Units 1 and 2 
and Montour Units 1 and 2. Homer City did not provide a revised 
threshold for Unit 3, so the same threshold from

[[Page 53400]]

the proposal was used for the final calculation for that unit. Key-Con 
also did not provide updated thresholds for Keystone and Conemaugh, 
though their thresholds from the proposal were based on comments from 
Key-Con on the recommendation submitted to EPA by the Ozone Transport 
Commission (OTC) under CAA Sec.  184(c).\54\ \55\ Table 7 of this 
preamble shows the thresholds used for the final calculation. As 
previously discussed, based on additional information received during 
the public comment period, the thresholds for Homer City Units 1 and 2 
increased slightly, while the thresholds for Montour increased more 
significantly, as compared to the proposal.
---------------------------------------------------------------------------

    \54\ CAA section 184(a) establishes a commission for the OTR, 
the OTC, consisting of the Governor of each state or their 
designees, the Administrator or their designee, the Regional 
Administrators for the EPA regional offices affected (or the 
Administrator's designees), and an air pollution control official 
representing each state in the region, appointed by the Governor. 
Section 184(c) specifies a procedure for the OTC to develop 
recommendations for additional control measures to be applied within 
all or a part of the OTR if the OTC determines that such measures 
are necessary to bring any area in the OTR into attainment for ozone 
by the applicable attainment deadlines. On June 8, 2020, the OTC 
submitted a recommendation to EPA for additional control measures at 
certain coal-fired EGUs in Pennsylvania. See 85 FR 41972; July 13, 
2020.
    \55\ Conemaugh and Keystone submitted data in response to the 
OTC's CAA section 184(c) recommendation identifying the MW input at 
which it typically operates or can operate the SCRs. EPA reviewed 
the historic operating data for these facilities as it did for Homer 
City, Montour, and Cheswick, and found that Keystone and Conemaugh's 
stated thresholds were consistent with the data. EPA thus relied 
upon the stated values for Keystone and Conemaugh in the development 
of this action's proposed rates.

                             Table 7--SCR Thresholds Used In Weighted Rates Analysis
                                              [Proposal vs. final]
----------------------------------------------------------------------------------------------------------------
                                                                              SCR threshold,     SCR threshold,
                       Facility name                             Unit         proposal  (MW)      final  (MW)
----------------------------------------------------------------------------------------------------------------
Conemaugh.................................................               1                450                450
Conemaugh.................................................               2                450                450
Homer City................................................               1                320                340
Homer City................................................               2                320                335
Homer City................................................               3                320                320
Keystone..................................................               1                660                660
Keystone..................................................               2                660                660
Montour...................................................               1                380                440
Montour...................................................               2                380                440
----------------------------------------------------------------------------------------------------------------

    The threshold changes result in some changes to the data underlying 
the weighted rate calculation for Homer City Units 1 and 2 and Montour 
Units 1 and 2 from the proposal.\56\ The changes to the SCR thresholds 
changed the SCR-on and -off rates for these units very slightly, as 
some hours went from being classified as SCR-on to SCR-off. The SCR-on 
and -off rates for the other units do not change from the proposal, and 
EPA is still using the rate based on the EGU's third-best ozone season 
average from 2003 to 2021 (second-best ozone season average for 
Conemaugh due to its more limited years of SCR data as compared to 
other units). The threshold changes altered the SCR-on and -off weights 
slightly for the Homer City units and substantially for the Montour 
units.
---------------------------------------------------------------------------

    \56\ See Appendix 2 of the TSD for the proposal to compare the 
proposed weights and rates to the final values in Table 6 of this 
preamble.
---------------------------------------------------------------------------

