Our Recent Posts



New Guidance on Ozone and Fine Particulate Modeling

On July 29, 2022, EPA’s Air Quality Assessment Division and Air Quality Policy Division jointly issued new guidance, Guidance for Ozone and Fine Particulate Matter Permit Modeling, addressing how in final form to state, local and tribal permitting authorities and the public. The guidance sets forth “EPA’s recommendations for how a stationary source seeking” a PSD permit “may demonstrate that it will not cause or contribute to a violation” of the Ozone NAAQS, the PM2.5 NAAQS or the PM2.5 increments. Unfortunately, the Final Guidance omits discussion of some common occurrences, which may lead to delays as applicants, permitting authorities and EPA Regions may take different views in the absence of more general guidance. In some cases, the Final Guidance also uses loose language that could result in the Final Guidance being applied in a way inconsistent with the applicable PSD and modeling rules. Thus, permitting authorities and sources will need to approach the Final Guidance with some caution as uncritical reliance on its text may lead to legal deficiencies in the final permit.


EPA begins the Final Guidance by noting that in the 2017 Guideline on Air Quality Modeling EPA had created a two-tiered demonstration approach for “single source impacts” on ozone and secondary PM2.5. EPA thereafter issued draft guidance on February 10, 2020 and then revised draft guidance on September 20, 2021. In the February 2020 draft guidance, EPA recommended that compliance demonstrations only include the direct emissions or precursors that exceeded the significant emission rate. The September 2021 revised guidance required that “sources should provide a full accounting of the combined impacts of their allowable precursor (and direct component, in the case of PM2.5) emissions on ambient concentrations of the relevant NAAQS.” EPA contends that “this holistic approach to the PSD compliance demonstration for O3 and PM2.5 is supported both scientifically and legally.” EPA announced that “we are maintaining the holistic compliance demonstration approach and all other recommendations from the revised draft guidance. We have a made a few clarifications and associated updates.” Finally, EPA states that it is fully replacing both the February 2020 and September 2021 guidance documents.

New elements in the Final Guidance that NSR Law Blog does not remember at the same level of specificity in earlier iterations include:

  • For Ozone secondary contributions, the Final Guidance states that design values alone should not be used but that the “characterization should take into consideration episodic high O3 concentrations and any trends that may be occurring.” This could be read as opening the door to arguments that long-term climatic effects should be considered in setting source impacts for today. The Final Guidance does not provide direction on how such trends should be considered or sideboards on how far “trends” may allow the disregarding of present conditions.

  • For PM2.5 secondary contributions, the Final Guidance similarly states that “the characterization should take into consideration the seasonality and speciated composition of current PM2.5 concentrations and any long-term trends that may be occurring” and that “it may also be important to describe the typical background concentrations of certain chemical species that participate in the photochemical reactions that form O3 and secondary PM2.5.” The Final Guidance does note that “it is possible that there are mitigating factors for secondary PM2.5 formation given limitations of other chemical species important in the photochemical reactions, e.g., minimal ammonia (NH3) in the ambient environment….” Again, the Final Guidance does not provide any direction on how seasonality should be considered – does the Final Guidance contemplate abandonment of the annual average for the highest seasonal average? It is impossible to tell.

  • EPA notes that it does not agree with the NACAA approach of converting secondary PM2.5 to equivalent direct PM2.5 that may allow use of AERMOD to assist with assessing secondary formation. See Final Guidance at 32-33. EPA also states that it agrees with NACAA’s rejection of a simple Q/D metric and 100 percent conversion of SO2 and NOx concentrations to their particulate counterparts.

The Final Guidance does provide some useful specifics on how compliance demonstrations should be made if it is necessary to go to “Tier 2” (e.g., chemical transport modeling). Highlights include:

  • Ozone. Compare “the highest of the multi-season (or episodic) averages of the maximum modeled daily 8-hour O3 concentrations predicted each season (or episode) at each grid cell or location … to the appropriate O3 SIL, since [sic] this metric represents the maximum potential daily 8-hour O3 impact from the proposed source or modification. For Tier 1, EPA states that the MERP guidance is acceptable. For Tier 2, if a chemical transport model (CTM) is used, then each cell must be less than the SIL. Final Guidance at III.5.1.

  • PM2.5 Direct Emissions. The Final Guidance uses a variable approach depending upon the meteorological data used to support the primary AERMOD “direct” PM2.5 emissions run:

  • If 5 years of National Weather Service data are used, then “the highest of the 5-year averages of the maximum modeled annual 24-hour PM2.5 concentrations (for the 24-hour PM2.5 NAAQS) or highest of the 5-year averages of the annual average PM2.5 concentration (for the annual PM2.5 NAAQS) predicted each year at each receptor.

