Air Plan Approval; Florida; Second Planning Period Regional Haze Plan

Federal RegisterJun 5, 2025

Ask Donna

What actually matters in this document.

Text

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Part 52

[EPA-R04-OAR-2021-0930; FRL-10403-02-R4]

Air Plan Approval; Florida; Second Planning Period Regional Haze Plan

AGENCY:

Environmental Protection Agency (EPA).

ACTION:

Final rule.

SUMMARY:

The Environmental Protection Agency (EPA) is approving a regional haze State Implementation Plan (SIP) revision submitted by the Florida Department of Environmental Protection (FDEP) on October 8, 2021, and supplemented on June 14, 2024, and October 28, 2024, as satisfying applicable requirements under the Clean Air Act (“CAA” or “Act”) and EPA's Regional Haze Rule (RHR) for the program's second planning period. Florida's SIP submissions were submitted to address the requirement that states must periodically revise their long-term strategies for making reasonable progress toward the national goal of preventing any future, and remedying any existing, anthropogenic impairment of visibility, including regional haze, in mandatory Class I Federal areas (hereinafter referred to as “Class I areas”). These SIP submissions also address other applicable requirements for the second planning period of the regional haze program. EPA is taking this action pursuant to sections 110 and 169A of the Act.

DATES:

This rule is effective July 7, 2025.

ADDRESSES:

EPA has established a docket for this action under Docket Identification No. EPA-R04-OAR-2021-0930. All documents in the docket are listed on the

regulations.gov

website. Although listed in the index, some information may not be publicly available,

i.e.,

Confidential Business Information 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 either electronically through

www.regulations.gov

or in hard copy at the Air Regulatory Management Section, Air Planning and Implementation Branch, Air and Radiation Division, U.S. Environmental Protection Agency, Region 4, 61 Forsyth Street SW, Atlanta, Georgia 30303-8960. EPA requests that, if at all possible, you contact the person listed in the

FOR FURTHER INFORMATION CONTACT

section to schedule your inspection. The Regional Office's official hours of business are Monday through Friday, 8:30 a.m. to 4:30 p.m., excluding Federal holidays.

FOR FURTHER INFORMATION CONTACT:

Pearlene Williams-Miles, Multi-Air Pollutant Coordination Section, Air Planning and Implementation Branch, Air and Radiation Division, U.S. Environmental Protection Agency, Region 4, 61 Forsyth Street SW, Atlanta, Georgia 30303-8960. Ms. Williams-Miles can be reached via telephone at (404) 562-9144 or electronic mail at

williamsmiles.pearlene@epa.gov.

SUPPLEMENTARY INFORMATION:

I. Background

On October 8, 2021, June 14, 2024, and October 28, 2024, FDEP submitted revisions to its SIP to address regional haze for the second planning period (“Haze Plan”).

1 2 3

FDEP made these SIP submissions to satisfy the requirements of the CAA's regional haze program pursuant to CAA sections 169A and 169B and 40 Code of Federal Regulations (CFR) 51.308. EPA has determined that the Haze Plan meets the applicable statutory and regulatory requirements and is thus approving Florida's submissions into its SIP.

“Haze Plan” collectively refers to the October 8, 2021, June 14, 2024, and October 28, 2024, SIP submissions. The phrase “2021 Plan” refers to the October 8, 2021, SIP submission; “2024 Supplement” refers to the June 14, 2024, SIP submission, which supplements the 2021 Plan; and “Second 2024 Supplement” refers to the October 28, 2024, SIP submission, which also supplements the 2021 Plan.

The 2021 Plan requests removal of source-specific and best available retrofit technology (BART) limits and conditions from the Florida SIP that address source-specific reasonable progress and BART control measures during the first planning period. On June 14, 2024, FDEP withdrew this request from its SIP revision, and thus, there is no action for EPA to take on this request.

The October 28, 2024, submission contains permits and a Four Factor Analysis (FFA) for the Georgia-Pacific—Foley Cellulose Perry Mill (Foley). In a letter dated April 8, 2025, FDEP withdrew its request to incorporate permit conditions for Foley from its SIP revision,and thus, there is no action for EPA to take on this request. This letter is included in the docket for this rulemaking.

Through a notice of proposed rulemaking (NPRM) published on December 27, 2024 (89 FR 105506), EPA proposed to approve Florida's Haze Plan as satisfying the regional haze requirements for the second planning period contained in the CAA and 40 CFR 51.308. EPA described its rationale for proposing approval of the Haze Plan in the December 27, 2024, NPRM. Comments on the December 27, 2024, NPRM were due on or before January 27, 2025. EPA received two sets of comments on the NPRM, one of which was a request for an extension to the public comment period. These comments are available in the docket for this action.

II. Response to Comments

In response to the NPRM, EPA received a comment letter dated January 27, 2025, and signed by the National Parks Conservation Association (NPCA), Sierra Club, the Coalition to Protect America's National Parks, and Friends of the Everglades (collectively referred to as the “Conservation Groups”). Additionally, EPA received a request for an extension to the public comment period dated January 9, 2025, signed by Laumann Legal, LLC, NPCA, Sierra Club, and the Coalition to Protect America's National Parks. All comments received are available in the docket for this action. A summary of the significant comments received from the Conservation Groups and EPA's responses to these comments is below.

Comment 1:

The Conservation Groups contend that EPA's proposal to approve Florida's reliance on the Visibility Improvement State and Tribal Association of the Southeast's (VISTAS) visibility modeling is arbitrary and capricious because the Agency ignored significant flaws in this modeling. They state that they informed VISTAS and EPA of significant errors in the visibility modeling through a 2021 letter and that EPA did not acknowledge these errors in the NPRM. They contend these errors affected the source selection process for all of the VISTAS states. Consequently, they assert that Florida improperly excluded major sources of haze-forming pollution from FFAs. These alleged errors are addressed in Comments 1.a through 1.c below.

Comment 1.a:

The Conservation Groups contend that the VISTAS modeling significantly underpredicted the contribution of sulfates to visibility impairment at Class I areas on the 20 percent most impaired days and that this underprediction was largest during the summer months when sulfate extinction is known to be a major contributor to visibility impairment, and when visibility impairment is most problematic.

They also assert that these errors resulted in the modeling not meeting VISTAS' model performance

goals and modeling acceptance criteria for a number of Class I areas. They provide examples of specific Class I areas in Florida where they contend the visibility modeling “failed to meet the acceptance criteria for sulfate” at Chassahowitzka National Wilderness Area (Chassahowitzka) by −30.37 percent, and at St. Marks National Wilderness Area (St. Marks) by −40.16 percent. They further assert that, although Florida claims that it corrected for these underpredictions through the use of relative response factors (RRFs) for its 2028 future year projections, neither Florida nor EPA assessed whether use of RRFs adequately corrected for errors in the modeling. They state that according to EPA's 2018 modeling guidance, the effectiveness of RRFs is dependent on the type of data used to calculate them.

Areas statutorily designated as mandatory Federal Class I areas consist of national parks exceeding 6,000 acres, wilderness areas and national memorial parks exceeding 5,000 acres, and all international parks that were in existence on August 7, 1977. CAA section 162(a). There are 156 mandatory Class I areas. The list of areas to which the requirements of the visibility protection program apply is in 40 CFR part 81, subpart D.

EPA's

Modeling Guidance for Demonstrating Air Quality Goals for Ozone, PM

2.5

and Regional Haze

(November 29, 2018) (“2018 Modeling Guidance”) is in the docket for this rulemaking and is also available at:

https://www.epa.gov/sites/default/files/2020-10/documents/o3-pm-rh-modeling_guidance-2018.pdf.

Response 1.a:

EPA disagrees that there are significant flaws in Florida's 2028 visibility modeling that resulted in excluding major sources of haze-forming pollution from evaluation via FFAs for the second planning period. As the Conservation Groups state, Florida relied upon the photochemical visibility modeling performed by VISTAS to project the impact of the State's 2028 sulfur dioxide (SO

) and nitrogen oxide (NO

) emissions on visibility in both in-state and out-of-state Class I areas. VISTAS performed the modeling in accordance with the principles described within EPA's “Modeling Guidance for Demonstrating Air Quality Goals for Ozone, PM

2.5

and Regional Haze” (2018 Modeling Guidance).

In 2018, EPA approved the Quality Assurance Project Plan

prepared by VISTAS for performing the modeling and reviewed and provided comments on the VISTAS Modeling Protocol. EPA also reviewed the VISTAS final modeling reports and data relied upon by Florida and found them acceptable.

Id.

The April 3, 2018, Quality Assurance Project Plan for the VISTAS II Regional Haze Project is located in appendix A-1 of the 2021 Plan.

Regarding sulfate predictions, figure 6-7 of Florida's Haze Plan shows the results of the normalized mean bias and normalized mean error statistical model performance tests for sulfates across the VISTAS region. Figure 6-7 does show that the modeled sulfate levels are biased low, with some values falling outside of the model performance criteria. However, as discussed below, these biases are not uncommon in photochemical modeling analyses and can be addressed with additional analyses.

Model bias and error, either high or low, is not uncommon in photochemical modeling analyses due to uncertainties in model inputs and the scientific model formulation, and the fact that all air quality models are simplified approximations of the complex phenomena of atmospheric chemistry, fate, and transport of pollutants. Section 6.0 of EPA's 2018 Modeling Guidance discusses uncertainties that may affect model results and provides recommendations to mitigate modeling bias and uncertainty. Florida acknowledges that model performance is biased low on the 20 percent most impaired days and provided an explanation of why this modeling was appropriate for its regulatory determinations in the 2021 Plan (which references the 2018 Modeling Guidance in several instances). The 2018 Modeling Guidance states that it is not appropriate to use a “bright-line test” for distinguishing between adequate and inadequate photochemical model performance for a single performance test statistic.

EPA's 2018 Modeling Guidance instead recommends using a “weight of evidence” approach for evaluating model performance holistically.

See

2018 Modeling Guidance at 69 (“Further, even with a single performance test, it is not appropriate to assign “bright line” criteria that distinguish between adequate and inadequate model performance.”).

Id.

(“[T]he EPA recommends that a “weight of evidence” approach be used to determine whether a particular modeling application is valid for assessing the future attainment status of an area.”).

As discussed in section 5.2(d) of EPA's “Guideline on Air Quality Models” contained in 40 CFR part 51, appendix W, there are no specific levels of any model performance metric that indicate acceptable model performance. The decision regarding acceptability is heavily influenced by professional judgment of the reviewing authority, which is EPA in this case. Based upon the overall performance of the model for all pollutants affecting visibility, considered holistically, Florida's conclusions that the modeling is acceptable for use in the regional haze SIP analyses are reasonable, and Florida provided a reasonable explanation for the model bias.

Just as importantly, Florida took appropriate steps to correct for this model bias. The Haze Plan explains that the model is applied in a relative sense through the calculation of RRFs following the procedures in 2018 Modeling Guidance for calculating 2028 future year visibility impacts, which mitigates concerns about the low bias in the sulfate model predictions. As described in EPA's 2018 Modeling Guidance, RRFs are “the fractional change in air quality concentrations that is simulated due to emissions changes between a base and a future year emissions scenario.”

Id.

at 103.

EPA agrees with Florida that applying the model in a relative sense using the RRFs is an important tool in mitigating the impacts of the sulfate modeling underpredictions in the 2011 baseline year on the model projections for the 2028 future year. Section 4.1 of the 2018 Modeling Guidance provides a detailed explanation of why EPA recommends photochemical modeling be applied in a relative sense and explains that problems posed by model bias are expected to be reduced when using the relative approach. Section 6.5 of Florida's 2021 Plan explains the calculation of 2028 visibility estimates using the RRF approaches contained in EPA's 2018 Modeling Guidance. Using the RRF approach with an average of five years of Interagency Monitoring of Protected Visual Environments (IMPROVE)

data on the 20 percent most impaired days and 20 percent clearest days along with the relative percent modeled change in all the particulate matter (PM) species between 2011 and 2028 reduces the influence of the bias in sulfate-modeled (and other PM species) values in the 2011 baseline year. The 2028 visibility impairment projection is derived primarily from the five-year average of actual IMPROVE monitoring data in 2009-2013 that was then scaled in a relative sense by the modeling results. If the model were being applied in an absolute sense, the low bias in the sulfate modeled values would have a larger impact on the 2028 visibility projections. For these reasons, Florida's use of the VISTAS model results to inform source selection was reasonable due to the use of RRFs to minimize the impacts of model bias.

IMPROVE visibility monitoring data is available at:

https://vista.cira.colostate.edu/Improve/.

Comment 1.b:

The Conservation Groups state that VISTAS relied on an “outdated” 2011 baseline year for its 2028 future year emissions projections and assumed that electric generating units (EGUs) would operate in the exact same manner in 2028 as they did in 2011. Thus, they assert that the model assumptions and results are incorrect because EGUs are likely to have different load utilization in 2028 than in 2011.

Response 1.b:

Florida's use of a 2011 base emissions inventory year to project emissions out to 2028 (the end of the second planning period) is reasonable in this instance. Although it is always preferable to use the most recent information available for modeling, the 2011 baseline year inventory used by VISTAS was the latest region-wide inventory available at the time that Florida's SIP submittal was being developed during the VISTAS technical work, which took place from December 2017 to February 26, 2021.

In EPA's experience, coordination among states such as those in the VISTAS region takes time, and the modeling involved is time-consuming, highly technical, and resource intensive. The modeling generally requires hundreds of hours of time to gather the model input data (

e.g.,

emissions, meteorology, land-use, etc.), prepare modeling protocols, perform the modeling, and analyze the results. The computational resources to run photochemical models are also very large. “Mainframe” clusters of a large number of computer processors are required to run the models, and even using these powerful computers, it takes weeks of computer run-time for a full-year model simulation. Additionally, EPA's newer 2016-based modeling platform only became available in September 2019,

after VISTAS had already invested a considerable amount of time and money into the regional haze modeling analysis, including the Comprehensive Air Quality Model with Extensions (CAMx) Particulate Matter Source Apportionment Technology (PSAT) source apportionment modeling that was used to identify sources to evaluate or reasonable progress.

See

“Timeline” for the VISTAS II Regional Haze Project at:

https://www.metro4-sesarm.org/content/vistas-regional-haze-project-intro.

See

“Technical Support Document for EPA's Updated 2028 Regional Haze Modeling” at:

https://www.epa.gov/visibility/technical-support-document-epas-updated-2028-regional-haze-modeling.

As discussed in section 6.6 of Florida's 2021 Plan, Florida evaluated the results of EPA's 2016 modeling for Everglades due to issues with model performance in the VISTAS 2028 modeling results for Everglades.

See

the September 29, 2018, memo from EPA (Richard Wayland) regarding Availability of Modeling Data and Associated Technical Support Document for the EPA's Updated 2028 Visibility Air Quality Modeling in the docket for this rulemaking. Due to these uncertainties, Florida instead relied on the results of EPA's 2016 modeling. EPA's 2016 modeling did not include PSAT tagging of individual sources like the VISTAS modeling, so for selecting sources to evaluate for control analyses, Florida used the VISTAS PSAT modeling results at Everglades like it did for Chassahowitzka and St. Marks. As discussed in section 7.4 of the 2021 Plan, both the VISTAS Modeling and EPA's 2016 Modeling show that EGU and non-EGU point sources contribute approximately five percent to total light extinction at Everglades, indicating that model performance for evaluating the impacts of EGU and non-EGU point sources is similar. Therefore, Florida's use of the PSAT modeling for source selection is acceptable to EPA.

EPA develops the National Emissions Inventory (NEI) suitable for use in such models every three years.

By design, the regional haze program requires states to spend significant time in the planning phase, and this generally necessitates the use of a baseline year that is substantially earlier than the date the state submits its SIP to EPA. There is no RHR requirement regarding the baseline year for regional photochemical modeling (nor is photochemical modeling required). Florida justifies the use of this particular baseline year and states that the 2011 emissions inventory was the most recently available quality-assured statewide emissions inventory when the VISTAS project began for the second planning period.

Moreover, prior to using this data, Florida discussed the selection of this baseline year emissions inventory and received confirmation from EPA to use this emissions inventory.

Given the aforementioned reasons, EPA finds the use of the 2011 baseline year by VISTAS, and thus Florida, reasonable.

For more information on the NEI,

see https://www.epa.gov/air-emissions-inventories/national-emissions-inventory-nei.