    Second, while EPA is retaining the use of the third-best weight 
(the ozone season in which the EGU had its third highest proportion of 
heat input spent above the SCR threshold) from the period 2011 to 2021 
for Homer City, Keystone, and Montour, EPA is using the second-best 
weight (the ozone season in which the EGU had its second highest 
proportion of heat input spent above the SCR threshold) for Conemaugh. 
As discussed previously in this action and in the proposal, Conemaugh 
installed its SCR much later than the other sources. In response to 
comments pointing out that Conemaugh's proposed limit was the highest 
despite having the newest SCR as well as to account for the more 
limited set of ozone seasons from which to draw data, the Agency 
believes it is reasonable to use the second-best weight instead of the 
third-best. EPA believes that the atypical result pointed out by the 
commenter stems mainly from the fact that using a third-best weight 
from a 7-year data set (as opposed to a third-best weight from an 11-
year data set used for the other sources with more years of SCR data) 
would be more analogous to a mean rate, rather than the lowest rate the 
source was capable of achieving as RACT requires. Given EPA's 
determination, informed by the Court decision, that RACT should 
represent a better rate than a mean rate, we believe that for 
Conemaugh, the second-best weight would provide a more comparable 
weight, while still excluding the low end. This results in a tightening 
of Conemaugh's final limit, as compared to the proposal. EPA still 
believes it is reasonable to use the time period 2011 to 2021 from 
which to draw the weights for Homer City, Keystone, and Montour for the 
final limit. EPA re-examined the occurrence of cycling at these 
facilities and found that the drop in time spent above the SCR 
threshold begins within this time period for these sources.
    Third, as discussed in section III of this preamble, because of the 
unit-specific nature of EPA's weighted rate analysis, the EPA expects 
that the unit-specific rates already represent RACT for each unit, and 
that the most appropriate basis for a facility-wide average would be 
the weighted rates for each of the units at the facility. Therefore, 
EPA is calculating the final facility-wide 30-day limits as an 
arithmetic average of the results of the weighted rates calculation for 
each unit at the facility, instead of applying the best unit-specific 
weighted rate facility-wide.

[[Page 53401]]



     Table 8--Revised Unit-Specific Daily NOX Mass Emissions Limits
------------------------------------------------------------------------
                                                         Unit-specific
            Facility name                   Unit        mass limit  (lb/
                                                              day)
------------------------------------------------------------------------
Conemaugh............................               1             14,308
Conemaugh............................               2             14,308
Homer City...........................               1             15,649
Homer City...........................               2             15,649
Homer City...........................               3             16,727
Keystone.............................               1             15,691
Keystone.............................               2             15,691
Montour..............................               1             17,912
Montour..............................               2             17,721
------------------------------------------------------------------------

    The final daily limits in Table 8, which complement the facility-
wide 30-day rate and further ensure RACT is applied continuously, are 
calculated using the same methodology as the proposal but with the 
updated final 30-day limits as shown in Table 5 of this preamble. The 
final 30-day limits are multiplied by each unit's maximum permitted 
heat input (in MMBtu/hr) by 24 hours.

V. Final Action

    Based on the considerations outlined at proposal, consideration of 
all public comments, and for the reasons described in this action, EPA 
is establishing the 30-day NOX emission rate limits in Table 
5 of this preamble, Daily NOX mass emission limits in Table 
8 of this preamble, and accompanying regulatory language added to 40 
CFR 52.2065, as major stationary source NOX RACT 
requirements for the 1997 and 2008 ozone NAAQS at four facilities in 
Pennsylvania: Conemaugh; Homer City; Keystone; and Montour.

VI. Statutory and Executive Order Reviews

    Additional information about these statutes and Executive Orders 
can be found at https://www.epa.gov/laws-regulations/laws-and-executive-orders.

A. Executive Order 12866: Regulatory Planning and Review and Executive 
Order 13563: Improving Regulation and Regulatory Review

    This final action is a rule of particular applicability and 
therefore is exempt from Office of Management and Budget (OMB) review.

B. Paperwork Reduction Act

    This proposed action does not impose an information collection 
burden under the provisions of the Paperwork Reduction Act (PRA).\57\ A 
``collection of information'' under the PRA means ``the obtaining, 
causing to be obtained, soliciting, or requiring the disclosure to an 
agency, third parties or the public of information by or for an agency 
by means of identical questions posed to, or identical reporting, 
recordkeeping, or disclosure requirements imposed on, ten or more 
persons, whether such collection of information is mandatory, 
voluntary, or required to obtain or retain a benefit.'' \58\ Because 
this proposed rule includes RACT reporting requirements for four 
facilities, the PRA does not apply.
---------------------------------------------------------------------------

    \57\ 44 U.S.C. 3501 et seq.
    \58\ 5 CFR 1320.3(c) (emphasis added).
---------------------------------------------------------------------------

C. Regulatory Flexibility Act

    I certify that this action will not have a significant economic 
impact on a substantial number of small entities under the RFA. This 
action does not affect small governmental jurisdictions or small 
organizations, and the affected entities are not small businesses as 
defined by the Small Business Administration's (SBA) regulations at 13 
CFR 121.201. Therefore, this action will not impose any requirements on 
small entities.