  • If 3 years of prognostic data are used, then “the highest of the 3-year averages of the maximum modeled annual 24-hour PM2.5 concentrations (for the 24-hour PM2.5 NAAQS) or highest of the 3-year averages of the annual average PM2.5 concentration (for the annual PM2.5 NAAQS) predicted each year at each receptor.

  • If on-site data are used, then then “the highest of the X-year averages of the maximum modeled annual 24-hour PM2.5 concentrations (for the 24-hour PM2.5 NAAQS) or highest of the X-year averages of the annual average PM2.5 concentration (for the annual PM2.5 NAAQS) predicted each year at each receptor, where X equals the number of years of site-specific data used.

  • PM2.5 Combined Emissions. The Final Guidance states that the maximum direct emissions may be combined with the maximum precursor emissions as a “conservative estimate” of combined impacts “since [sic[ as noted above, peak impacts associated with a source’s direct PM2.5 and precursor emissions are not likely well-correlated in time or space.” Final Guidance at 41. EPA cautions that full temporal pairing may not be appropriate given the different data sets/dates that may be required and differences in grid models versus AERMOD “receptor” locations. As a result, EPA recommends “seasonal or monthly basis with consideration of spatial pairing that reflects the general lack of correlation between primary and secondary impacts i.e., primary impacts being higher near the source while secondary impacts being higher at some distance away from the source.” Final Guidance at 42.

  • Ozone cumulative modeling. EPA recommends use of background concentrations as the basis for considering “nearby” sources. Final Guidance at 43-44. The MERPs or modeled value is then compared to the background design value to determine whether the NAAQS is exceeded. Final Guidance at 49. If Tier 2 done, then compare fourth high to design value. Episodic requires consultation. Final Guidance at 49.

  • PM2.5 cumulative modeling. While not explicit, the Final Guidance appears to recommend use of AERMOD modeling of nearby sources’ primary PM2.5 emissions if they would be expected to result in a concentration gradient in the area affected by the modeled source and use of background concentration otherwise. For precursors, EPA generally recommends the use of background concentrations. Final Guidance at 44-45. For comparison to the PM2.5 NAAQS, EPA recommends use of the “PM2.5 design value appropriate to the area” rather than the overall maximum monitored background. The direct PM2.5 concentration (from AERMOD) and the secondary (from MERPs or CTM) are then added to the design value using the modeled 98th percentile value (for the 24-hour NAAQS) or the annual average value (for the annual NAAQS) based on the 5/3/X years of AERMOD meteorological data used (see discussion above for the 5/3/X). Final Guidance at 51-52. The Final Guidance suggests consideration of seasonal or monthly background but not more granular analysis. Final Guidance at 54.

  • PM2.5 increment modeling is mostly similar to past practice but difficult to summarize. Readers are directed to the Final Guidance Section V.3 for details.

The Final Guidance represents a substantial effort by EPA to address the challenges of considering the impact of precursors. There are generally well thought out approaches for typical situations and a welcome recognition that the impacts of the primary and secondary emissions of PM2.5 are poorly correlated in time and space, which should provide permitting authorities some flexibility in addressing situations where a source just barely exceeds a SIL or NAAQS.


The Final Guidance is not an unexpected outcome. It was hoped, however, that it might provide guidance to address situations, common to modifications, that are presently not well addressed. Two examples come to mind:

  • Project emissions decreases. In some modifications, emissions of a direct or precursor pollutant may decrease. The Final Guidance is silent on how to address this situation and leaves it to permitting authorities to decide how to proceed. In NSR Law Blog’s view, in Project Impact Analysis, it would be appropriate to treat the emissions decrease as “negative” values for purposes of AERMOD modeling or MERPs analysis and then to combine the positive and negative emissions as the Final Guidance states to determine whether the major modification warrants a cumulative modeling analysis. Such an approach would recognize the ambient benefits of emissions reductions. A more explicit statement addressing the handling of emissions reductions would have been helpful.

  • "Affected sources." As the NAAQS and increments become tighter, it would be useful if EPA clarified that for unmodified units “affected” by a modification elsewhere (e.g., those units that may see an increase in utilization), the increase that should be modeled is the expected increase and not the “allowables.” Current guidance is vague or suggests going to potential emissions or allowables, which is inappropriate if the new or modified unit could not require such an increase. An example of such an inappropriate use of allowable emissions would be requiring a diesel water pump that runs a few hours a day supplying a large water tank to be modeled as running 24 hours a day because it is “affected” by the installation of a new unit that requires a few gallons of water an hour. Only the emissions from the additional time needed to meet the demand should be required.