See

2021 Plan at 56 (“The year 2011 was selected as the modeling base year because the VISTAS 2028 emissions inventory is based on the 2011 Version 6 EPA modeling platform. For the analyses in this SIP, this period consists of those years surrounding 2011 (

i.e.,

2009-2013)”).

See also

2021 Plan at 83 (“Calendar year 2011 satisfies the criteria in EPA's modeling guidance episode selection discussion and is consistent with the base year modeling platform. Specifically, EPA's guidance recommends choosing a time period which reflects the variety of meteorological conditions that represent visibility impairment on the 20 percent clearest and 20 percent most-impaired days in the Class I areas being modeled (high and low concentrations). This is best accomplished by modeling a full calendar year. In addition, the 2011/2028 modeling platform was the most recent available platform when VISTAS started their modeling work. EPA's 2016-based platform became available at a later date after VISTAS had already invested a considerable amount of time and money into the modeling analysis. Using the 2016-based platform was not feasible from a monetary perspective, nor could such work be done in a timely manner.”).

See

the January 29, 2018, email from EPA (Richard Wayland) regarding use of a 2011 base year by VISTAS for regional haze in the docket for this rulemaking.

The 2011 emissions inventory was used to estimate emissions of visibility impairing pollutants in 2028. VISTAS applied reductions expected from Federal and state regulations to the visibility impairing pollutants NO

, PM, and SO

. Florida's 2028 emissions projections are based on the State's technical analysis of the anticipated emission rates and level of activity for EGUs, other point sources, non-point sources, on-road sources, and off-road sources based on their emissions in the 2011 base year, considering growth and additional emissions controls to be in place by 2028. In addition, the VISTAS emissions inventory for 2028 accounts for post-2011 emission reductions from promulgated Federal, state, local, and site-specific control programs.

Although Florida used 2011 as its emissions inventory base year, as required by the RHR at 40 CFR 51.308(f)(2)(iii), Florida also examined more recent emissions inventory information for SO

and NO

for the years 2017, 2018, and 2019 and compared these emissions to the 2028 emission projections that were used for modeling purposes in section 7.6.5, table 7-28 of its Haze Plan. This helped to ensure that the State adequately considered more recent emissions inventory information when developing its long-term strategy (LTS). The technical information provided in the docket demonstrates that the emissions inventory in the Haze Plan adequately reflects projected 2028 conditions. Given the aforementioned reasons, EPA finds the use of the 2011 baseline year by VISTAS (and thus Florida) reasonable.

Comment 1.c:

The Conservation Groups state that VISTAS used “outdated” monitoring data for its 2028 future year projections that did not reflect an observed shift in nitrate contribution to visibility impairment in the southeastern United States in the recent past. They therefore contend that this resulted in the exclusion of major NO

sources from the modeling results.

Response 1.c:

Regarding the Conservation Groups' comment that the 2009-2013 modeling base period did not reflect more recent changes in nitrate contributions, EPA discussed its views on this issue in detail in the NPRM. Nitrates are also discussed in Response 4, below. EPA agrees that after the 2009-2013 timeframe, nitrate impacts have become more significant on some of the 20 percent most impaired days, especially considering the significant decrease in SO

emissions and measured sulfate

concentrations as acknowledged in the NPRM. EPA nonetheless agrees with Florida's conclusion that for the second planning period, sulfates remain the dominant visibility-impairing pollutant at the Class I areas affected by Florida and that it is therefore reasonable for Florida to focus on SO

-emitting sources during this period.

Comment 2:

The Conservation Groups state that the purported errors in the VISTAS modeling discussed in Comment 1 were carried forward into the source selection process for VISTAS states, including Florida, and that those errors caused VISTAS, and the states that relied on the VISTAS process, to improperly exclude sources from FFAs. In addition to the modeling errors, they state that Florida adopted VISTAS' “unreasonable” source screening process that uses Area of Influence (AoI) and PSAT analyses, and also applied unreasonably high source selection thresholds. Based on these reasons, they conclude that EPA's proposal to approve the State's source selection method is arbitrary and capricious. The Conservation Groups' specific comments on this topic are addressed in Comments 2.a through 2.f, below.

Comment 2.a:

The Conservation Groups comment that Florida employed unreasonably high source selection thresholds for the AoI analysis, which were too restrictive and resulted in the identification of only 13 Florida sources at the AoI step. Specifically, they assert that by using a percentage source selection threshold, the State's calculated threshold in absolute terms was higher for Class I areas with the most severe visibility impairment, meaning that fewer sources were identified at the AoI step for Class I areas with the worst impairment. The Conservation Groups state that for the areas with the worst visibility impairment, more sources should be selected to make progress toward the natural visibility goal. In addition, the Conservation Groups state that neither Florida nor EPA have provided justification to support the application of a five percent threshold for in-state and out-of-state sources at the AoI step.

Response 2.a:

EPA disagrees with this comment. The RHR does not require states to consider controls for all sources, all source categories, or any or all sources in a particular source category. Nor does the RHR expressly specify criteria for minimum source selection thresholds.

These flexibilities are, however, not unbounded. The RHR requires that “[t]he State should consider evaluating major and minor stationary sources or groups of sources, mobile sources, and area sources. The State must include in its implementation plan a description of the criteria it used to determine which sources or groups of sources it evaluated and how the four factors were taken into consideration in selecting the measures for inclusion in its long-term strategy.”

In addition, the technical basis for source selection must also be documented, as required by 40 CFR 51.308(f)(2)(iii). Thus, states must utilize a reasonable source selection methodology, and whatever choices states make regarding source selection should be reasonably explained.

Florida met these requirements. Specifically, Florida discussed the criteria it used to determine which sources or groups of sources were evaluated by the State, including the use of AoI analysis, photochemical modeling (

e.g.,

PSAT), and associated source selection thresholds for AoI and PSAT tagging in its Haze Plan. Florida documented its use of these approaches in extensive detail within section 7.5 of the Haze Plan and appendix D-1 of the Haze Plan (relating to AoI analysis) and section 7.6 and appendices E-1a, E-1b, E-2a, E-2b, E-2c, E-2d, E-2e, E-2f, E-3, E-4, E-5, E-6, E-7a, and E-8 of the Haze Plan (relating to PSAT analysis).

See

40 CFR 51.308(f)(2)(i).

See

89 FR 47481, 47493 (June 3, 2024).

Florida's documentation adequately demonstrates why its source selection methodology—including the use of an AoI threshold of five percent of sulfate and nitrate for in-state and out-of-state sources for follow-up PSAT tagging and a one percent PSAT threshold on a pollutant-by-pollutant basis for source selection—is reasonable. For the reasons stated in the NPRM, EPA finds that Florida's source selection methodology was reasonable and resulted in the selection of a reasonable set of sources contributing to visibility impairment at Class I areas affected by Florida's sources. The State's methods for selecting sources for a control analysis and the State's AoI and PSAT analyses identified sources in Florida having the highest impact on visibility at Class I areas at the end of the second planning period and identified sources outside of Florida having the largest impacts on visibility at Class I areas in the State. A specific source selection approach is not required by the RHR.

Both of these approaches (AoI and PSAT) are example methods in EPA's August 20, 2019, guidance titled: “Guidance on Regional Haze State Implementation Plans for the Second Implementation Period” (“2019 Guidance”), which is available at:

https://www.epa.gov/sites/default/files/2019-08/documents/8-20-2019_-_regional_haze_guidance_final_guidance.pdf. See

subsection “b) Estimating baseline visibility impacts for source selection” on pages 12-15 of the 2019 Guidance. PSAT is a type of photochemical modeling which is item 4 on page 13 of the 2019 Guidance. VISTAS' AoI analyses involve items 1-3 on page 13 of the 2019 Guidance.

The results of this methodology were reasonable as well. Florida selected for further analysis the 12 sources with the largest visibility impacts (accounting for both SO

/sulfate and NO

/nitrate

) at Florida and nearby Class I areas. On the whole, SO

emissions from the 12 in-state sources selected by Florida for further analysis of reasonable progress—Duke Crystal River Power Plant (Duke-Crystal River), Georgia-Pacific, Foley Cellulose, LLC (Foley), Jacksonville Electric Authority—JEA Northside Generating Station (JEA Northside), Lakeland CD McIntosh Jr. Power Plant (CD McIntosh),

Mosaic Fertilizer, LLC—Bartow (Mosaic-Bartow), Mosaic Fertilizer, LLC—New Wales (Mosaic-New Wales), Mosaic Fertilizer, LLC—South Pierce (Mosaic-South Pierce), Nutrien White Springs Agricultural Chemicals, Inc. (Nutrien), Seminole Generating Station (Seminole), Tampa Electric Company—Big Bend Power Station (TECO-Big Bend), WestRock Fernandina Beach Paper Mill (WestRock-Fernandina), and WestRock Panama City Paper Mill (WestRock-Panama City)

—are projected to impact visibility at Class I areas as described in table 1 below.

Florida selected sources for PSAT modeling based on the combined impact of sulfate plus nitrate. Sulfates and nitrates were modeled together in the PSAT modeling with the other PM species that impact visibility (

e.g.,

direct PM, organic carbon, elemental carbon, etc.). There were no sources with a sulfate impact below the PSAT threshold(s), but a sulfate plus nitrate impact above the threshold(s).

The fossil fuel steam generating unit No. 3 (EU006) at CD McIntosh was permanently shut down in 2021.

See

appendices G-3d and G-5h of the 2021 Plan.

In June 2022, the WestRock-Panama City facility announced its intention to permanently cease operations.

See

section 7.8.4 of the 2024 supplement. FDEP included documentation for closure of the WestRock-Panama City facility in its 2024 Supplement. In addition, on October 18, 2024, FDEP sent a site inspection report and other supporting documentation for the WestRock-Panama City closure as an addendum to the 2024 Supplement. The inspection report documents the permanent closure and inoperable status of the facility and notes that any project to restore the facility would be subjected to mandatory New Source Review (NSR) and that multiple new source performance standards would inevitably apply. This additional documentation may be found in the docket for this rulemaking.

Table 1—Sulfate PSAT Contributions (Percent) for the 12 Sources Selected for Further Analysis in Five Class I Areas on the 20 Percent Most Impaired Days *

Sources ** sulfate PSAT

contributions to Class I areas

Chassahowitzka National

Wilderness Area

(FL)

St. Marks National Wilderness Area

(FL)

Everglades

National Park

(FL)

Okefenokee

National

Wilderness Area

(GA)

Wolf Island

National

Wilderness Area ***

(GA)

Duke-Crystal River

6.45

Foley

2.23

JEA Northside

1.34

CD McIntosh

Mosaic-Bartow

2.68

Mosaic-New Wales

2.66

Mosaic-South Pierce

Nutrien

2.87

Seminole

TECO-Big Bend

1.32

3.38

WestRock-Fernandina

1.36

2.43

WestRock-Panama City

4.74

* Note that fields with a “-“indicate that visibility impacts are below one percent.

** The Class I areas listed in table 1, above, are included because the Florida facilities in this table have a sulfate PSAT contribution of one percent or more at one or more of these areas.

*** Wolf Island National Wilderness Area (Wolf Island) has no IMPROVE monitor. Visibility at Wolf Island is assumed to be the same as the nearest Class I area monitor located at Okefenokee National Wilderness Area (Okefenokee).

Although these 12 sources are the largest contributors within Florida to visibility impairment at Class I areas, table 1 shows sulfate PSAT visibility impacts from these sources range from approximately 1.3 to 6.5 percent at the selected Class I areas. This is due to the fact that most anthropogenic impacts to visibility at these Class I areas come from outside of Florida. In fact, these anthropogenic impacts primarily originate from outside the VISTAS states. This is illustrated in figures 7-22, 7-23, and 7-24 of the 2021 Haze Plan, which provide the contributions from 2028 SO

and NO

emissions to visibility impairment from all source sectors for the 20 percent most impaired days in units of inverse megameters (Mm

−

). The entries in table 2, below, show the contributions made from Florida, all other VISTAS states, and other Regional Planning Organizations to Florida's Class I areas.

Table 2—Contributions of 2028 SO

and NO

Emissions From All Source Sectors to Visibility Impairment for the 20 Percent Most Impaired Days for Class I Areas in Florida

[Mm

−

] *

Class I area **

Projected 2028

impairment

on 20% most

impaired days

All other

VISTAS

states

CENRAP region ***

LADCO

region ***

MANE-VU

region ***

WRAP region

within

VISTAS

modeling

domain ***

CHAS

53.92

4.13

4.09

3.21

1.76

0.22

2.22

SAMA

52.91

2.86

4.60

5.26

2.21

0.39

3.44

EVER

47.70

1.49

0.22

0.68

0.17

0.03

2.05

OKEF

54.67

2.76

6.99

2.27

3.60

1.02

2.84

WOLF

53.59

1.69

7.44

2.15

3.44

1.15

3.41

* The columns to the right of “Projected 2028 Impairment on 20% Most Impaired Days” do not add up to the values in the “Projected 2028 Impairment on 20% Most Impaired Days” column due to international emissions and boundary emissions visibility impacts not shown in this table.

** “CHAS” refers to Chassahowitzka National Wilderness Area; “SAMA” refers to St. Marks National Wilderness Area; “EVER” refers to Everglades National Park (Everglades); “OKEF” refers to Okefenokee National Wilderness Area; and “WOLF” refers to Wolf Island National Wilderness Area.

*** “CENRAP” refers to Central Regional Air Planning Association (which is associated with the Central States Air Resource Agencies (CENSARA)); “LADCO” refers to Lake Michigan Air Directors Consortium; “MANE-VU” refers to Mid-Atlantic/Northeast Visibility Union; “WRAP” refers to Western Regional Air Partnership.

See also https://www.epa.gov/visibility/visibility-regional-planning-organizations.

Table 2, above, illustrates that Florida's in-state emissions account for a relatively small fraction of total visibility impairment at Class I areas impacted by Florida sources. This fraction is approximately 7.66 percent

for Chassahowitzka, 5.41 percent for St. Marks, 3.12 percent for Everglades, 5.05 percent for Okefenokee, and 3.15 percent for Wolf Island.

These percentages were calculated by dividing the “FL” column by the “Projected 2028 20% Most Impaired Days Column” and multiplying by 100.

Likewise, the PSAT Tag Results spreadsheet referenced in section 6.3 of appendix E-7a of the 2021 Plan shows the visibility impacts on a facility-by-facility basis due to SO

emissions. Specifically, the spreadsheet referenced in Attachment A of appendix E-7a shows the following SO

visibility impacts to Class I areas impacted by Florida sources on the 20 percent most impaired days in units of Mm

−

Table 3—2028 SO

Visibility Impacts to Florida Class I Areas on the 20 Percent Most Impaired Days

[Mm

−

]

Source

Chassahowitzka

St. Marks

Everglades

Okefenokee

Wolf Island

Duke-Crystal River

0.629

0.047

0.006

0.028

0.025

Foley

0.066

0.112

0.001

0.289

0.064

JEA Northside

0.095

0.012

0.006

0.113

0.167

CD McIntosh *

Mosaic-Bartow

0.080

0.064

0.035

0.018

0.012

Mosaic-New Wales

0.073

0.069

0.035

0.018

0.011

Mosaic-South Pierce *

Nutrien

0.050

0.015

0.002

0.372

0.087

Seminole *

TECO-Big Bend

0.129

0.098

0.044

0.029

0.016

WestRock-Panama City

0.058

0.540

0.002

0.060

0.041

WestRock-Fernandina

0.090

0.014

0.008

0.176

0.304

Total of Florida Selected Sources

1.27

0.971

0.139

1.103

0.727

Florida Total Contribution

3.35

2.40

0.89

2.30

1.42

All Sources (including out-of-state Contribution)

12.54

15.84

2.61

16.39

16.22

* CD McIntosh, Mosaic-South Pierce, and Seminole were not tagged for PSAT modeling, so no PSAT contribution information is available.

The above data in table 3 further supports that Florida's source selection thresholds and source selection methodology were reasonable. Specifically, on the 20 percent most impaired days, Florida's selected in-state sources are responsible for approximately 37.9 percent of Florida's total in-state SO

visibility impairment at Chassahowitzka, 40.5 percent of total in-state SO

visibility impairment at St. Marks, 34.1 percent of total in-state SO

visibility impairment at Everglades, 48.0 percent of total in-state SO

visibility impairment at Okefenokee, and 51.2 percent of total in-state SO

visibility impairment at Wolf Island.