D. Unfunded Mandates Reform Act (UMRA)

    This action does not contain an unfunded mandate of $100 million or 
more as described in UMRA, 2 U.S.C. 1531-1538, and does not 
significantly or uniquely affect small governments.

E. Executive Order 13132: Federalism

    This action does not have federalism implications. It will not have 
substantial direct effects on the states, on the relationship between 
the national government and the states, or on the distribution of power 
and responsibilities among the various levels of government.

F. Executive Order 13175: Consultation and Coordination With Indian 
Tribal Governments

    Executive Order 13175, entitled ``Consultation and Coordination 
with Indian Tribal Governments,'' requires the EPA to develop an 
accountable process to ensure ``meaningful and timely input by tribal 
officials in the development of regulatory policies that have tribal 
implications.'' \59\ This rule does not have tribal implications, as 
specified in Executive Order 13175. It will not have substantial direct 
effects on tribal governments. Thus, Executive Order 13175 does not 
apply to this rule.
---------------------------------------------------------------------------

    \59\ 65 FR 67249, 67250 (November 9, 2000).
---------------------------------------------------------------------------

G. Executive Order 13045: Protection of Children From Environmental 
Health Risks and Safety Risks

    The EPA interprets Executive Order 13045 as applying only to those 
regulatory actions that concern environmental health or safety risks 
that the EPA has reason to believe may disproportionately affect 
children, per the definition of ``covered regulatory action'' in 
section 2-202 of the Executive Order. This action is not subject to 
Executive Order 13045 because it implements a previously promulgated 
health-based Federal standard. Further, the EPA believes that the 
ozone-related benefits from this final rule will further improve 
children's health.

H. Executive Order 13211: Actions Concerning Regulations That 
Significantly Affect Energy Supply, Distribution, or Use

    This action is not subject to Executive Order 13211 (66 FR 28355 
(May 22, 2001)), because it is not a significant regulatory action 
under Executive Order 12866.

I. National Technology Transfer and Advancement Act

    This rulemaking does not involve technical standards.

[[Page 53402]]

J. Executive Order 12898: Federal Actions To Address Environmental 
Justice in Minority Populations and Low-Income Populations

    Executive Order 12898 establishes Federal executive policy on 
environmental justice.\60\ Its main provision directs Federal agencies, 
to the greatest extent practicable and permitted by law, to make 
environmental justice part of their mission by identifying and 
addressing, as appropriate, disproportionately high and adverse human 
health or environmental effects of their programs, policies and 
activities on minority populations and low-income populations in the 
United States.
---------------------------------------------------------------------------

    \60\ Executive Order 12898 can be found 59 FR 7629 (February 16, 
1994).
---------------------------------------------------------------------------

    The EPA believes that this action does not have disproportionately 
high and adverse human health or environmental effects on minority 
populations, low-income populations and/or indigenous peoples, as 
specified in Executive Order 12898. EPA reviewed the Regulatory Impact 
Analysis (RIA) prepared for the recently proposed 2015 Ozone NAAQS 
transport FIP, and in particular the Ozone Exposure Analysis at section 
7.4 of the RIA.\61\ Although that analysis projected reductions in 
overall AS-MO3 ozone concentrations in each state for all affected 
demographic groups resulting from newly proposed limits on EGUs and 
non-EGUs (See Figure 7-3 of the RIA), it also found that emission 
reductions from only EGUs would result in national reductions in AS-MO3 
ozone concentrations for all demographic groups analyzed (See Figure 7-
2 of the RIA). In summation, based on the analysis contained in that 
RIA, EPA has concluded that the FIP is expected to lower ozone in many 
areas, including residual ozone nonattainment areas, and thus mitigate 
some pre-existing health risks of ozone across all populations 
evaluated (RIA, p. 7-32). Further, EPA reviewed an analysis of 
vulnerable groups near the Conemaugh, Homer City, and Keystone EGUs 
found in the TSD for EPA's proposed disapproval of the SO2 
attainment plan for the Indiana, PA SO2 nonattainment 
area.\62\
---------------------------------------------------------------------------

    \61\ The RIA for that separate EPA action can be found at 
www.regulations.gov under the docket number EPA-HQ-OAR-2021-0668. 
Section 7.4 begins on page 7-9.
    \62\ See www.regulations.gov, Docket EPA-R03-OAR-2017-0615-0059, 
pp. 14 -17.
---------------------------------------------------------------------------

K. Congressional Review Act (CRA)

    This rule is exempt from the CRA because it is a rule of particular 
applicability.