NSR Law Blog hopes that EPA will address these situations in a future iteration of the guidance.

More troubling, however, is loose wording in several areas of the Final Guidance that could be read inconsistently with the underlying regulations and Part 51, Appendix W. Areas of such loose language identified by NSR Law Blog include the following:

  • Final Guidance II.4, at 18. The Final Guidance states “”instead, for major modifications, the definition of “actual emissions” at 40 CFR 52.21(b)(21) continues to apply and post-project emissions should be based on potential to emit or allowable emissions.16” This statement is correct only if the underlying unit has been modified, which often will not be the case because sources of direct PM2.5 and the precursors NOx and SO2 may be different. Fortunately, footnote 16 gives the correct guidance that “actual emissions” should be used if the underlying unit is not modified or affected by the project.

  • Final Guidance II.5.1, footnote 18. The footnote states that a new major source must model “all direct and precursor pollutants if the source has the potential to emit any direct or precursor pollutant in an amount greater than or equal to the SER.” This is incorrect and should read “all direct and precursor pollutants of a pollutant if the source has the potential to emit any direct or precursors of that pollutant in an amount greater than or equal to the SER.” Similarly, for major modifications, the footnote should read “for a major modification to an existing major stationary source, it includes all direct and precursor pollutants of a pollutant, if the modification would result in a significant emissions increase and a significant net emissions increase of any director or precursor pollutant of that pollutant.”

  • Final Guidance III.1, Table III-1 “EPA Recommended Approaches for Assessing O3 Impacts by Assessment Case.” Table III-1 presents the assessment case for a new major stationary source. It does not present the case for a major modification. In the case of a major modification, the question is whether there is both an emissions increase and a significant net emissions increase of either NOx or VOC emissions that exceeds the 40 tpy SER. The Final Guidance should have said “increase” and not “emissions” as all emissions at a new source are an “increase.”

  • Final Guidance III.2, Table III-2 “EPA Recommended Approaches for Assessment Primary and Secondary PM2.5 Impacts by Assessment Case.” Like Table III-1, Table III-2 presents the assessment case for a new major stationary source. It omits the term “both an emissions increase and a significant net increase” in excess of the relevant SER in determining whether a major modification of a major stationary source would be subject to further analysis. Again, it would have been clearer to simply say “increase” rather than “emissions.”

  • Final Guidance IV.1 Modeling Inventory. The Final Guidance states that for a “new or modifying source” the “maximum allowable emissions rate” is used. This is true for a new source but only partially true for a modifying source. Appendix W, Table 8-2, “Point Source Model Emission Inputs for NAAQS Compliance in PSD Demonstrations,” states that a “Nearby Source” includes “existing facility to which modification is proposed if the emissions from the existing facility will not be affected by the modification. Otherwise use the same parameters as the major modification.” 40 CFR Part 51, App. W, Table 8-2, note 4. For major modifications, the “new or modified emissions units” must use the maximum allowable emission rates. Unaffected units should use the Nearby Source rates.

  • Final Guidance, Table V-2. Again, the table should use “increase” rather than “emissions” in its “Description of Assessment Case.”

  • Final Guidance, V.3.2. Inset bullet 1 states that “direct and precursor allowable emissions from the proposed new or modifying source” should be used but this is corrected in Inset bullet 2, which notes that “Direct and precursor actual emissions changes that have occurred to existing sources (including the existing source at which a major modification is being proposed, where applicable)…. NSR Law appreciates EPA including the underlined language in bullet #2.

In summary, the Final Guidance presents a useful implementation tool for applicants and permitting authorities to use in assessing primary and secondary emissions impacts for ozone and PM2.5. The Final Guidance misses some opportunities to provide useful direction on project emissions reductions and “affected” sources. Users are cautioned that they must think critically about whether they are applying the Final Guidance to a “new source,” where it mostly applies as written, or to a “major modification” where it applies to the increase in emissions for applicability and project impact analysis and then to a mix of allowable and actual emissions for cumulative modeling. In the major modification context, only emissions from new and modified units are modeled at allowables and emissions from unaffected units are modeled as set forth in Appendix W, Table 8-2. Lack of attention to these details could result in a legally deficient analysis.