Additionally, section 7.6.4 of the 2021 Plan states that the selected sources represent 12 of the top 18 SO

emitting sources in Florida in 2019,

which account for approximately 35,000 tons of SO

emissions, which is the vast majority of all the point source emissions in Florida. States are not required by the RHR to select every source in the state, and Florida selected the in-state sources with the largest visibility impacts on in-state and nearby Class I areas. The selection of the above sources captured sufficient visibility-impairing emissions to allow Florida to ensure that FFAs conducted for this planning period had the potential to meaningfully reduce emissions (and thus, associated visibility impacts at Class I areas) from in-state sources.

These percentages were calculated by dividing the “Total of Selected Florida Sources” row in table 3 by the “Florida Total Contribution” row and multiplying by 100.

Florida's 2021 Plan states that 11 of the top 18 sources were selected. Florida later added the Mosaic-South Pierce facility to the list of selected sources for a total of 12 sources, not 11, as discussed in the 2024 Supplement on pages 4-6. FDEP determined that increases in SO

emissions from the Mosaic-South Pierce facility since the 2011 baseline period warranted a reasonable progress analysis.

Table 3 also shows that most emissions of visibility-impairing sulfates that impact Florida's Class I areas on the 20 percent most impaired days are emitted from outside of Florida. The same general pattern holds for the 20 percent least impaired days as well. Florida does not have jurisdiction through its SIP to regulate sources outside of state boundaries. Florida did, however, request FFAs from other states for an additional two facilities outside of Florida through the interstate consultation process.

The “regional” nature of the regional haze program necessarily requires Florida to rely on reasonable progress made by other states, just as other states must rely on Florida to make reasonable progress.

See

2021 Plan at section 7.6.

The Conservation Groups also argue that neither Florida nor EPA provided justification for the five percent AoI threshold for out-of-state sources. In its 2021 Plan, Florida explained that use of an AoI contribution of five percent or more to tag sources for PSAT captures large sources outside of Florida. When selecting out-of-state sources, 40 CFR 51.308(f)(2)(ii) applies. The regulation at 40 CFR 51.308(f)(2)(ii) requires states to “consult with those States that have emissions that are reasonably anticipated to contribute to visibility impairment in the mandatory Class I Federal area.” The use of the five percent AoI threshold allowed Florida to identify the most important individual out-of-state point sources that “that are reasonably anticipated to contribute to visibility impairment”

at Class I areas within Florida.

See

40 CFR 51.308(f)(2)(ii).

Turning to the Conservation Groups' other source selection comments, they assert that by using a percentage threshold for AoI and PSAT, the calculated threshold in absolute visibility impact terms was higher for Class I areas with the most severe visibility impairment, which resulted in fewer sources being evaluated for reasonable progress for the most visibility-impaired Class I areas. Thus, the Conservation Groups assert that the use of a percentage threshold was unreasonable.

EPA disagrees with this comment. As noted above, states have flexibility to adopt any source selection methodology so long as the methodology is reasonable, and their choices are reasonably explained. A percentage threshold, rather than one using an

absolute visibility threshold (Mm

−

or deciviews), allowed Florida—like every other VISTAS state—to select sources with the largest visibility contributions to each Class I area regardless of the magnitude of visibility impairment at a Class I area. This approach is reasonable. Use of a percentage-based threshold produced a relative ranking of visibility impairment to allow the State to focus on the sources contributing to the largest amount of visibility impact at each individual Class I area. These sources have the potential to reduce visibility impacts the most. The use of a percentage threshold is therefore consistent with the requirement to make reasonable progress toward remedying visibility impairment in each Class I area. EPA finds that Florida's source selection method is reasonable and adequately explained for the reasons discussed above and within our proposal.

Comment 2.b:

The Conservation Groups state that VISTAS considered sulfate and nitrate separately in the PSAT model analyses, which the Conservation Groups allege does not align with how these pollutants actually function in the atmosphere, where sulfate and nitrate act in combination, along with other precursors, to contribute to visibility impairment. As a result, they argue that VISTAS likely underestimated the overall visibility impact of individual sources in its PSAT analysis.

Response 2.b:

EPA disagrees with Conservation Groups' assertion that VISTAS' separate consideration of sulfate and nirate undermines its analysis of visibility impacts. Sulfates and nitrates were modeled together in the PSAT modeling with the other PM species that impact visibility (

e.g.,

direct PM, organic carbon, elemental carbon, etc.). Section 7.6.2 of Florida's 2021 Plan summarizes the results of the PSAT modeling. This section states: “[t]he adjusted PSAT results were used to calculate the percent contribution of each tagged facility to the total sulfate and nitrate point source (EGU + non-EGU) contribution at each Class I area.” Tables 7-16 through 7-18 of the 2021 Plan contain the specific PSAT results for each of Florida's Class I areas. Florida considered the PSAT modeled results for sulfate and nitrate separately only to compare against its selected one percent threshold for each of these pollutants to identify a reasonable number of sources for reasonable progress analyses. The State's approach is reasonable for the reasons discussed above, and it was adequately justified in the Haze Plan and in EPA's NPRM.

Comment 2.c:

The Conservation Groups state that VISTAS used an outdated 2028 emissions projection to “tag” sources. They note that although VISTAS documented that the initial 2028 emission inventory projections were updated for the final modeling, the associated PSAT modeling did not use the final 2028 inventory. The Conservation Groups state that VISTAS scaled predicted sulfate and nitrate to the corresponding changes in SO

and NO

emissions in the updated 2028 inventory using a linear relationship between sulfate and nitrate concentrations. They argue ample evidence shows that there is a non-linear relationship between emissions and sulfate/nitrate concentrations, and that this resulted in additional errors into the modeling. Citing the 2025 Kordzi Report, the Conservation Groups contend that Florida significantly underestimated future 2028 emissions for multiple sources, and that some of these estimates are unjustified and unexplained by Florida.

They also contend that Florida did not explain the 2028 decreases from Foley, Breitburn Operating LP (Breitburn), Mosaic-South Pierce, Monarch Hill, and Gulf Clean Energy Center (Plant Crist).

The Conservation Groups state that nothing in the SIP revision indicates that there have been federally enforceable changes to the Florida facilities' operating parameters that would justify the differences between recent actual emissions and future 2028 projections, and thus, EPA must either present information as a SIP enforceable mechanism to justify these emission reductions or disapprove Florida's source selection process.

The 2025 Kordzi Report is Exhibit 1 to the January 27, 2025, letter from the Conservation Groups and is included in the docket for this rulemaking.

2025 Kordzi Report at 6-7.

Id.

Response 2.c:

EPA disagrees with this comment. VISTAS used the original 2028 emissions inventory to perform the PSAT modeling, and the original PSAT results were linearly scaled to reflect the updated 2028 emissions. Although linear scaling introduces some uncertainty to the final PSAT results, EPA agrees with VISTAS and Florida that adjusting the results to account for VISTAS' updated 2028 emissions inventory using linear scaling is a reasonable approach to account for VISTAS' updated 2028 emissions projections and is a better approach than relying on the original PSAT modeling.

Linear scaling of photochemical modeling results to account for changes in emissions is, in most cases, reasonable and is an accepted practice by EPA. For example, EPA guidance recommends using EPA's Modeled Emission Rates for Precursors (MERPs) for evaluating the impacts of secondary particulate matter of 2.5 micrometers or less in diameter (PM

2.5

) in Prevention of Significant Deterioration (PSD) modeling analyses and allows for and recommends scaling of photochemical modeling results based on emissions.

This guidance recommends an approach where the PM

2.5

impacts are estimated using an archived national-scale photochemical modeling analysis, performed using CAMx and Community Multiscale Air Quality (CMAQ)

photochemical models, that uses hypothetical emissions sources, and then linearly scaling the photochemical modeling results using the ratio of the PSD project-specific source emissions to the modeled emissions from the hypothetical source (

see

equation 1 on page 3 of the referenced April 30, 2024, MERPs memorandum). This approach is widely used and accepted by state air quality agencies and EPA to account for secondarily formed PM

2.5

from precursor emissions (SO

and NO

) for PSD modeling analyses. Since the VISTAS analyses used for regional haze modeling use linear scaling with CAMx and for the same PM

2.5

precursors (SO

and NO

) as the MERPs analyses, EPA finds the method of linear scaling of PM precursor emissions conducted by VISTAS to be acceptable practice.

See

“Clarification on the Development of Modeled Emission Rates for Precursors (MERPs) as a Tier 1 Demonstration Tool for Ozone and PM

2.5

under the PSD Permitting Program,” April 30, 2024, Memorandum from Tyler Fox to Regional Office Modeling Contacts is available at:

https://www.epa.gov/sites/default/files/2020-09/documents/epa-454_r-19-003.pdf.

See https://www.epa.gov/cmaq

for further information on CMAQ.

With respect to the Conservation Groups' comments that Florida significantly underestimated and did not explain future 2028 emission projections for multiple sources, including for Foley, Breitburn, Mosaic-South Pierce, Monarch Hill, and Plant Crist, EPA also disagrees with the Conservation Groups' statements. Florida used the best assumptions available at the time of SIP development to project the 2011 base year emissions out to 2028, including for the facilities noted by the Conservation Groups. The State compared 2017, 2018, and 2019 actual SO

emissions to 2028 projected emissions in table 7-28 of its 2021 Plan. The methodology used to make the 2028 projections is also discussed in appendix B (Emissions Preparation and

Processing) of the 2021 Plan. As discussed in section 4.1 of the 2021 Plan, VISTAS' initial emissions projections for 2028 were completed in June 2018 and the initial modeling was completed in October 2019. After comparing those results to EPA's projected 2028 emissions inventory, which was based on a 2016 base year and was released in September 2019, VISTAS noted differences between their projected inventory and EPA's projected 2028 inventory. VISTAS chose to update their 2028 emissions inventory to incorporate the best estimates of future emissions. This inventory was ultimately used in the 2028 remodeling. The data provided in table 7-28 of the 2021 Plan shows that the projected 2028 emissions for some facilities increased (

e.g.,

Mosaic-South Pierce) and some decreased (

e.g.,

Plant Crist) for using the best information that was available at the time Florida was preparing the final 2021 Plan.

As discussed in Comment/Response 6.e and 8.a below, Florida selected Mosaic-South Pierce and Foley, respectively, for reasonable progress evaluations.

Additionally, Florida considered Breitburn and Plant Crist, and as discussed in section 7.6.4 of the 2021 Plan, concluded that these sources did not need to be evaluated for reasonable progress.

Florida selected Foley for an FFA; however, the facility has fully shut down.

See

Response 8.a.

EPA also disagrees that “a SIP enforceable mechanism” must be put in place for these and other sources that were not selected for evaluation of reasonable progress using an FFA. As discussed in Response 2.a, Florida's source selection methodology is reasonable and is adequately documented in its Haze Plan. The fact that certain sources, including the five sources identified by the Conservation Groups, were not selected for FFAs for either SO

or NO

for this planning period is the result of the reasonable application of Florida's source selection process and source selection thresholds.

Comment 2.d:

The Conservation Groups further claim that Florida did not justify its application of the one percent PSAT threshold for either NO

or SO

and that Florida's use of a percentage-based threshold at the PSAT step biased the process against heavily polluted Class I areas. They note that the U.S. National Park Service's (NPS') comments on the draft SIP revision in 2021 explained that reliance on the percent-based threshold required source impacts to be 80 times larger for the most visually impaired Class I areas versus the least visually impaired Class I areas in order to be selected for an FFA. They also argue that PSAT tagging was unnecessary because the AoI step already identified the sources that contributed to impairment at Class I areas.

Response 2.d:

EPA disagrees with the Conservation Groups' contention that Florida did not justify its application of the one percent PSAT threshold. Section 7.6.4 of the 2021 Plan explains the State's rationale for using a one percent PSAT threshold to select sources for a reasonable progress evaluation. Using a percentage-based threshold enabled the State to identify the sources that contribute most to visibility impairment at the Class I areas, regardless of the magnitude of visibility impairment at each Class I area. Use of a percentage-based threshold produced a relative ranking of impacts on visibility impairment, allowing the State to focus on the sources with the greatest visibility impacts on each individual Class I area. Regardless of whether a relative or absolute threshold was used, Florida's source contribution threshold identified the largest sources to evaluate emissions measures using an FFA. Therefore, the methodology is reasonable and was adequately documented in its Haze Plan.

Regarding the Conservation Groups' assertion that the PSAT tagging process was unnecessary because the AoI step already identified the sources that contributed to impairment at Class I areas, EPA disagrees with the premise of this comment. The standard is not whether the State's source selection approach is necessary or required, but rather, whether the approach is reasonable and is reasonably explained.

The two-step process of screening with the AoI analysis and then applying the more refined PSAT source apportionment modeling to sources that met the initial AoI screening criteria is a sound technical approach for identifying sources to evaluate for reasonable progress. Elements of Florida's AoI approach are discussed in EPA's 2019 Guidance as a viable method to assess sources' visibility impacts to Class I areas.

Florida, along with many of the VISTAS states, also relied upon the AoI initial screening approach in its first planning period Haze Plan. VISTAS used the AoI analysis as an initial screening step because it is a much simpler and less resource intensive approach than using PSAT tagging to model hundreds to thousands of potential sources. The AoI screening approach identified a smaller subset of sources that could undergo refined analysis using PSAT modeling. EPA finds the two-step process of first screening with the AoI analysis followed by use of the more refined PSAT source apportionment modeling to sources is valid and reasonable. Also, as discussed above, states have discretion under the RHR regarding choice of source selection methodology.

See

40 CFR 51.308(f)(2)(i), (iii); 89 FR 105506, 105518 (December 27, 2024).

EPA's 2019 Guidance at 12-14 describes components of Florida's AoI approach, including Q/d (emissions (Q) divided by distance to a Class I area (d)), trajectory analyses, residence time analyses, and source apportionment photochemical modeling (

e.g.,

CAMx PSAT).

Comment 2.e:

The Conservation Groups contend that EPA did not address “significant flaws” in the VISTAS modeling and source selection process and that EPA improperly concluded that Florida's selection of eighteen in-state sources was reasonable because it enabled the identification of sources with the largest visibility impacts. They argue that this is contrary to EPA's guidance which states that a source selection threshold that captures only a small portion of a state's contribution to visibility impairment in Class I areas is more likely to be unreasonable and contrary to the CAA which does not authorize states or EPA to select only the largest contributors to visibility impairment. They assert that Florida should have used a different selection method with a lower threshold, such as a “Q/d” with a threshold of five or lower, to capture the largest portion of in-state sources.

Response 2.e:

EPA disagrees with the assertion that Florida's selection of the 12 largest sources contributing to visibility impairment at Class I areas is contrary to EPA's guidance. The PSAT modeling performed by VISTAS found that the three sources selected by Florida for FFAs have the largest contribution to visibility impairment of any point sources in the State. As discussed in Response 2.a, the PSAT modeling results show that the total cumulative contribution to visibility impairment on the 20 percent most impaired days at Florida's Class I areas from all SO

and NO

emitting sources in the State are relatively small, ranging from 3.12 percent for Everglades to 7.66 percent for Chassahowitzka based on table 2, above.

Given state discretion in selecting sources to evaluate for emissions controls, and since the SO

and NO

emissions from all point sources in Florida contribute a relatively small amount to the visibility impairment at its Class I areas, the State's selection of the three largest

source contributors to visibility impairment is reasonable.

See

footnote 24.

Regarding the Conservation Groups' claim that the State should have adopted a different selection method (such as Q/d) with a lower threshold to select more sources in Florida, as discussed above, a state is not required to evaluate all sources of emissions in each planning period. Instead, a state may reasonably select a set of sources for an analysis of control measures. Selecting a set of sources for analysis of control measures in each planning period is also consistent with the RHR, which sets up an iterative planning process and anticipates that a state may not need to analyze control measures for all sources in a given SIP revision. Moreover, use of Q/d (which simply involves dividing the quantity of emissions by the distance to a Class I area) does not consider transport direction/pathway, dispersion and photochemical processes, or the particular days that have the most anthropogenic impairment due to all sources. Therefore, compared to photochemical modeling, using a simple Q/d technique, as the Conservation Groups suggest, would have resulted in a less accurate quantification of visibility impacts on Class I areas. As discussed in detail above, Florida's reliance on VISTAS modeling and the State's source selection methodology are well documented within the SIP submittal and therefore reasonable.