VII. Petitions for Judicial Review

    Under section 307(b)(1) of the CAA, petitions for judicial review 
of this action must be filed in the United States Court of Appeals for 
the appropriate circuit by October 31, 2022. Filing a petition for 
reconsideration by the Administrator of this final rule does not affect 
the finality of this action for the purposes of judicial review nor 
does it extend the time within which a petition for judicial review may 
be filed, and shall not postpone the effectiveness of such rule or 
action.
    This action setting RACT limits for certain EGUs in Pennsylvania 
may not be challenged later in proceedings to enforce its requirements. 
(See section 307(b)(2).)

List of Subjects in 40 CFR Part 52

    Environmental protection, Air pollution control, Continuous 
emission monitoring, Electric power plants, Incorporation by reference, 
Nitrogen oxides, Ozone, Reporting and recordkeeping requirements.

Michael S. Regan,
Administrator.

    For the reasons stated in the preamble, the EPA amends 40 CFR part 
52 as follows:

PART 52--APPROVAL AND PROMULGATION OF IMPLEMENTATION PLANS

0
1. The authority citation for part 52 continues to read as follows:

    Authority:  42 U.S.C. 7401 et seq.

Subpart NN--Pennsylvania

0
2. Section 52.2065 is added to subpart NN to read as follows:


Sec.  52.2065   Federal implementation plan addressing reasonably 
available control technology requirements for certain sources.

    (a) Applicability. This section shall apply to Conemaugh, Homer 
City, Keystone, and Montour, as defined in this section, as well as any 
of their successors or assigns. Each of the four listed facilities are 
individually subject to the requirements of this section.
    (b) Effective date. The effective date of this section is September 
30, 2022.
    (c) Compliance date. Compliance with the requirements in this 
section shall commence on March 29, 2023, except the Facility-wide 30-
Day Rolling Average NOX Emission Rate Limit requirement in 
(f)(1) of this section will commence for the Facility on the day that 
Facility has operated for thirty (30) Operating Days after, and 
possibly including, the compliance date of March 29, 2023.
    (d) General provisions. This section is not a permit. Compliance 
with the terms of this section does not guarantee compliance with all 
applicable Federal, state, or local laws or regulations. The emission 
rates and mass emissions limits set forth in this section do not 
relieve the facility from any obligation to comply with other State and 
Federal requirements under the Clean Air Act, including the Facility's 
obligation to satisfy any State requirements set forth in the 
applicable SIP.
    (e) Definitions. Every term expressly defined by this section shall 
have the meaning given to that term within this section. Every other 
term used in this section that is also a term used under the Act or in 
Federal regulations in this chapter implementing the Act shall mean in 
this section what such term means under the Act or the regulations in 
this chapter.
    CEMS or Continuous Emission Monitoring System, means, for 
obligations involving the monitoring of NOX emissions under 
this section, the devices defined in 40 CFR 72.2 and installed and 
maintained as required by 40 CFR part 75.
    Clean Air Act or Act means the Federal Clean Air Act, 42 U.S.C. 
7401-7671q, and its implementing regulations in this chapter.
    Conemaugh means, for purposes of this section, Keystone Conemaugh 
Project LLC's Conemaugh Generating Station consisting of two coal-fired 
units designated as Unit 1 (8,280 MMBtu/hr) and Unit 2 (8,280 MMBtu/
hr), located in West Wheatfield Township, Indiana County, Pennsylvania.
    Day or daily means calendar day unless otherwise specified in this 
section.
    EGU means electric generating unit.
    EPA means the United States Environmental Protection Agency.
    Facility means each of the following as defined in this section: 
Conemaugh; Homer City; Keystone; and Montour.
    Facility-wide 30-Day Rolling Average NOX Emission Rate for the 
Facility shall be expressed in lb/MMBtu and calculated in accordance 
with the following procedure: first, sum the total pounds of 
NOX emitted from all Units during the current Operating Day 
and the previous twenty-nine (29) Operating Days; second, sum the total 
heat input from all Units in MMBtu during the current Unit Operating 
Day and the previous twenty-nine (29) Operating Days; and third, divide 
the total number of pounds of NOX emitted from all Units 
during the thirty (30) Operating Days by the total heat input during 
the thirty

[[Page 53403]]