Comment 2.f:

The Conservation Groups state that EPA asserts in its proposal that Florida's source selection method is reasonable because: (1) SO

and NO

emissions have decreased since the first planning period and are projected to continue decreasing, (2) visibility conditions at in-state Class I areas are projected to improve and have improved since the baseline period, and (3) Florida sources do not contribute to any Class I areas above their respective Uniform Rate of Progress (URP). They argue, however, that projected visibility condition improvement at Florida's Class I areas and the fact that those areas are below their respective URPs are not a valid basis to approve the State's flawed selection method. They cite to EPA guidance stating that the URP is not a safe harbor and that states cannot avoid requiring sources to install reasonable controls merely because there have been emissions reductions due to ongoing air pollution controls since the first planning period or because visibility is projected to improve at Class I areas.

Response 2.f:

As required by the RHR, States must evaluate and determine the emissions reduction measures that are necessary for reasonable progress by considering the four statutory factors.

See

40 CFR 51.308(f)(2)(i). However, we note that emissions from Florida are not reasonably anticipated to contribute to visibility impairment in any Class I areas that are above the 2028 URP, which is relevant to whether a state needs to perform a “robust demonstration” based on the requirements in 40 CFR 51.308(f)(3)(ii)(A) and (B). Therefore, a comparison of the URP to projected visibility impairment in 2028 is needed to inform that requirement. Additionally, other information about measured progress towards natural conditions can be relevant in evaluating the source selection process. For example, significant improvements in visibility at impacted Class I areas since the beginning of the second planning period (starting in 2018) are relevant to whether a state is making progress towards natural conditions and may provide information that could influence the selection of sources to be analyzed for emissions controls in the second planning period. Regardless of the visibility information listed in the proposed rule, EPA independently evaluated Florida's SIP documentation and came to the conclusion that Florida's source selection methodology and thresholds for this second planning period are reasonable for the reasons stated earlier in this response.

Comment 3:

The Conservation Groups contend that Florida arbitrarily and unlawfully refused to conduct FFAs for nine facilities despite their undisputed contribution to visibility impairment in numerous Class I areas. The nine facilities are Duke-Crystal River; JEA Northside; Mosaic-Bartow; Mosaic-New Wales; Mosaic-South Pierce; Nutrien; Seminole; TECO-Big Bend; and Breitburn. They provide the following arguments to support this contention.

First, they state that the text of the CAA and the RHR require the State to evaluate the four statutory factors for any source reasonably anticipated to cause or contribute to any visibility impairment at any Class I area. The Conservation Groups contend that Florida improperly rewrites the statute and regulation to require consideration of the four factors only when a source “significantly contributes” to visibility impairment. Second, they state the structure of the CAA makes clear the requirement to implement emission reductions to ensure reasonable progress is not contingent on whether a source significantly contributes to visibility impairment. They note that Congress expressly uses the modifier “significant” in numerous sections of the CAA and argue that the modifier is conspicuously absent from CAA section 169A. Third, they argue that the purpose of the CAA's visibility provisions to reduce and ultimately eliminate “any impairment of visibility” make clear that Congress “intended for the term `contributes' as used in 7491(b)(2) to encompass

smaller

impacts than would be required to regulate only those sources that contribute `significantly.' ” (emphasis in original). They state that Florida “effectively rewrites those provisions of the Act and requires only the evaluation of emissions that it deems significant or large enough.” The Conservation Groups acknowledge that there is no bright line test for assessing contribution under the RHR, but state that EPA has “made clear that a state's reasonable progress analysis must consider a meaningful set of sources and controls that impact visibility” and that if a state fails to do so, EPA must disapprove the SIP revision and issue a Federal Implementation Plan.

Response 3:

EPA disagrees with the assertion that the CAA and RHR require the State to evaluate the four statutory factors for any source that is “reasonably anticipated to cause or contribute to visibility impairment.” Section 169A(b)(2) of the CAA uses that language, but not for the purposes that the Conservation Groups assert. The CAA requires an implementation plan from a state if emissions from the state “may reasonably be anticipated to cause or contribute to any impairment of visibility in any such area” (referring to out-of-state Class I areas that are impacted by emissions from the state). This is not an individual source requirement. The phrase “may reasonably be anticipated to cause or contribute to any impairment” is only applied to the identification of individual stationary sources in the BART provisions in CAA section 169A(b)(2)(A). But these BART provisions are not applicable in this second planning period SIP evaluation. BART evaluations and emissions limits were only required as part of first planning period regional haze SIPs, and Florida has already met the BART requirements. Additionally, EPA agrees that CAA section 169A and the RHR do not use the phrases “significant contribution” or “significantly contribute” when discussing the four factors. The CAA and RHR do not explicitly list factors that a state must or may not consider when selecting the sources for which it will determine what control measures are necessary to make reasonable progress. The

appropriate threshold for selecting sources may reasonably differ across states and Class I areas due to varying circumstances. In setting a threshold, a state may consider the number of emissions sources affecting the Class I areas at issue, the magnitude of the individual sources' impacts, and the amount of anthropogenic visibility impairment at the Class I areas. As discussed in Response 2.a, Florida considered the magnitude of the individual sources' impacts at Class I areas using AoI screening and PSAT modeling, which is a reasonable approach to identify sources in the State that are reasonably anticipated to cause or contribute to visibility impairment in any Class I area.

Comment 4:

The Conservation Groups assert that EPA incorrectly endorses Florida's decision to exclude consideration of NO

controls in any FFAs. They contend that VISTAS' modeling did not accurately reflect the shift in the 20 percent most impaired days and the corresponding increase in the contribution of nitrate to visibility impairment at Southeastern Class I areas. They state that more of the 20 percent most impaired days now occur in the winter, when nitrate plays a bigger role in visibility impairment, and they note that Florida explained in its SIP that “occasionally nitrate is the predominant visibility impairing pollutant on certain days, generally in winter months.” They also note EPA's general expectation that states will, at a minimum, consider both SO

and NO

in this planning period and assert that there are multiple sources of significant NO

emissions that Florida should have analyzed for NO

controls.

Response 4:

EPA disagrees with this comment. The RHR does not prescribe which visibility impairing pollutants must be evaluated in the FFAs. When selecting sources for analysis of control measures, a state may focus on the PM species that dominate visibility impairment at the Class I areas affected by emissions from the state and then select only sources with emissions of those dominant pollutants and their precursors. EPA has recommended that states that do not evaluate SO

and NO

in both source selection and control evaluations show why consideration of these pollutants would be unreasonable, especially if the state considered both of these pollutants in the first planning period.

Florida considered SO

for FFAs conducted in the first planning period.

Florida followed these recommended approaches here. Florida considered both SO

emissions (via sulfates visibility impacts) and NO

emissions (via nitrates visibility impacts) in the source selection process. As part of the 2021 Plan, FDEP presented the results of PSAT modeling conducted by VISTAS to estimate the projected impact of statewide SO

and NO

emissions across all emissions sectors in 2028 on total light extinction for the 20 percent most impaired days in all Class I areas in the VISTAS modeling domain. The result of this process was that while sources were selected for SO

control analysis determinations, no sources in Florida met the State's nitrate source selection thresholds. Therefore, Florida did not select any sources for a NO

emissions control evaluation. Contrary to the Conservation Groups' assertion that Florida made a “decision” not to consider NO

controls in any FFAs, it was Florida's application of its source selection process, in combination with data and modeling showing that SO

is the dominant visibility impairing pollutant, that resulted in Florida only selecting sources for SO

emissions control analyses and not NO

emissions control analyses.

Additionally, in order to better understand the trends in PM species contributions to visibility impairment, Florida examined more recent IMPROVE monitoring data. More recent IMPROVE monitoring data shows that ammonium sulfate remains the dominant visibility impairing pollutant at Florida's Class I areas as discussed in section 2.5.2 of the 2021 Plan (particularly figures 2-6 through 2-8 for the 2009-2013 period) and in section 2.6.2 (particularly figures 2-9 through 2-11 for the 2014-2018 period). The 2015-2019 IMPROVE monitoring data (the most recent data available at the time) from the IMPROVE website identifies the relative contributions of PM species contributing to the total visibility impairment at the Florida Class I areas, which are shown in table 4, below. In spite of increased nitrate contributions on the 20 percent most impaired days (as the Conservation Groups note, often on winter days), as indicated in that table, ammonium nitrate contributions to regional haze at the State's Class I areas remain relatively low at eight to nine percent of the total visibility impairment as compared to ammonium sulfate at 57 to 60 percent.

See

the spreadsheet containing the 2015-2019 speciated IMPROVE monitoring data for Florida's Class I areas included in the docket for this rulemaking.

Table 4—2015-2019 Speciated IMPROVE Monitoring Data (Percent) for Florida's Class I Areas

Ammonium sulfate

Ammonium nitrate

Organic

carbon

Coarse

mass

Elemental

carbon

Fine

sea salt

Fine

soils

Chassahowitzka

Everglades

St. Marks

Furthermore, in tables 7-21 through 7-23 of the 2021 Plan, the State provided a calculation of the sulfate and nitrate extinction weighted residence time (EWRT) used in the AoI analysis for the Florida Class I areas for the 20 percent most impaired days, demonstrating that the sulfate EWRT are significantly higher than the nitrate EWRT. This further supports the importance of focusing on SO

emissions reductions for this planning period. The State's rationale for focusing on SO

controls in the FFAs is summarized in Florida's SIP submittal and the NPRM.

See

2021 Haze Plan, section 2 (particularly figures 2-9 through 2-13), section 7 (particularly figures 7-17, 7-18, 7-20 through 7-24), and section 10 (particularly figures 10-1); 89 FR 105518-105519.

With respect to the Conservation Groups' assertion that nitrate is the biggest contributor to light extinction on multiple of the 20 percent of most impaired days for these Florida Class I areas during the 2014-2018 period (especially on winter days), as described above, the average nitrate contribution across the 20 percent most impaired days is still relatively small. Thus, while nitrate impairment may be relatively high on a particular day, the data that states are required to use for regional haze as specified in 40 CFR 51.301 and 40 CFR 51.308(f)(1) show ammonium nitrate only contributed eight to nine percent of the total visibility

impairment (during the 2015-2019 period).

For these reasons, Florida's justification for not evaluating sources selected for SO

emission control analyses for a separate NO

emission control analysis is reasonable for this planning period. The trends in PM species' contributions to visibility impairment will continue to be evaluated in future planning periods. If the data warrants consideration of NO

controls in future planning periods, EPA expects that Florida will address potential NO

controls in future regional haze SIP revisions.

Comment 5:

The Conservation Groups assert that EPA ignores that Florida unreasonably excluded sources from FFAs. They state that to correct errors in the source selection method, EPA must require Florida to assess additional EGU and non-EGU emission sources identified by NPS and the Conservation Groups, which have emissions that likely contribute to impairment in Class I areas. Furthermore, the Conservation Groups assert that EPA must find that the State arbitrarily and unlawfully refused to consider cost-effective control upgrades or measures improving efficiency of existing controls, and refused to conduct FFAs on additional sources that contribute to visibility impairment in Class I areas. The comments regarding specific sources identified by the Conservation Groups are addressed in Comments 6 and 7 below.

Response 5:

As explained in Response 2.a and in the NPRM (89 FR 105511), the RHR does not require states to select and consider controls for all sources, all source categories, or any or all sources in a particular source category. Nor does the RHR expressly specify criteria for minimum source selection thresholds. States have discretion to choose reasonable source selection criteria, and sources that meet the state's criteria are selected for an evaluation of potential control options for specific visibility impairing pollutants by considering the four statutory factors in CAA section 169A(g)(1).

As discussed in Response 2.a, Florida's source selection methodology is reasonable and is adequately documented in its Haze Plan. The fact that certain sources, including the 28 sources identified by the Conservation Groups, were not selected for FFAs for either SO

or NO

for this planning period is the result of the reasonable application of Florida's source selection process and source selection thresholds.

As discussed in Response 4, NO

impacts were considered by the State, but no sources were selected for a NO

control evaluation (including these sources highlighted by the Conservation Groups) because visibility impacts for NO

did not exceed the State's source selection threshold. To the extent that the 28 sources identified by the Conservation Groups were not selected by Florida, the Responses to Comments 2 (source selection) and 4 (nitrates/NO

controls) generally address why these sources were not selected and why EPA agrees with the State that it was reasonable to not select these sources for this planning period. To summarize, Florida selected a sufficient number of sources under Florida's jurisdiction to ensure that sources responsible for the largest visibility impacts to Class I areas completed FFAs (or, alternatively, demonstrated that sources have existing, effective controls) for this planning period. Florida has discretion under the RHR to determine its source selection methodology. EPA has found the sources that Florida selected were reasonable and that its Haze Plan complied with the CAA and RHR for this planning period. While Florida could have used its discretion to select other sources in addition to those screened in during its source selection process, including some or all of the sources that the Conservation Groups highlight, Florida was not required to do so. As EPA has stated elsewhere in this notice in Responses 2 and 4, and here in this response, EPA finds Florida's approach to source selection reasonable and appropriate for the second planning period.

In this case, the 28 sources refer to the eight sources that Florida provided an existing, effective controls demonstration for (

see

89 FR 105522) and the 20 sources recommended by the Conservation Groups in the January 27, 2025, Comment Letter.

Comment 6:

The Conservation Groups assert that Florida failed to demonstrate that the nine facilities it eliminated from FFAs on the basis that they are effectively controlled (Duke-Crystal River; JEA Northside; Mosaic-Bartow; Mosaic-New Wales; Mosaic-South Pierce; Nutrien; Seminole; TECO-Big Bend; and Breitburn

) are in fact effectively controlled with existing measures. The Conservation Groups maintain that there are likely cost-effective controls that can be implemented to reduce emissions for each source, and that such controls are therefore necessary to make reasonable progress during the second planning period. Therefore, they assert that EPA's proposal to approve Florida's Haze Plan based on the State's “effectively controlled” determinations for the facilities violates the CAA and the RHR.

Breitburn did not meet Florida's source selection criteria.

See

2021 Plan at 252. Florida therefore did not provide an existing effective SO

controls demonstration in section 7.6.4.1 of the 2021 Plan. The Conservation Groups' comment on Breiturn is addressed in Response 7.

The Conservation Groups comment that the plain language of the CAA and RHR do not allow EPA or the State to eliminate sources from analysis based on the assertion that sources are “effectively controlled.” Instead, they comment that the CAA and RHR require states to consider the four statutory reasonable progress factors for any existing source that is reasonably anticipated to cause or contribute to any impairment of visibility in any Class I area. They assert that Florida determined that the nine facilities mentioned above all contribute to the impairment of visibility in mandatory Class I areas but failed to conduct FFAs for them as required by the Act. The Conservation Groups note that the concept of “effectively controlled” sources only appears in EPA's 2019 Guidance and 2021 Clarification Memo, which they assert is nonbinding and cannot override the plain language of the CAA and RHR. They also assert that EPA has repeatedly explained that states cannot categorically exclude sources from an FFA simply because the source has existing controls and must provide source-specific explanations as to why their decisions for excluding sources from FFAs are reasonable.

The Conservation Groups contend that instead of making the required demonstration in accordance with EPA's guidance, Florida merely claimed that the RHR does not require best available controls, but only measures necessary for reasonable progress. They argue that there are likely feasible and cost-effective controls available for each of the nine facilities, including Breitburn; that the controls are likely reasonable and therefore necessary for reasonable progress; and that EPA must therefore require the inclusion of these controls in Florida's SIP. The Conservation Groups' comments regarding the eight (excluding Breitburn as discussed in footnote 42) sources with an effective SO

controls demonstration are addressed in Responses 6.a through 6.h below. The comments regarding Breitburn are addressed in Response 7.

Response 6:

EPA finds Florida's determination that the affected units at Duke-Crystal River, JEA Northside, Mosaic-Bartow, Mosaic-New Wales, Mosaic-South Pierce, Nutrien, Seminole, and TECO-Big Bend are effectively controlled to be reasonable.

Florida determined that these sources have existing, effective SO

measures and concluded that it would be reasonable to not select such sources for an FFA because an FFA would likely result in the conclusion that no further controls are necessary.

This is consistent with the discretion and flexibilities states have within the CAA and RHR to develop their regional haze SIPs.

FDEP proposed existing SO

measures as necessary for reasonable progress for incorporation into the Florida SIP for the affected units at the following eight facilities: Duke-Crystal River, JEA Northside, Mosaic-Bartow, Mosaic-New Wales, Mosaic-South Pierce, Nutrien, Seminole, and TECO-Big Bend.

See

section 7.6.4.1 of the 2021 Plan.

EPA disagrees that the State cannot rely on existing controls for these eight facilities. The RHR provides flexibility in how its requirements may be addressed, and thus, it may be reasonable for a state not to select a source because the source may already have effective controls in place as a result of a previous regional haze SIP or to meet another CAA requirement. Thus, conducting an FFA would likely result in no new measures found necessary for reasonable progress at the eight aforementioned sources. In Responses 6.a through 6.h, EPA evaluates whether new measures would likely be found reasonable had an FFA been completed for the affected sources.

EPA agrees that guidance cannot override the plain language of the CAA and RHR. EPA's citations to guidance documents in the NPRM were intended to provide further context on what is generally considered to be a reasonable approach to fulfill the statutory and regulatory requirements addressing regional haze for the second planning period. EPA acknowledges that the suggestions in those guidance documents are not binding but are generally assumed to be reasonable. States can deviate from the suggestions within EPA guidance documents. However, they must do so in a reasonable way, accompanied by sufficient justification.

Comment 6.a:

The Conservation Groups argue that Florida should have conducted an FFA for Duke-Crystal River, which is located approximately 20 kilometers (km) north of Chassahowitzka. They assert that the facility has the highest cumulative Q/d value (624.09) of any facility in the State and note EPA proposed to find reasonable Florida's adoption of the Mercury and Air Toxics Standards (MATS) SO

limit,

0.20 pounds per million British thermal unit (lb/MMBtu), for the Fossil Fuel Steam Generating Units 4 and 5 at Duke-Crystal River, and the permit requirements that allow the Citrus Combined Cycle Station Units 1A, 1B, 2A, and 2B to consume only pipeline natural gas. They contend that EPA wrongly takes at face value Florida's assertion that no other controls are likely to be available or cost-effective for this facility.

The MATS Rule is located at 40 CFR part 63, subpart UUUUU.

The Conservation Groups state that Units 4 and 5 are capable of operating “well below” the SO

MATS limit on a continuous basis. According to the Conservation Groups, Unit 5 consistently operated at emission rates below 0.10 lb/MMBtu between 2010 and 2013, with Unit 4 having similar performance. They argue that the only reason these units do not currently operate at these lower rates is because they are not constrained by an enforceable limit. The Conservation Groups maintain that instead of properly responding to public comments regarding the need for Florida to conduct an FFA for the facility, the State merely referred to its general position that EPA's 2019 Guidance notes that the MATS SO

limit of 0.20 lb/MMBtu is “low enough that it is unlikely that an analysis of control measures . . . would conclude that even more stringent control of SO

is necessary to make reasonable progress.” Furthermore, they acknowledge that their analysis of 2017-2023 scrubber efficiencies from these units are “very close to those EPA completed,” but argue that EPA failed to assess the results of its scrubber control efficiency work and recognize that the scrubbers at Units 4 and 5 “are operating sub-optimally” with large swings in the efficiencies, particularly in 2022 and 2023. The Conservation Groups conclude that an SO

FFA would almost certainly result in cost-effective additional SO

controls and that EPA cannot condone the State's lack of support for its determination that a full FFA will likely lead to the conclusion that no further controls are necessary. They maintain that EPA must require an FFA for Units 4 and 5.

Response 6.a:

Regarding Duke-Crystal River's visibility impacts to Class I areas, as discussed in Response 2, Florida considered Q/d as part of its AoI analysis and then further applied PSAT modeling, which differs from the method used by the Conservation Groups. EPA finds Florida's source selection methodology and its results reasonable and has evaluated the PSAT data from Florida as the basis for this action as explained in Response 2.

EPA disagrees that Florida must conduct an FFA for Duke-Crystal River Units 4 and 5 for the reasons discussed in Response 6. EPA also disputes that there was lack of supporting documentation for EPA's proposed conclusions for Units 4 and 5.

Scrubber systems are widely considered the best control technology for reducing SO

emissions, as they can achieve very high removal efficiencies, making them highly effective at capturing SO

from industrial flue gases.

The MATS Rule is a fairly recent CAA requirement with co-benefits for reducing SO

. For the purpose of SO

control measures, an EGU that has add-on flue gas desulfurization (FGD)

and that meets the applicable alternative SO

emission limit of the MATS Rule for power plants is one example of a scenario in which it may be reasonable for a state not to select a particular source for further analysis because the two limits in the rule (0.20 lb/MMBtu for coal-fired EGUs or 0.30 lb/MMBtu for EGUs fired with oil-derived solid fuel) are low enough that it is unlikely that an analysis of control measures for a source already equipped with a scrubber and meeting one of these limits would conclude that even more stringent control of SO

is necessary to make reasonable progress.

See

89 FR 105527 and EPA's analysis of EGUs in Florida found in the spreadsheet file called “FL EGU scrubber efficiency analysis 2017-2023” (hereinafter EGU scrubber efficiency spreadsheet) included in the docket for this rulemaking.

See

section 5, chapter 1, of EPA's “Air Pollution Cost Control Manual” (CCM), available at

https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-reports-and-guidance-air-pollution.

FGD is a type of scrubber system.

See

2019 Guidance at 23.

EPA disagrees that the Agency took at face value Florida's conclusion that no other controls are likely available or cost effective for this facility. EPA analyzed the controls and confirmed that Duke-Crystal River Units 4 and 5 are not uncontrolled or lightly controlled for SO

, are subject to the MATS Rule alternative SO

emission limit of 0.20 lb/MMBtu, and are equipped with wet scrubber systems that routinely achieve a high SO

control effectiveness (with the yearly averages fluctuating between 96.2 to 98.9 percent).

The typical SO

removal efficiency range for wet scrubbers ranges from 90 to 98 percent.

Thus, it is unlikely that an

FFA would result in the conclusion that further SO

emissions controls (including FGD upgrades) are necessary for reasonable progress in the second planning period. Therefore, EPA finds FDEP's effective controls demonstration for Duke-Crystal River to be reasonable. The “swings” in the control efficiencies alleged by the Conservation Groups are attributable to their use of data points that are not representative of the units' normal operation. For example, the Conservation Groups determined the control efficiency for Unit 4 was 72.7 percent during February 2023.

However, the facility did not consume coal that month and did consume a small amount of distillate fuel oil, resulting in a slightly lower average yearly SO

removal efficiency that is not representative of that unit's normal operation.

See

89 FR 105527. Between 2017 to 2023, the yearly average FGD SO

control efficiencies for Duke-Crystal River Unit 4 ranged from 96.2 to 98.9 percent and Unit 5 ranged from 96.4 to 97.6 percent.

See

EGU scrubber efficiency spreadsheet.

See

table 1.1 on page 1-3, section 5, chapter 1 of the CCM.

See

Exhibit 42 attached to the Conservation Groups' comments.

Regarding the assertion that Florida failed to properly respond to the Conservation Groups' comments,

see

Response 12.

Comment 6.b:

The Conservation Groups note that Florida's 2021 Plan incorporates a permitted SO

limit of 0.15 lb/MMBtu for JEA Northside Units 1 and 2 and that Florida later supplemented its “effectively controlled” demonstration by incorporating the SO

MATS limit of 0.20 lb/MMBtu, which applies continuously on a heat input-weighted 30-boiler operating day rolling average.

However, they assert that the State cannot exempt a source from an FFA by relying on controls implemented under other CAA programs. The Conservation Groups argue that the application of the higher SO

MATS limit fails to demonstrate that the units are effectively controlled because Units 1 and 2 have achieved lower SO

emission rates than the MATS limits. Furthermore, the Conservation Groups state that neither JEA Northside nor the State provided adequate documentation to assess the SO

removal efficiency of existing scrubbers, and that the State has not required the facility to conduct an analysis of potential NO

controls for Units 1 and 2. They conclude that EPA cannot approve a SIP that does not require JEA Northside to conduct a full FFA for SO

and NO

See

2024 Supplement at 7.

The Conservation Groups also argue that EPA failed to distinguish “optimized scrubber efficiencies” from “scrubber efficiencies.” They assert that even small improvements in control efficiencies may lead to significant reductions in SO

emissions, with the associated costs being primarily the additional reagent used, electricity for additional spray pumps, and potentially minor capital costs from improving the liquid to gas ratio. The Conservation Groups maintain that EPA relied on Florida's conclusion instead of conducting an independent analysis and that EPA's lack of independent assessment has allowed Florida to wrongly claim that many sources in the State are effectively controlled without considering the most stringent controls achievable.

Response 6.b:

EPA disagrees that the Agency did not independently assess Florida's effectively controlled analysis for JEA Northside Units 1 and 2. EPA prepared and analyzed a spreadsheet providing FGD control efficiencies for the selected Florida power plants, including JEA Northside Units 1 and 2, discussed the data in the NPRM, and included the spreadsheet in the docket.

See

NPRM at 89 FR 105527.

Regarding the assertion that Florida must conduct a NO

FFA for JEA Northside Units 1 and 2,

see

Response 4. With respect to the contention that neither the State nor JEA Northside provided adequate documentation to assess the SO

removal efficiency of existing scrubbers, the NPRM included EPA's assessment of the SO

removal efficiencies for both units to augment the documentation that Florida provided. Regarding the fact that JEA Northside Units 1 and 2 have achieved lower SO

emission rates than the MATS SO

limit of 0.20 lb/MMBtu, these units are each subject to a 0.15 lb/MMBtu SO

limit and EPA expects that these units will operate in compliance with their permitted emissions limits, and thus, actual emissions will routinely be below 0.15 lb SO

/MMBtu.

Scrubber systems are widely considered the best control technology for reducing SO

emissions, as they can achieve very high removal efficiencies, making them highly effective at capturing SO

from industrial flue gases.

The MATS Rule is a fairly recent CAA requirement with co-benefits for reducing SO

. For the purpose of SO

control measures, an EGU that has add-on FGD and that meets the applicable alternative SO

emission limit of the MATS Rule for power plants is one scenario in which it may be reasonable for a state not to select a particular source for further analysis because the two limits in the rule (0.20 lb/MMBtu for coal-fired EGUs or 0.30 lb/MMBtu for EGUs fired with oil-derived solid fuel) are low enough that it is unlikely that an analysis of control measures for a source already equipped with a scrubber and meeting one of these limits would conclude that even more stringent control of SO

is necessary to make reasonable progress.

EPA's analysis confirms that JEA Northside Units 1 and 2 are not uncontrolled or lightly controlled for SO

; are subject to the MATS Rule alternative SO

emission limit of 0.20 lb/MMBtu (30-boiler operating day rolling average), a SO

emission limit of 0.15 lb/MMBtu (30-day rolling average), and a SO

emission limit of 0.2 lb/MMBtu (24-hour block average); and are equipped with wet scrubber systems that routinely achieve a high SO

control effectiveness (approximately

94.8

96.6 percent).

As discussed above, the typical SO

removal efficiency range for wet scrubbers ranges from 90 to 98 percent.

Thus, it is unlikely that an FFA would result in the conclusion that further SO

emissions controls (including FGD upgrades) are necessary for reasonable progress. Therefore, EPA finds FDEP's effective controls determination for Northside Units 1 and 2 to be reasonable.

See

section 5, chapter 1, of the CCM, available at

https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-reports-and-guidance-air-pollution.

See

2019 Guidance at 23.

See

89 FR 105527. Between 2017 to 2023, the yearly average FGD SO

control efficiencies for JEA Northside Unit 1 ranged from 94.9 to 96.3 percent and Unit 2 ranged from 94.8 to 96.6 percent.

See

EGU scrubber efficiency spreadsheet that is included in the docket for this rulemaking.

See

table 1.1 on page 1-3, section 5, chapter 1 of the CCM.

Comment 6.c:

The Conservation Groups state that the Mosaic-Bartow, located 105 km from Chassahowitzka, is a significant source of SO

and NO

based on NPCA's analysis of 2020 NEI data (showing 2,907 tpy of SO

emissions and 153 tpy of NO

emissions). They also state that the facility likely impacts six Class I areas and has a “very high” cumulative Q/d of 85.69. The Conservation Groups conclude that despite the facility's significant SO

emissions, the State determined it was effectively controlled and failed to conduct an FFA.

The Conservation Groups argue that while Florida proposed to exempt the SO

emissions from Sulfur Acid Plants (SAPs) 4-6 because they are limited to 4 pounds of SO

per ton of 100 percent sulfuric acid produced (lbs/ton), the State failed to specify averaging periods, provide a monitoring plan, and provide an opportunity for the emission limits and monitoring, recordkeeping, and reporting requirements to be reviewed and commented on. Furthermore, they

claim that Florida failed to demonstrate the facility is effectively controlled and that an FFA would not identify additional necessary controls. They also state that, by only looking at EPA's “incomplete” Reasonably Available Control Technology (RACT), Best Available Control Technology (BACT), and Lowest Achievable Emission Rate (LAER) Clearinghouse (RBLC) database,

the State did not conduct a meaningful search of control options. The Conservation Groups add that the State did not document the information it found in the database, which they explain is contrary to the RHR's requirements to document the information it relied on for its SIP revision, preventing the public from being able to meaningfully review and comment on the State's analysis for the facility. Additionally, they state that “the range of 3.0 to 4.0 lbs/ton represents a

potential increase

of 33% in the SO

emissions.”

EPA's RBLC is available at:

https://www.epa.gov/catc/ractbactlaer-clearinghouse-rblc-basic-information.

EPA's Clean Air Technology Center (CATC) maintains a permit database called the RBLC. The RBLC contains information about recent control technology determinations submitted by state and local agencies.

Citing to 2021 Kordzi Report at 12-13.

The Conservation Groups also state that they provided these comments to the State and the State failed to meaningfully engage and respond. The Conservation Groups conclude that nothing in the record supports EPA's assertion that units are effectively controlled for SO

and that additional reasonable controls are unlikely to be found, stating that EPA merely replicates the State's assertions and fails to document supporting information. Due to these issues, as well as others highlighted in the Kordzi Reports, they contend that the State has not demonstrated that the controls are equivalent to the best performing controls or conduct/require an FFA. Therefore, they argue that EPA must require Mosaic-Bartow to complete an FFA.

Response 6.c:

Regarding Mosaic-Bartow's visibility impacts to Class I areas, as discussed in Response 2, Florida considered Q/d as part of its AoI analysis and then further conducted PSAT modeling, which differs from the method used by the Conservation Groups. EPA finds Florida's source selection methodology and its results reasonable and has evaluated the PSAT data from Florida as the basis for this action as explained in Response 2.

EPA disagrees that Florida proposed to exempt the SO

emissions from SAP 4-6 from an FFA because the units are limited to a 4 lbs/ton production limit. Florida determined that SAPs 4-6 are effectively controlled and that additional reasonable controls are unlikely to be found through an FFA because (1) SAPs 4, 5, and 6 utilize double-absorption technology with vanadium promoted catalysts for the first three converter beds and a cesium-promoted catalyst for the fourth bed that oxidize SO

generated from the sulfur furnace to form sulfur trioxide (SO

) at a conversion efficiency of 99.7 percent or higher,

and (2) the SIP contains a three-unit cap of 1,100 pounds per hour (lbs/hr) on a 24-hour block average (as determined by a continuous emission monitoring system (CEMS)), which is more stringent than the production limit of 4 lbs/ton,

a permit limit that the State determined is consistent with BACT determinations in the RBLC for sulfur burning, double-absorption SAPs with cesium-promoted catalysts.

See

AP 42, Fifth Edition, Volume I Chapter 8.10 available at:

https://www.epa.gov/sites/default/files/2020-09/documents/8.10_sulfuric_acid.pdf; see also

“Background Report AP-42 Section 5.17 Sulfuric Acid” available at:

https://www.epa.gov/sites/default/files/2020-09/documents/final_background_document_for_sulfuric_acid_section_8.1_1.pdf;

2021 Plan at 255 and appendix G-2e.

See

Air Plan Approval and Designation of Areas; FL; Source-Specific SO

Permit Limits & Redesignation of Hillsborough-Polk 2010 1-Hr SO

Nonattainment Area to Attainment & Mulberry Unclassifiable Area to Attainment/Unclassifiable, 84 FR 47216, 47219 (September 9, 2019) (noting that the 1,100 lbs/hr limit reduced potential SO

emissions from 5,694 tpy (pursuant to the 4 lbs/ton production limit) to 4,818 tpy).