(30) Operating Days. A new Facility-wide 30-Day Rolling Average 
NOX Emission Rate shall be calculated for each new Operating 
Day. Each 30-Day Rolling Average NOX Emission Rate shall 
include all emissions that occur during all periods within any 
Operating Day, including, but not limited to, emissions from startup, 
shutdown, and malfunction.
    Fossil fuel means any hydrocarbon fuel, including coal, petroleum 
coke, petroleum oil, fuel oil, or natural gas.
    Homer City means, for purposes of this section, Homer City 
Generation LP's Homer City Generating Station consisting of three coal-
fired units designated as Unit 1 (6,792 MMBtu/hr), Unit 2 (6,792 MMBtu/
hr), and Unit 3 (7,260 MMBtu/hr), located in Center Township, Indiana 
County, Pennsylvania.
    Keystone means, for purposes of this section, Keystone Conemaugh 
Project LLC's Keystone Generating Station consisting of two coal-fired 
units designated as Unit 1 (8,717 MMBtu/hr) and Unit 2 (8,717 MMBtu/
hr), located in Plumcreek Township, Armstrong County, Pennsylvania.
    lb/MMBtu means one pound per million British thermal units.
    Montour means, for purposes of this section, Talen Energy 
Corporation's Montour Steam Electric Station consisting of two coal-
fired units designated as Unit 1 (7,317 MMBtu/hr) and Unit 2 (7,239 
MMBtu/hr), located in Derry Township, Montour County, Pennsylvania.
    ``NOX'' means oxides of nitrogen, measured in accordance with the 
provisions of this section. ``NOX emission rate'' means the number of 
pounds of NOX emitted per million British thermal units of 
heat input (lb/MMBtu), calculated in accordance with this section.
    Operating day means any calendar day on which a Unit fires Fossil 
Fuel.
    Title V Permit means the permit required for major sources pursuant 
to Subchapter V of the Act, 42 U.S.C. 7661-7661e.
    Unit means collectively, the coal pulverizer, stationary equipment 
that feeds coal to the boiler, the boiler that produces steam for the 
steam turbine, the steam turbine, the generator, the equipment 
necessary to operate the generator, steam turbine, and boiler, and all 
ancillary equipment, including pollution control equipment and systems 
necessary for production of electricity. An electric steam generating 
station may be comprised of one or more Units.
    Unit-specific daily NOX mass emissions shall be expressed in lb/day 
and calculated as the sum of total pounds of NOX emitted 
from the Unit during the Unit Operating Day. Each Unit-specific Daily 
NOX Mass Emissions shall include all emissions that occur 
during all periods within any Operating Day, including emissions from 
startup, shutdown, and malfunction.
    (f) NOX emission limitations. (1) The Facility shall achieve and 
maintain their Facility-wide 30-Day Rolling Average NOX 
Emission Rate to not exceed their Facility limit in Table 1 to this 
paragraph (f)(1).

  Table 1 to Paragraph (f)(1)--Facility-Wide 30-Day Rolling Average NOX
                          Emission Rate Limits
------------------------------------------------------------------------
                                                   Facility-wide 30-day
                                                   rolling average NOX
                    Facility                     emission rate limit (lb/
                                                          MMBtu)
------------------------------------------------------------------------
Conemaugh......................................                    0.072
Homer City.....................................                    0.096
Keystone.......................................                    0.075
Montour........................................                    0.102
------------------------------------------------------------------------

    (2) The Facility shall achieve and maintain their Unit-specific 
Daily NOX Mass Emissions to not exceed the Unit-specific 
limit in Table 2 to this paragraph (f)(2).

   Table 2 to Paragraph (f)(2)--Unit-Specific Daily NOX Mass Emissions
                                 Limits
------------------------------------------------------------------------
                                                         Unit-specific
                                                         daily NOX mass
               Facility                     Unit        emissions limit
                                                            (lb/day)
------------------------------------------------------------------------
Conemaugh............................               1             14,308
Conemaugh............................               2             14,308
Homer City...........................               1             15,649
Homer City...........................               2             15,649
Homer City...........................               3             16,727
Keystone.............................               1             15,691
Keystone.............................               2             15,691
Montour..............................               1             17,912
Montour..............................               2             17,721
------------------------------------------------------------------------