See also

FDEP, Proposed Revision to State Implementation Plan, Submittal Number 2017-04, Incorporation of SO

Emissions Limits for Two Facilities in Polk County (December 1, 2017) at 11-12, available as document EPA-R04-OAR-2018-0510-0008 in the

www.regulations.gov

docket for the September 9, 2019 action (hereinafter 2017 Florida SIP Revision).

The facility recently upgraded its catalyst beds and accepted the 1,100 lbs/hr SIP-approved limit to bring the Hillsborough-Polk nonattainment area into attainment for the 2010 SO

National Ambient Air Quality Standards (NAAQS).

The 4 lbs/ton limit, from New Source Performance Standards (NSPS) subpart H (

see

40 CFR 60.82(a)), is included in the facility's operating permit,

and is comparable to a SO

conversion efficiency of 99.7 percent. Florida's Haze Plan does not contain the 4 lbs/ton limit because the SIP already contains the more stringent 1,100 lbs/hr cap.

Florida evaluated SO

BACT determinations for sulfur burning, double-absorption SAPs with cesium-promoted catalysts in the RBLC and determined that they are in the range of 3.0 to 4.0 lbs/ton. EPA performed a search of SAPs from 2000 to 2025 in the RBLC and found many instances where the double absorption process is associated with BACT.

Regardless, a source is not required to meet BACT to satisfy the RHR, and EPA is not performing a BACT analysis in this action. Because Mosaic-Bartow utilizes double-absorption technology with catalytic enhacement and is achieving a SO

conversion efficiency greater than 99.7 percent under the 1,100 lbs/hr cap, EPA finds that it is unlikely that an FFA would conclude that even more stringent control is necessary for reasonable progress and finds Florida's effective controls determination for SAPs 4-6 to be reasonable.

See

85 FR 9666.

See

Condition E.6. of Permit No.1050046-091-AV.

See

85 FR 9666.

The statutory considerations for selecting BACT are similar to, if not more stringent than, the four statutory factors for reasonable progress.

See

2019 Guidance at 23.

See

also EPA's RBLC search result included in the docket for this rulemaking.

With respect to the comment regarding the State's alleged failure to specify averaging periods, provide a monitoring plan, and provide an opportunity for public comment for the 4 lbs/ton limit, the State was not required to perform these tasks because the State is not adding this limit to its SIP. Although the Conservation Groups do not address the 1,100 lbs/hr limit, EPA notes that the public had the opportunity to comment on that limit during the 2019 rulemaking incorporating the limit into the SIP,

during the state-level public comment period on the draft Haze Plan, and during the public comment period for EPA's proposed rulemaking on the Haze Plan.

The final rule at 85 FR 9666 was preceded by a notice of proposed rulemaking published on September 9, 2019 (84 FR 47216). The public comment period closed on October 9, 2019.

The comment that “the range of 3.0 to 4.0 lbs/ton of sulfuric acid produced represents a potential increase of 33% in the SO

emissions” is unclear as is its relationship to the 1,100 lbs/hr SIP-approved limit, but it appears to relate to the change between these two values to support the contention that “such a wide range should not be used to characterize the acceptable range of best performing controls.” There is no requirement in the CAA or the RHR for second period regional haze plans to evaluate and/or select the most stringent (“best”) control option for selected sources, and as discussed above, the 1,100 lbs/hr limit results in a SO

conversion efficiency greater than 99.7 percent.

EPA disagrees that Florida is required to demonstrate that Mosaic-Bartow has the “best performing controls” as part of its existing, effective controls

demonstration. There is no statutory or regulatory requirement to have each selected source evaluate and/or adopt the most stringent controls or emission limits. Rather, states are required to include in the LTS the measures necessary for reasonable progress, which Florida did for Mosaic-Bartow.

The 2021 Plan explains that existing SO

measures identified in the existing controls analysis are already adopted into the Florida SIP for Mosaic-Bartow (85 FR 9666 (February 20, 2020)).

See

2021 Plan at 255.

Regarding the comment that Florida needs to explain a statement in the State's December 2017 SIP revision that “the production-based emissions limits at the 3 sulfuric acid plants of 4 lbs SO

/ton of 100% H

are effectively lowered to 3.4 lbs SO

/ton of 100% H

”

and how this affects the limits discussed on page 255 of the 2021 Plan, it is self-evident that the 3.4 lbs SO

/ton production-based limit is the effective production-based limit equivalent of the SIP-approved 1,100 lbs/hr cap. The effectively equivalent limit does not affect the limits discussed on page 255 of the 2021 Plan because Florida is relying on the 1,100 lbs/hr cap for regional haze purposes, not the 4 lbs SO

/ton production-based limit

See

2017 Florida SIP Revision at 12.

With respect to the assertion that Florida's response to comments failed to meaningfully engage and respond to the Conservation Groups' comments,

see

Response 12.

Regarding Florida's use of the RBLC and documentation of information from the RBLC, Florida evaluated SO

BACT determinations for sulfur burning, double-absorption SAPs with cesium-promoted catalysts in the RBLC and determined that they are in the range of 3.0 to 4.0 lbs/ton. The RBLC is publicly available and may be searched by any member of the public. Although the RHR requires the State to document the technical basis, including modeling, monitoring, cost, engineering, and emissions information, on which the State is relying to determine the emission reduction measures that are necessary to make reasonable progress in each mandatory Class I area it affects, it does not specify what that documentation must consist of.

See

40 CFR 51.308(f)(2)(iii). It was unnecessary in this instance for the State to provide additional documentation because Florida identified the RBLC as the source of information and provided the criteria necessary for the public to replicate the analysis. As discussed above, EPA conducted its own search of the RBLC for SAPs using a timeframe from 2000 to 2025 to independently assess Florida's findings.

EPA used the industry process type code of 62.015 for “sulfuric acid plants,” selecting “SOX” as the pollutant name.

See

EPA's RBLC search result.

Comment 6.d:

The Conservation Groups state that Mosaic-New Wales, located 103 km from Chassahowitzka, is a significant source of SO

and NO

based on NPCA's analysis of 2020 NEI data (showing 4,002 tpy of SO

emissions and 218 tpy of NO

emissions). They also state that the facility likely impacts 11 Class I areas and has a “very high” cumulative Q/d of 147.51. The facility was one of the sources the State determined was effectively controlled; however, the Conservation Groups argue that the State failed to demonstrate that the controls are equivalent to the best performing controls or conduct an FFA.

The Conservation Groups reference their state-level comments to Florida regarding Mosaic-New Wales, including a comment that the range of 3.0 to 4.0 lbs/ton is potentially a 33 percent increase in SO

emissions, and claim that the State failed to meaningfully engage and respond to their comments. They argue that EPA is rubber stamping Florida's approach, fails to recognize the issues the Conservation Groups identified, and provides no separate justification to exempt Mosaic-New Wales from the FFA requirement. Therefore, they argue that EPA must require Mosaic-New Wales to be subjected to an FFA.

Response 6.d:

Regarding Mosaic-New Wales' visibility impacts to Class I areas, as discussed in Response 2, Florida considered Q/d as part of its AoI analysis and then further conducted PSAT modeling, which differs from the method used by the Conservation Groups. EPA finds that Florida's source selection methodology and its results are reasonable and has evaluated the PSAT data from Florida as the basis for this action, as explained in Response 2.

Florida proposed to exempt the SO

emissions from SAPs 1-5 from an FFA because (1) these units utilize double-absorption technology with vanadium-promoted catalyst for the first three converter beds and cesium-promoted catalyst for the fourth bed that oxidize SO

generated from the sulfur furnace to form SO

at a conversion efficiency of 99.7 percent or higher,

and (2) the SIP contains a five-unit cap of 1,090 lbs/hr on a 24-hour block average (as determined by CEMS), which is more stringent than the production limit of 3.5 lbs/ton, a limit included in the facility's operating permit for SAPs 1-3, and the production limit of 4 lbs/ton for SAPs 4-5,

which is a limit that the State determined is consistent with BACT determinations in the RBLC for sulfur burning, double-absorption SAPs with cesium-promoted catalysts.

See

AP 42, Fifth Edition, Volume I Chapter 8.10 available at:

https://www.epa.gov/sites/default/files/2020-09/documents/8.10_sulfuric_acid.pdf; see

also “Background Report AP-42 Section 5.17 Sulfuric Acid” available at:

https://www.epa.gov/sites/default/files/2020-09/documents/final_background_document_for_sulfuric_acid_section_8.1_1.pdf;

2021 Plan at 255 and appendix G-2f.

See

84 FR 47216, 47219 (September 9, 2019) (noting that the 1,090 lbs/hr limit reduced potential SO

emissions from 10,750 tpy (under the 4 lbs/ton and 3.5 lbs/ton production limits) to 4,774 tpy).

See

2017 Florida SIP Revision at 11.

The facility recently upgraded its catalyst beds and accepted the 1,090 lbs/hr SIP-approved limit to bring the Hillsborough-Polk nonattainment area into attainment for the 2010 SO

NAAQS.

The 4.0 lbs/ton limit is from NSPS subpart H (

see

40 CFR 60.82(a)), is included in the facility's operating permit, and is comparable to a SO

conversion efficiency of 99.7 percent. Florida's Haze Plan does not contain the 3.5 lbs/ton or 4 lbs/ton limit because the SIP already contains the more stringent 1,090 lbs/hr cap.

Florida evaluated SO

BACT determinations for sulfur burning, double-absorption SAPs with cesium-promoted catalysts in the RBLC and determined that they are in the range of 3.0 to 4.0 lbs/ton. EPA performed a search of SAPs from 2000 to 2025 in the RBLC and found many instances where double absorption technology is associated with BACT (

e.g.,

Mississippi Phosphates Company in Jackson, Mississippi; PCS Phosphate Company in Beaufort, North Carolina; Plant City Phosphate Complex in Hillsoborough, Florida).

Regardless, a source is not required to meet BACT to satisfy the RHR, and EPA is not performing a BACT analysis in this action. Because Mosaic-New Wales utilizes double-absorption technology with catalytic enhacement and is achieving a SO

conversion efficiency greater than 99.7 percent under the 1,090 lbs/hr cap, EPA finds that it is unlikely that an FFA would conclude that even more stringent control is necessary for reasonable progress and finds Florida's effective control determination for SAPs 1-5 to be reasonable.

See

85 FR 9666.

See

footnote 70.

The statutory considerations for selecting BACT are similar to, if not more stringent than, the four statutory factors for reasonable progress.

See

2019 Guidance at 23.

EPA disagrees that Florida is required to demonstrate that Mosaic-New Wales has the “best performing controls” as

part of its existing, effective controls demonstration. There is no statutory or regulatory requirement to have each selected source evaluate and/or adopt the most stringent controls or emission limits for second planning period haze plan. Rather, states are required to include in the LTS the measures necessary for reasonable progress, which Florida did for Mosaic-New Wales.

EPA finds that Florida's conclusion that Mosaic-New Wales SAPs 1-5 has existing effective SO

controls to be reasonable.

The 2021 Plan explains that existing SO

measures deemed necessary for reasonable progress for the second planning period are already adopted into the Florida SIP for Mosaic-New Wales (85 FR 9666 (February 20, 2020)).

The comment that “the range of 3.0 to 4.0 lbs/ton of sulfuric acid produced represents a potential increase of 33% in the SO

emissions” is unclear, as is its relationship to the 1,090 lbs/hr SIP-approved limit, but it appears to relate to the change between these two values to support the contention that “such a wide range should not be used to characterize the acceptable range of best performing controls.” There is no requirement in the CAA or the RHR for second period regional haze plans to evaluate and/or select the most stringent (“best”) control option for selected sources, and as discussed above, the 1,090 lbs/hr limit results in a SO

conversion efficiency greater than 99.7 percent.

Regarding the comment that Florida needs to explain a statement in the State's December 2017 SIP revision that “the production-based emissions limits at the five sulfuric acid plants of 3.5 and 4 lbs SO

/ton of 100 percent H

are effectively lowered to 1.6 & 1.8 lbs SO

/ton of 100 percent H

, respectively,”

and how this affects the limits discussed on page 255 of the 2021 Plan, it is self-evident that these 1.6 and 1.8 lbs SO

/ton production-based limits are the effective production-based limit equivalents of the SIP-approved 1,090 lbs/hr cap. These effectively equivalent limits do not affect the limits discussed on page 255 of the 2021 Plan because Florida is relying on the 1,090 lbs/hr cap for regional haze purposes, not the 3.5 lbs SO

/ton or 4 lbs SO

/ton production-based limit.

See

2017 Florida SIP Revision at 11.

With respect to the assertion that Florida's response to comments failed to meaningfully engage and respond to the Conservation Groups' comments,

see

Response 12.

Comment 6.e:

The Conservation Groups state that Mosaic-South Pierce, located 114 km from Chassahowitzka, is a significant source of SO

and NO

based on NPCA's analysis of 2020 NEI data (showing 1,739 tpy of SO

emissions and 66 tpy of NO

emissions). They also state that the facility likely impacts three Class I areas and has a “very high” cumulative Q/d of 35.81.

The Conservation Groups contend that their comments to Florida identified numerous issues with the State's determination that SAPs 10 and 11 are effectively controlled and that Florida failed to meaningfully consider and respond to their comments. They also contend that EPA failed to recognize “Florida's short-comings and misrepresentations” identified in NPS' comments to the State. According to the Conservation Groups, NPS identified several facilities in the RBLC with additional post-process controls, including scrubbers (hydrogen peroxide or caustic scrubbers) and/or mist elimination with emission limits as low as 0.15 lb SO

/ton of sulfuric acid and noted that Idaho's second planning period Regional Haze SIP found wet flue gas desulfurization (WFGD), hydrogen peroxide scrubbers, and dry sorbent injection (DSI)/caustic scrubbers to be technically feasible. The Conservation Groups assert that Florida failed to consider this information and responded by stating that it had reviewed the information regarding the use of post-process scrubbers and determined it would not be cost-effective. They argue the Second 2024 Supplement provided “no actual documented and reasoned determination for this position.” The Conservation Groups conclude that EPA must require Mosaic-South Pierce to conduct an FFA.

Response 6.e:

Regarding Mosaic-South Pierce's visibility impacts to Class I areas, as discussed in Response 2, Florida considered Q/d as part of its AoI analysis and then further conducted PSAT modeling, which differs from the method used by the Conservation Groups. EPA finds Florida's source selection methodology and its results reasonable and has evaluated the PSAT data from Florida as the basis for this action.

Regarding the comment that additional scrubbers could have been evaluated at SAPs 10 and 11 at Mosaic-South Pierce, there is no statutory or regulatory requirement to evaluate all technically feasible control measures each period (

see

also Response 6). Rather, states are required to include in the LTS the measures necessary for reasonable progress, which Florida did for Mosaic-South Pierce.

However, the RHR does not require states to evaluate or implement all possible cost-effective controls including all available permutations of each control option. EPA's RBLC search identified many instances where double absorption process is associated with BACT for SAPs.

The 2024 Supplement explains that existing SO

measures deemed necessary for reasonable progress for the second planning period are already adopted into the Florida SIP for Mosaic-South Pierce (

see

88 FR 51702, August 4, 2023).

See

EPA's RBLC search result.

SAPs 10 and 11 utilize double-absorption technology with vanadium-promoted catalyst for the first three converter beds and cesium-promoted catalyst for the fourth bed that oxidize SO

generated from the sulfur furnace to form SO

at a conversion efficiency of 99.7 percent or higher.

Appendix B-2 of Florida's 2024 Supplement contains the RBLC results from Mosaic showing that the use of double absorption technology with cesium promoted catalyst represents BACT in the range of 3.0 to 4.0 lbs/ton for SO

emissions.

Regardless, a source is not required to meet BACT to satisfy the RHR, and EPA is not performing a BACT analysis in this action. Because Mosaic-South Pierce utilizes double-absorption technology with catalytic enhacement and is achieving a conversion efficiency of 99.7 percent or higher, EPA finds that it is unlikely that an FFA would conclude that even more stringent control is necessary for reasonable progress and finds Florida's effective control determination to be reasonable.

See

2024 Supplement and appendix B-2 of the 2024 Supplement.

The statutory considerations for selecting BACT are similar to, if not more stringent than, the four statutory factors for reasonable progress.

See

2019 Guidance at 23.

EPA is aware that NPS identified facilities in the RBLC with additional post-process controls and that Idaho's second planning period regional haze SIP found WFGD, hydrogen peroxide scrubbers, and DSI/caustic scrubbers to be technically feasible for Itafos Conda. However, there is no CAA or RHR requirement for second period haze plans that specifies the scope of technically feasible control options to evaluate for each unit type and pollutant. Thus, Florida reasonably applied this discretion by focusing on the main form of SO

control in use currently at other SAPs: double absorption technology with catalytic enhancement.

Regarding Itafos Conda,

the Conservation Groups did not acknowledge the subsequent analysis the facility submitted to the Idaho Department of Environmental Quality that addressed issues with the original analysis and determined that no additional controls beyond dual-absorption technology with catalytic enhancement were reasonable.

See

the RBLC search results included in the 2021 Plan (appendices G-2e, G-2f, and G-2g) and 2024 Supplement (appendix B-2).

See

also RBLC

results documented in the Conservation Groups' Exhibit 36 and EPA's RBLC search result.

See

Idaho's “Regional Haze State Implementation Plan for the 2nd Planning Period” (June 2022) (erroneously marked with a “draft” watermark) included in the docket for this rulemaking; “Supplement to Idaho Regional Haze State Implementation Plan for the Second Planning Period” (August 2024) included in the docket; 90 FR 13516, 13531-32 (March 24, 2025).

With respect to the assertion that Florida's response to comments failed to meaningfully engage and respond to the Conservation Groups' comments,

see

Response 12.

Comment 6.f:

The Conservation Groups state that Nutrien, located 37 km from Okefenokee, is a significant source of SO

and NO

based on NPCA's analysis of 2020 NEI data. They also state that the facility likely impacts four Class I areas and has a “very high” cumulative Q/d of 77.26.

The Conservation Groups argue that Florida incorrectly exempted Nutrien from an FFA based on its determination that seven-year-old upgrades to the SAPs, required by a consent decree, are consistent with recent BACT determinations for similar SAPs. They also argue that Florida did not meaningfully engage and respond to their comments, EPA's suggestion that a 10-year-old consent decree is “recent” is “not true,” and EPA's proposal contradicts the record. The Conservation Groups assert that assuming upgrades required by a consent decree are adequate is not a substitute for an FFA and that the record shows that there are other plants with much lower limits than those for the facility. The Conservation Groups conclude that EPA must require Nutrien be subjected to an FFA.

Response 6.f:

Regarding Nutrien's visibility impacts to Class I areas, as discussed in Response 2, Florida considered Q/d as part of its AoI analysis and then further conducted PSAT modeling, which differs from the method used by the Conservation Groups. EPA finds Florida's source selection methodology and its results reasonable and has evaluated the PSAT data from Florida as the basis for this action as explained in Response 2.

With respect to the 2015 consent decree, as stated in the NPRM (89 FR 105523), Nutrien completed upgrades to its catalysts on SAP E and SAP F which enabled these SAPs to meet new SO

emission limits of 2.6 lbs/ton of sulfuric acid on a three-hour rolling average (excluding startups and shutdowns) and 2.3 lbs/ton of sulfuric acid on a 365-day rolling average (including startups and shutdowns), pursuant to the consent decree.

Nutrien came into compliance with these limits on January 1, 2018, for SAP F and January 1, 2020, for SAP E. Thus, compliance with the consent decree occurred five to seven years ago, within the second planning period.

SAPs E and F utilize the double-absorption process with vanadium-promoted catalyst for the first three converter beds and vanadium/cesium-based catalyst for the fourth bed that oxidize SO

to SO

at a conversion effiency greater than 99.7 percent.

See

appendix G-2g of the 2021 Plan.

With respect to the comments regarding the 2015 consent decree, EPA does not, as the Conservation Groups suggest, generally assume that upgrades required by a consent decree are an adequate substitute for an FFA. EPA and the State evaluated the specific requirements of the 2015 consent decree for Nutrien and determined that they reflect existing effective controls for this facility. The consent decree resolved allegations that Nutrien (and/or their predecessors in interest) constructed or modified, and then operated, the Nutrien facility without, among other things, installing BACT.

This suggests that the limits in the consent decree represented BACT for the Nutrien SAPs at the time of execution in 2015.

The 2020 effective control analysis for the facility in appendix G-2 of the 2021 Plan also evaluated the RBLC for BACT determinations made on sulfur-burning SAPs with catalyst enhancement and concluded that the emission limits for SAPs E and F are “consistent with, and equivalent to the most recent BACT determinations made for similar double-absorption, sulfur-burning sulfuric acid plants.” The analysis also states that “there have been no new developments in catalyst technology and/or strategies for operating SAPs since these BACT determinations have been made.” Regardless, a source is not required to meet BACT to satisfy the RHR, and EPA is not performing a BACT analysis in this action. Because Nutrien utilizes double-absorption technology with catalytic enhancement and is achieving a SO

conversion efficiency of greater than 99.7 percent under the consent decree's SO

emission limits, EPA finds that it is unlikely that an FFA would conclude that even more stringent control is necessary for reasonable progress and finds Florida's effective control determination to be reasonable.

See

2015 Consent Decree at 2.

As discussed above, the statutory considerations for selecting BACT are similar to, if not more stringent than, the four statutory factors for reasonable progress.

See

2019 Guidance at 23. EPA believes that a BACT determination within eight years of a SIP submission for the second planning period should be consistent with up-to-date, effective, and reasonable control measures.

Id.

Regarding the comment that other SAPs are achieving lower emissions limits, there is no statutory or regulatory requirement to achieve the lowest possible emissions rate for reasonable progress (

see

also Response 6). Rather, states are required to include in the LTS the measures necessary for reasonable progress, which Florida did for Nutrien. EPA finds Florida's determination that Nutrien SAPs E and F are effectively controlled for SO

reasonable.

With respect to the assertion that Florida's response to comments failed to meaningfully engage and respond to the Conservation Groups' comments,

see

Response 12.

Comment 6.g:

The Conservation Groups argue that EPA wrongly proposes to find Florida's determination that Seminole's Unit 1 and Unit 2 are effectively controlled for SO

(including the determination to accept the MATS limit of 0.20 lb/MMBtu) reasonable and merely repeats Florida's flawed contentions without scrutiny. Furthermore, they claim the RHR has consistently acknowledged that scrubber upgrades are cost-effective and should be considered to ensure reasonable progress; EPA's guidance that FGD systems installed since 2007 should achieve at least 95 percent effectiveness does not exempt states from evaluating feasible and cost-effective reductions; and Florida arbitrarily ignores achievable emission reductions and fails to consider technically and economically feasible upgrades to scrubbers and selective catalytic reduction (SCR) systems. They state that EPA's data evaluation from 2017 to 2023 showed that Seminole is historically capable of achieving 96.5 to 97.3 percent year average SO

control efficiencies, with a seven-year average SO

removal efficiency of 96.8 percent, and acknowledge that their analysis of 2017-2023 scrubber efficiencies from these units are “very close” to EPA's evaluation. However, they argue that EPA fails to recognize that the wet scrubber systems for Seminole Unit 2 is operating sub-optimally with large swings in efficiencies, particularly in 2022 and 2023. The Conservation Groups state that the scrubber at Unit 2 has historic efficiency levels as high as 97.2 percent but a 95.8 percent control efficiency in 2023. Therefore, they claim an SO

FFA would almost certainly lead to additional cost-effective controls for

emissions because currently the wet scrubber systems are “exhibiting large swings in their scrubber efficiencies, particularly in 2022 and 2023.”

They state that since Florida did not provide analysis demonstrating that Seminole Unit 2 is effectively controlled, EPA wrongly proposes to find the State's determination reasonable. Instead, they assert that it is necessary to conduct a full FFA of the unit.

See

2025 Kordzi Report at 24.

Response 6.g:

EPA disagrees that the Agency did no further analysis of Florida's demonstration that Seminole Units 1 and 2 have existing effective controls for SO

In fact, as acknowledged by the Conservation Groups, EPA prepared and analyzed scrubber control efficiency data for Seminole Units 1 and 2 using 2017-2023 data.

See

89 FR 105528 and the EGU scrubber efficiency spreadsheet.

The Conservation Groups state that Seminole Unit 1 appears to have retired. According to EPA's

Clean Air Markets Program Data (

CAMPD) website

and the United States Energy Information Administration (EIA),

Unit 1 is retired as of December 2023.

See

CAMPD data available at:

https://campd.epa.gov/data/custom-data-download.

See

Preliminary Monthly Electric Generator Inventory (based on Form EIA-860M as a supplement to Form EIA-860), specifically the January 2024 spreadsheet available at:

https://www.eia.gov/electricity/data/eia860m/.

Scrubber systems are widely considered the best control technology for reducing SO

emissions, as they can achieve very high removal efficiencies, making them highly effective at capturing SO

from industrial flue gases.

The MATS Rule is a fairly recent CAA requirement with co-benefits for reducing SO

. For the purpose of SO

control measures, an EGU that has add-on FGD and that meets the applicable alternative SO

is necessary to make reasonable progress.

See

section 5, chapter 1, of the CCM available at

https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-reports-and-guidance-air-pollution.

See

2019 Guidance at 23.

Regarding comments that scrubber upgrades are not expensive and that 95 percent scrubber control efficiency does not exempt an EGU from further analysis as effectively controlled for SO

, EPA's analysis confirms that the units are not uncontrolled or lightly controlled, are subject to the MATS Rule alternative SO

emission limit of 0.20 lb/MMBtu, and are equipped with WFGD that routinely achieve a high SO

control effectiveness. As stated in the NPRM (89 FR 105528), EPA calculated FGD control efficiencies at Units 1 and 2 at Seminole during periods when coal is one of the fuel sources consumed over the 2017-2023 period and calculated that the existing FGD systems routinely achieve 96.5 to 97.3 percent yearly average SO

removal efficiencies, with a seven-year average (2017-2023) SO

removal efficiency of 96.8 percent.

As stated above, the typical SO

removal efficiency range for wet scrubbers ranges from 90 to 98 percent.

Thus, it is unlikely that were an FFA completed, these existing control efficiencies could be improved cost-effectively and result in meaningful emissions reductions. Therefore, EPA finds FDEP's effective controls determination for Seminole Units 1 and 2 to be reasonable. The “swings” in the control efficiencies alleged by the Conservation Groups are attributable to the their use of data points that are not representative of unit's normal operation. For example, the Conservation Groups determined the control efficiency for Unit 2 was 84.9 percent during March 2023.

However, the facility did not consume coal that month and did consume a small amount of distillate fuel oil, resulting in a slightly lower average yearly SO

removal efficiency that is not representative of unit's normal operation.

See

89 FR 105528. Between 2017 to 2023, the yearly average FGD SO

control efficiencies for Seminole Unit 1 ranged from 96.6 to 97.0 percent and Unit 2 ranged from 96.5 to 97.3 percent.

See

EGU scrubber efficiency spreadsheet that is included in the docket for this rulemaking.

See

table 1.1 on page 1-3, section 5, chapter 1 of the CCM.

See

Exhibit 42 attached to the Conservation Groups' comments.

EPA disagrees with the comment that Florida did not adequately demonstrate that Unit 2 has existing, effective controls for SO

for the second planning period. Florida describes both Units 1 and 2 as subject to the MATS SO

limit of 0.20 lb/MMBtu on page 254 of the 2021 Plan and includes both units in table 7-27 of the 2021 Plan when comparing actual SO

emissions rates to this MATS limitation. Also, in the “Materials to be Incorporated into the SIP” section of the 2021 Plan, both Units 1 and 2 are listed as affected units for which the MATS SO

limit of 0.20 lb/MMBtu is proposed for incorporation into the SIP.

Comment 6.h:

The Conservation Groups argue that EPA's proposal to approve Florida's determination that TECO-Big Bend Unit 4 is effectively controlled for SO

and that no additional reasonable controls are likely to be identified because the SIP will incorporate the SO

MATS limit of 0.20 lb/MMBtu is insufficient. The Conservation Groups add that the State did not require a detailed FFA from TECO-Big Bend for Unit 4, or require the facility to provide supporting documentation to explain why it is effectively controlled for SO

emissions with wet scrubbers.

The Conservation Groups assert that it is difficult to determine the performance potential of the SCR and scrubber systems for TECO-Big Bend Unit 4 because it is permitted to consume multiple fuel types and periods of low SO

and NO

could reflect the partial use of natural gas. They also allege that the SCR system was not being used to its full capacity and is minimally operated to achieve its 0.10 lb of NO

/MMBtu emission limit. They therefore assert that EPA must require the State to conduct FFAs of SO

and NO

emissions and must independently review the analyses, fill in the gaps where necessary, and then establish practically enforceable emission limits.

The Conservation Groups explain that while EPA evaluated data from 2017 to 2023 and calculated that the existing FGD system had yearly average SO

removal efficiencies ranging between 92.2 to 97.1 percent during periods when coal is one of the fuel sources consumed, EPA failed to assess the results of its work and did not provide any opinion on how these values relate to an achievable optimized control efficiency of a modern scrubber system. They contend that an SO

FFA of TECO-Big Bend Unit 4 would almost certainly result in additional cost-effective control for SO

The Conservation Groups argue that an FFA is necessary to determine if the scrubber and SCR systems can be cost-effectively upgraded or optimized, the scrubber system is underperforming, and EPA cannot approve a SIP that refuses to conduct an FFA because Florida failed to explain why an FFA would result in a conclusion that no further controls are necessary.

Response 6.h:

Regarding arguments that Florida must evaluate NO

controls for TECO-Big Bend Unit 4, see Response 4. Regarding SO

EPA disagrees with

the assertion that Florida must conduct a full FFA of Unit 4 for SO

Scrubber systems are widely considered the best control technology for reducing SO

emissions, as they can achieve very high removal efficiencies, making them highly effective at capturing SO

from industrial flue gases.

The MATS Rule is a fairly recent CAA requirement with co-benefits for reducing SO

For the purpose of SO

control measures, an EGU that has add-on FGD and that meets the applicable alternative SO

is necessary to make reasonable progress.

Unit 3 at TECO-Big Bend was permanently retired from electric generation service on April 26, 2023, and therefore, Florida's demonstration of existing, effective controls is no longer relevant and no further action is required by EPA. The Retired Unit Exemption Form for TECO-Big Bend Unit 3 is included in the docket for this rulemaking. On December 12, 2024, FDEP provided a letter removing the units from the Florida regional haze plan because the unit is permanently retired. This letter is in the docket for this rulemaking.

See

section 5, chapter 1, of the CCM available at

https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-reports-and-guidance-air-pollution.

See

2019 Guidance at 23.

EPA disagrees that the Agency did not assess the results of its work. EPA's analysis confirms that Unit 4 is not uncontrolled or lightly controlled, is subject to the MATS Rule alternative SO

emission limit of 0.20 lb/MMBtu, and is equipped with WFGD that routinely achieves a high SO

control effectiveness.

As stated in the NPRM (89 FR 105528), EPA evaluated data from 2017-2023 for Unit 4 and calculated that the existing FGD system routinely achieves 92.2 to 97.1 percent yearly average SO

removal efficiencies during periods when coal is one of the fuel sources consumed, with a seven-year average (2017-2023) SO

removal efficiency of 95.8 percent.

Additionally, except for two months at the end of 2023, the unit routinely achieved 95 percent or great removal efficiency.

Regarding the comment that EPA failed to provide any opinion as to how the 92.2 to 97.1 percent yearly average SO

removal efficiencies relate to the optimized control level that a modern scrubber system is capable of achieving, EPA believes that it is unlikely that an FFA would result in the conclusion that further SO

emissions controls (including FGD upgrades) are necessary for reasonable progress. Therefore, EPA finds FDEP's effective controls determination for TECO-Big Bend Unit 4 to be reasonable.

See

EPA's EGU scrubber efficiency spreadsheet.

Id.

The EIA data reports Big Bend Unit 4 as burning bituminous coal with a sulfur content between 2 and 3.15 percent for all months in the 2017-2023 time period through October 2023. However, the Unit is listed as burning low-sulfur coal with a sulfur content of 0.3 to 0.6 percent in November and December of 2023. The reported low sulfur content caused the calculation of removal efficiency to be low for those months (73 to 85 percent), which lowered the annual average for 2023. Assuming the data is accurate, the SO

emissions for November and December 2023 were similar or lower than the SO

emissions from other months in 2023, and the emissions rate never exceeded 0.16 lb/MMBtu in any month in 2023.