    (g) Monitoring of NOX emissions. (1) In determining the Facility-
wide 30-Day Rolling Average NOX Emission Rate, the Facility 
shall use CEMS in accordance with the procedures of 40 CFR parts 60 and 
75, appendix F, Procedure 1.
    (2) For purposes of calculating the Unit-specific Daily 
NOX Mass Emissions Limits, the Facility shall use CEMS in 
accordance with the procedures at 40 CFR part 75. Emissions rates, mass 
emissions, and other quantitative standards set by or under this 
section must be met to the number of significant digits in which the 
standard or limit is expressed. For example, an Emission Rate of 0.100 
is not met if the actual Emission Rate is 0.101. The Facility shall 
round the fourth significant digit to the nearest third significant 
digit, or

[[Page 53404]]

the sixth significant digit to the nearest fifth significant digit, 
depending upon whether the limit is expressed to three or five 
significant digits. For example, if an actual emission rate is 0.1004, 
that shall be reported as 0.100, and shall be in compliance with an 
emission rate of 0.100, and if an actual emission rate is 0.1005, that 
shall be reported as 0.101, and shall not be in compliance with an 
emission eate of 0.100. The Facility shall report data to the number of 
significant digits in which the standard or limit is expressed.
    (h) Recordkeeping and periodic peporting. (1) The Facility shall 
electronically submit to EPA a periodic report, within thirty (30) Days 
after the end of each six-month reporting period (January through June, 
July through December in each calendar year). The portion of the 
periodic report containing the data required to be reported by this 
paragraph (h) shall be in an unlocked electronic spreadsheet format, 
such as Excel or other widely-used software, and contain data for each 
Operating Day during the reporting period, including, but not limited 
to: Facility ID (ORISPL); Facility name; Unit ID; Date; Unit-specific 
total Daily Operating Time (hours); Unit-specific Daily NOX 
Mass Emissions (lbs); Unit-specific total Daily Heat Input (MMBtu); 
Unit-specific Daily NOX Emission Rate (lb/MMBtu); Facility-
wide 30-Day Rolling Average NOX Emission Rate (lb/MMBtu); 
Owner; Operator; Representative (Primary); and Representative 
(Secondary). In addition, the Facility shall maintain the following 
information for 5 years from the date of creation of the data and make 
such information available to EPA if requested: Unit-specific hourly 
heat input, Unit-specific hourly ammonia injection amounts, and Unit-
specific hourly NOX emission rate.
    (2) In any periodic report submitted pursuant to this section, the 
Facility may incorporate by reference information previously submitted 
to EPA under its Title V permitting requirements, so long as that 
information is adequate to determine compliance with the emission 
limits and in the same electronic format as required for the periodic 
report, and provided that the Facility attaches the Title V Permit 
report (or the pertinent portions of such report) and provides a 
specific reference to the provisions of the Title V Permit report that 
are responsive to the information required in the periodic report.
    (3) In addition to the reports required pursuant to this section, 
if the Facility exceeds the Facility-wide 30-day rolling average 
NOX emission limit on three or more days during any 30-day 
period, or exceeds the Unit-specific daily mass emission limit for any 
Unit on three or more days during any 30-day period, the Facility shall 
electronically submit to EPA a report on the exceedances within ten 
(10) business days after the Facility knew or should have known of the 
event. In the report, the Facility shall explain the cause or causes of 
the exceedances and any measures taken or to be taken to cure the 
reported exceedances or to prevent such exceedances in the future. If, 
at any time, the provisions of this section are included in Title V 
Permits, consistent with the requirements for such inclusion in this 
section, then the deviation reports required under applicable Title V 
regulations shall be deemed to satisfy all the requirements of this 
paragraph (h)(3).
    (4) Each report shall be signed by the Responsible Official as 
defined in Title V of the Clean Air Act, or his or her equivalent or 
designee of at least the rank of Vice President. The signatory shall 
also electronically submit the following certification, which may be 
contained in a separate document:

    ``This information was prepared either by me or under my 
direction or supervision in accordance with a system designed to 
assure that qualified personnel properly gather and evaluate the 
information submitted. Based on my evaluation, or the direction and 
my inquiry of the person(s) who manage the system, or the person(s) 
directly responsible for gathering the information, I hereby certify 
under penalty of law that, to the best of my knowledge and belief, 
this information is true, accurate, and complete. I understand that 
there are significant penalties for submitting false, inaccurate, or 
incomplete information to the United States.''

    (5) Whenever notifications, submissions, or communications are 
required by this section, they shall be made electronically to the 
attention of the Air Enforcement Manager via email to the following 
address: [email protected].

[FR Doc. 2022-18669 Filed 8-30-22; 8:45 am]
BILLING CODE 6560-50-P