EPA acknowledges that TECO-Big Bend can consume multiple fuel types; however, only coal and natural gas consumption were documented in the EIA data that is a part of the EGU scrubber efficiency spreadsheet between 2017-2023. EPA disagrees with the Conservation Groups' assertion that it is difficult to determine a scrubber's control efficiency if a unit is consuming multiple fuel sources. As indicated in the EGU scrubber efficiency spreadsheet included in the docket for this action, EPA determined a unit's control efficiency by calculating the sum of the uncontrolled tons of SO

emitted from each fuel source and comparing it to the measured SO

emitted after the controls (data from CAMPD). For example, if a unit consumed a mixture of coal and natural gas, the predicted uncontrolled tons of SO

emitted by the unit and CEMS SO

emissions values are only attributable to sulfur from the coal consumed. The contribution of SO

from natural gas to the predicted uncontrolled tons of SO

(by unit) would be nearly zero. Therefore, the partial consumption of natural gas does not meaningfully impact how the SO

scrubber efficiency was determined.

Comment 7:

The Conservation Groups assert that Florida unreasonably excluded several EGU and non-EGU sources “that EPA does not cover in its proposal” from FFA requirements. The Conservation Groups maintain that EPA wrongly proposes to approve the State's exclusion of Deerhaven Generating Station (Deerhaven) from the FFA requirement based on a fuel co-firing project that will allow the facility to burn up to 100 percent natural gas. They argue that the facility is not restricted to consume only natural gas and that it is capable of burning all natural gas, all coal, or a mixture of the two fuels. Hence, the Conservation Groups assert that a proper FFA must be conducted unless the SIP includes an enforceable commitment to burn only natural gas.

The Conservation Groups state that Breitburn, located 191 km from the Breton National Wilderness Area (Breton), is a significant source of haze pollution, emitting 778 tpy of SO

and 333 tpy of NO

in 2020. They maintain that the facility has a cumulative Q/d of 5.98; there are issues with Florida's determination that the facility is effectively controlled because that determination is based solely on the facility's distance from Breton; and VISTAS' projected 2028 decrease in emissions for this facility compared to more recent actual emissions was not explained. They also state that Florida did not meaningfully consider and respond to their 2021 comments to the State on Breitburn; EPA failed to consider Breitburn in its proposal; and this silence is arbitrary and capricious.

The Conservation Groups also assert that Florida failed to select the following 18 sources that “likely contribute to visibility impairment at in-state and out-of-state Class I areas” for FFAs: Orlando Utilities Commission Electric Generation facility; CEMEX Miami Cement Plant; Titan Florida Cement Plant; Department of Solid Waste Management, Miami-Dade; Rayonier Fernandina Plant; Hernando County CEMEX Plant; Florida Gas Transmission Company—Gadsden County; Mosaic Florida Phosphate Plant—Hillsborough County; Argos Facility—Alachua; Wheelabrator South Broward; Duke Energy—Pasco County; Florida Power & Light Company—Escambia County; International Paper Company—Escambia Mill; Pinellas County Landfill; Solid Waste Incinerator of Palm Beach; U.S. Sugar Corporation—Hendry County; Florida Power & Light—Lee County; and Sugar Cane Growers Co-op—Palm Beach County Mill. According to the Conservation Groups, each source has 1,000 tpy of either total combined emissions of SO

, particulate matter less than 10 micrometers (PM

), and NO

(based on 2020 NEI) or total combined NO

and SO

emissions (based on EPA's 2023 CAMPD), and nearly all of these sources emit more than 1,000 tpy of NO

alone. They also state that the Orlando Utilities Commission Electric Generation and CEMEX Miami Cement Plant facilities emit the second and third highest amounts of haze-generating pollutants in the State with Q/d values of 244.94 and 240.55,

respectively. Therefore, they argue that there are likely readily available, feasible, and cost-effective controls that can be implemented at the facilities. The Conservation Groups maintain that the State erred in not selecting these sources for FFAs and EPA failed to address them in its proposal. Thus, they argue that EPA must require Florida to conduct an FFA for each facility to ensure the State meets CAA requirements to make reasonable progress during the second planning period.

Response 7:

Regarding the 20 sources that the Conservation Groups contend must be evaluated by Florida, EPA disagrees that the Agency “must require Florida to conduct a Four-Factor Analysis of potential controls for each of these facilities to ensure that the State meets the Clean Air Act's requirements to make reasonable progress in the second planning period.”

See

Response 2, Response 5, and the NPRM (89 FR 105511) which explain that the RHR does not require states to select and consider controls for all sources, all source categories, or any or all sources in a particular source category. Nor does the RHR expressly specify criteria for minimum source selection thresholds. Florida has discretion under the RHR to determine its source selection methodology and Florida's source selection process, and the sources that Florida selected were reasonable and the Haze Plan complied with the CAA and RHR for this planning period. While Florida could have used its discretion to select other sources in addition to those screened in during its source selection process, including some or all of the sources that the Conservation Groups highlight, Florida was not required to do so. Also, sources that did not meet the State's reasonable source selection criteria (such as Deerhaven) were not selected for an FFA and were therefore not required to have emission limits and supporting conditions adopted into the LTS in the SIP to support reasonable progress for the second planning period. Regarding the assertion that Florida failed to meaningfully consider and respond to the Conservation Groups' comments concerning Breitburn,

see

Response 12.

Comment 8:

The Conservation Groups assert that EPA shirks its duty to review Florida's source-specific FFAs. They state that EPA proposes to “rubber stamp” the SIP submission without engaging in any meaningful and independent analysis of Florida's FFAs for the four facilities

100

to ensure they comply with the CAA and the RHR. Pointing to EPA's Technical Support Document (TSD), the Conservation Groups claim EPA merely restated what Florida did and that EPA entirely failed to grapple with the record before it and thus shirked its duties under the Act. They explain that EPA has stated in its 2021 Clarification Memo that it expects states to “undertake rigorous reasonable progress analyses that identify further opportunities to advance the national visibility goal.” They then assert that “[d]espite EPA's stated expectations for this planning period, in large part, Florida does not require any of the sources to adopt additional control measures to make reasonable progress” and that EPA accepts “Florida's decisions to ignore readily available, feasible, and cost-effective controls,” which they contend violates the CAA and RHR. The Conservation Groups' specific comments on the FFAs for Foley, JEA Northside, and WestRock-Fernandina are addressed in Comments 8.a through 8.c, below.

100

Initially, Florida completed FFAs for four facilities (Foley, JEA Northside Unit 3, WestRock-Fernandina, and WestRock-Panama City). As stated in the NPRM (

see

footnote 51 at 89 FR 105518), FDEP included documentation of the closure of WestRock-Panama City in the 2024 Supplement. In addition, on October 18, 2024, FDEP sent a site inspection report and other supporting documentation for the WestRock-Panama City closure as an addendum to the 2024 Supplement. Foley has also shut down as discussed in Response 8.a.

Response 8:

EPA's approval of Florida's Haze Plan is a proper exercise of EPA's authority under the CAA. Congress crafted the CAA intending for states to take the lead in developing implementation plans. However, Congress balanced that decision by requiring EPA to review the plans to determine whether a regional haze SIP revision meets the requirements of the CAA. When reviewing SIPs, EPA must consider not only whether the state considered the appropriate factors in making decisions, but acted reasonably in doing so. In undertaking such a review, EPA does not usurp the state's authority but ensures that such authority is reasonably exercised under the requirements of the CAA and RHR.

Contrary to the comment that the Agency “shirks” its CAA obligations, EPA has performed its duties with diligence. EPA carefully evaluated the Haze Plan and the associated record and engaged in a thorough analysis of each control option, including each of the underlying cost assumptions used in the calculations. Florida conducted extensive technical work in support of its SIP submittal, and EPA independently evaluated each FFA, including costs, and compared each FFA's control determination against EPA's CCM. In the TSD to the NPRM, EPA documented the cost assumptions that the State relied upon in its FFAs for transparency to the public.

101

Each of the FFAs are discussed in more detail in the responses to comments that follow.

101

The NPRM contains several erroneous cross-references to the TSD. On page 105516, the cross-reference to section I.F. should be to section I.E.; on page 105524, the cross-reference to section I.E. should be to section I.D.; on page 105524, the cross-reference to section I.D. should be to section I.C. and the cross-reference to Section I.F. should be to section I.E.; and on page 105529, the cross-reference to section I.F. should be to section I.E.

Comment 8.a:

The Conservation Groups contend that Foley, located 43 km from St. Marks, contributes a significant amount of SO

and NO

(emitting 2,087 tpy and 1,596 tpy of each pollutant in 2020, respectively) that likely impacts 15 Class I areas. Furthermore, they maintain that the facility has an “extremely high” cumulative Q/d of 288.37. Florida selected Foley for an emissions control analysis, and the facility conducted an FFA at the request of the State. The Conservation Groups assert that there were many significant technical issues with the FFA outlined in the 2021, 2024, and 2025 Kordzi Reports. They provide summaries intended to identify issues from these reports in table 1 of their comments on pages 39 through 41 of their Comment Letter.

The Conservation Groups also assert that neither Florida nor EPA indicate that the shutdown of Foley has been made federally enforceable through inclusion in the SIP and that EPA must therefore require that Florida make the Foley shutdown federally enforceable. Alternatively (

i.e.,

if the shutdown has not occurred and is not enforceable), the Conservation Groups provide specific comments regarding Foley's FFA. Because the facility has shut down, the Conservation Groups' specific comments on the Foley FFA are not reproduced here.

Response 8.a:

Regarding Foley's visibility impacts to Class I areas, as discussed in Response 2, Florida considered Q/d as part of its AoI analysis and then further conducted PSAT modeling, which differs from the method used by the Conservation Groups. EPA finds Florida's source selection methodology and its results reasonable as explained in Response 2.

Regarding the comments on Foley's FFA, these comments are no longer relevant because Foley has fully shut down, is incapable of restarting without undergoing applicable New Source Review permitting for new sources, and Florida has withdrawn the emission limits resulting from the Foley FFA. In

a letter dated May 22, 2024, Foley indicated that its mill “ceased production operations as a pulp and paper mill and will no longer operate the [] equipment that was included in the regional haze analysis.”

102

Foley also indicated in its letter that this equipment would “be air-gapped by December 31, 2024.” In a letter dated April 8, 2025, Florida notified EPA that Foley has permanently ceased operation of all emissions units. In its letter, Florida confirmed that these units “have been rendered inoperable and permanently retired,” and therefore, as of April 7, 2025, Florida revoked the title V Air Operation Permit (Permit No. 1230001-127-AV) for Foley.

103

In the letter, Florida also notes that “[t]he effect of this expiration is that all existing emissions units at Foley are considered retired and if Georgia Pacific (or any successor owner) were to restart these units, they would be treated as new units and subject to the Department's New Source Review program.” Further, in Florida's April 8, 2025, letter to EPA, the State withdrew its request to incorporate permit conditions for Foley into the SIP.

102

See

appendix C-2 to 2024 Supplement.

103

FDEP's April 8, 2025, letter to EPA containing Florida's Notice of Administratively Corrected Title V Air Operation Permit that administratively expires the facility's current Title V Air Operation Permit on April 7, 2025, is included in the docket for this action.

Regarding the assertion that this shutdown must be made federally enforceable by placing the shutdown into the SIP, EPA disagrees for the reasons discussed immediately above. Furthermore, Florida's New Source Review permitting requirements are in Florida's SIP and are therefore federally enforceable.

104

See

Rules 62-210.300 (General Preconstruction Review Requirements), 62-210.400 (Prevention of Significant Deterioration permitting), 62-210.500 (Preconstruction Review for Nonattainment Areas), and 40 CFR 52.520(c) (incorporating by reference Rules 62-210.300, 62-210.400, and 62-210.500 into Florida's SIP).

Comment 8.b:

The Conservation Groups contend that Florida conducted an inadequate FFA for Unit 3 at JEA Northside, a power plant located in North Jacksonville. Unit 3 is an EGU that consumes natural gas and a limited amount of fuel oil, but does not have add-on SO

controls. The FFA led Florida to make the determination that switching to lower sulfur No. 6 fuel oil would be cost-effective. The Conservation Groups highlight Florida's statement that “[g]iven that JEA can timely implement a fuel switch and there are no energy or non-air environmental impacts, [FDEP] has determined that switching to lower sulfur No. 6 fuel oil is necessary for achieving reasonable progress.”

105

However, the Conservation Groups assert that EPA, in its Proposed Rule, failed to meaningfully review the FFA for JEA Northside Unit 3 since it does not consider elimination of fuel oil altogether. Furthermore, they argue that EPA's analysis of Unit 3 “rubber stamps” the State's determination that prohibition of fuel oils with greater than 1.0 percent sulfur content is a measure necessary for reasonable progress. They claim that EPA condones Florida's conclusion that switching to fuel oil that is 1.0 percent or lower sulfur content is the most cost-effective control option. Instead, they maintain that EPA should employ a similar analytic approach across all sources (

e.g.,

the four boilers at WestRock-Fernandina) and evaluate a switch to No. 6 fuel oil with a 0.5 percent sulfur content. The Conservation Groups assert that EPA cannot approve a SIP that fails to evaluate eliminating fuel oil use or converting to ultra-low sulfur diesel (ULSD).

105

Citing the 2021 Plan at 268.

Response 8.b:

The Conservation Groups state that the “elimination of fuel oil altogether” or a conversion to ULSD were not assessed as potential measures for JEA Northside Generating Station. However, the JEA Northside FFA presented in appendix G-2 of Florida's October 8, 2021, SIP revision includes an assessment of fuel switching the Unit 3 to ULSD. The same analysis stated that fuel oil usage in unit 3 is extremely limited, as the unit meets the definition of a natural gas-fired electric utility steam generating unit, as defined in 40 CFR 63.10042, based on its limited use of oil. The JEA Northside FFA showed that from 2015 to 2019, Unit 3 fired fuel oil for a maximum of only 1.35 percent of the total annual heat input and a minimum of 0.03 percent. As Unit 3 is already almost entirely fueled by natural gas, the complete elimination of fuel oil was not selected as a potential SO

control as removing the already extremely limited use of fuel oil would result in negligible or little improvement in emissions and, therefore, visibility. Regarding the lack of an evaluation of a fuel switch to No. 6 fuel oil with a 0.5 percent sulfur content, the 2019 Guidance provides that “[a] state must reasonably pick and justify the measures that it will consider, recognizing that there is no statutory or regulatory requirement to consider all technically feasible measures or any particular measures. A range of technically feasible measures available to reduce emissions would be one way to justify a reasonable set.” As Florida has provided a range of technically feasible measures—including considering ULSD and selecting as a control requiring No. 6 fuel oil that does not exceed 1.0 percent sulfur—an evaluation of No. 6 fuel oil with a 0.5 percent sulfur content is not required.

Comment 8.c:

The Conservation Groups identify WestRock-Fernandina's FFA as flawed. First, they explain that WestRock-Fernandina is located 64 km from Okefenokee, and that according to the NPCA's 2020 NEI data analysis, which showed it emitted 633 tpy of SO

and 1,231 tpy of NO

, it is a significant source of haze-forming pollution that likely contributes to visibility impairment at five Class I areas. They note the same analysis shows the facility has a “very high” cumulative Q/d of 77.51.

The Conservation Groups summarize issues with the WestRock-Fernandina FFA identified in the 2025 Kordzi Report as well as their comment letters to Florida. They claim that Florida failed to meaningfully engage with and respond to their comments and that EPA's proposal rubber stamps WestRock-Fernandina's FFA. Specifically, they state that EPA proposes to find that Florida's determinations regarding the applicable controls for the sources at WestRock-Fernandina are reasonable despite Florida's failure to evaluate available and technically feasible SO

controls based on, where applicable, estimated values of capital costs, annualized costs, and cost per ton of emission reductions, consistent with recommendations in the CCM.

The Conservation Groups point to a general statement in the 2025 Kordzi Report that WestRock-Fernandina failed to provide adequate documentation for many claims in its analysis, including cost figures.

106

Specifically, they say that WestRock-Fernandina claims that the No. 7 Power Boiler is capable of running on 100 percent natural gas, but counter that WestRock did not explore the total conversion of this boiler to natural gas. The Conservation Groups claim that WestRock-Fernandina's claim that such a conversion would be a “fundamental change

This text is long and has been trimmed here. Open the source document for the complete record.

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.