Written by Jonathan B. Hill, CM, Managing Consultant - Raleigh, NC, George J. Schewe, CCM QEP, Principal Consultant - Covington, KY, Tony Schroeder, Managing Consultant - Indianapolis, IN
Originally published by Trinity Consultants on February 14, 2017.
On December 20, 2016, the U.S. EPA Administrator signed a final rule entitled Revisions to the Guideline on Air Quality Models: Enhancements to the AERMOD Dispersion Modeling System and Incorporation of Approaches to Address Ozone and Fine Particulate Matter. The rule provides EPA-recommended models and other techniques as well as guidance for their use in predicting ambient concentrations of air pollutants. EPA's finalized changes enhance the formulation and application of the agency's AERMOD dispersion model, prescribe modeling techniques for secondarily formed fine particle and ozone pollution for single sources, and make various editorial improvements. The Federal Register version of the Final Rule is available here. The rule, referred to hereafter as the 2017 Guideline, was published on January 17, 2017, and takes effect on March 21, 2017.
National and state air quality regulations rely upon this guidance from EPA for performing dispersion modeling studies as part of routine air permitting and other regulatory air quality studies. The 2017 Guideline will now direct application of dispersion models for major permit actions (e.g., new and modified PSD permits). Some states do not have any additional state-specific modeling guidance and interpret the guidelines on an as-needed basis for projects; other states will have their own interpretations and specific guidance. Some of the new guidance includes modelingof actual emissions instead of allowable emissions for regional inventory sources, clarification regarding the definition of background concentrations, use of adjusted turbulence (which will lower concentrations in most cases), low wind speed adjustments (may still be beta), and alternate NO2/NOx conversions.
This 2017 version of the Guideline replaces the previous version published by EPA in 2005, but continues to provide primary guidance to EPA regional offices and state/local environmental regulatory agencies for the dispersion modeling component of industrial air quality permitting, State Implementation Plan (SIP) development, and many other applications. In 2010, EPA promulgated new 1-hour probabilistic National Ambient Air Quality Standards (NAAQS) for both NO2 and SO2 that the approved 2005 Guideline models were not equipped to handle. Although interim memoranda on EPA's SCRAM website as well as needed updates were made available over the past 12 years, it was time for a comprehensive update of the Guideline and for the codification of new features of AERMOD.
Summary of the 2017 Guideline
The 2017 Guideline is organized into sections providing model overviews, recommendations, and requirements, and sections addressing the use of meteorological data sets, actual versus allowable emissions data, model accuracy, and model options. The Introduction describes its purpose, its regulatory applicability, effective date (30 days after publication in FR but delayed until March 21, 2017 by Executive Order of President), and regulatory integration requirement of one year (except for transportation conformity, which has a three-year transition period). Excerpts and summaries of finalized components of the 2017 Guideline are provided below.
The 2017 Guideline includes revisions to AERMOD,
EPA's preferred model for refined industrial modeling.
The Preface retains the discussion on the importance and need for the Guideline:
Industry and control agencies have long expressed a need for consistency in the application of air quality models for regulatory purposes. In the 1977 Clean Air Act (CAA), Congress mandated such consistency and encouraged the standardization of model applications. The Guideline on Air Quality Models (hereafter, Guideline) was first published in April 1978 to satisfy these requirements by specifying models and providing guidance for their use. The Guideline provides a common basis for estimating the air quality concentrations of criteria pollutants used in assessing control strategies and developing emissions limits.
EPA has been steadfast in its commitment to this mandate and is providing the modeling community with this next generation of guidance in an effort to improve model performance, consider new options, and meet the needs of implementing the many forms of the NAAQS.
The Introduction describes the purpose and applicability of the Guideline and remains essentially the same as in the 2005 version of the Guideline but reflecting the reorganized nature of the 2017 Guideline.
2. Overview of Model Use
Section 2 was updated to clarify the modeling process flow and provide consistent definitions of commonly used terms such as screening technique, screening model, and refined model. Model performance evaluation is described within the context of accuracy and model uncertainties as well as the suitability of a model for a particular application.
3. Preferred and Alternative Air Quality Models
The requirements ("rules of the game") for model use and approvals are detailed in Section 3. This section also references Appendix A of the Guideline, which lists "preferred" models that can be used in dispersion modeling application without further justification. Modeling tools developed for specific source types or by private companies may be proposed for use as "alternative" models, provided the approval criteria delineated in the Guideline can be met. The last portion of Section 3 includes the most substantive change in this section—codifying the role of the Modeling Clearinghouse in the model approval process, in regulatory interpretation, and in other case-specific modeling issues that may arise.
4. Models for Carbon Monoxide, Lead, Sulfur Dioxide, Nitrogen Dioxide and Primary Particulate Matter
Section 4 describes the screening and refined models approved for use to model inert criteria pollutants. The approved refined model remains AERMOD. AERSCREEN is formally introduced and codified as the industrial-source screening model for single sources. A revised "multi-tiered" approach for NO2 is presented, as outlined below.
- Tier 1 still assumes that all emitted NOx converts to NO2.
- Tier 2 (previously known as the Ambient Ratio Method [ARM]) is now replaced with ARM2 in AERMOD, which uses concentration-specific ambient ratios based on nationwide monitoring data, rather than the single, fixed value used previously. ARM2 is now a regulatory default. The minimum ambient ratio is now 0.5 (raised from 0.2), making the new regulatory default of ARM2 more conservative than the non-regulatory ARM2 in the previous version of AERMOD.
- Tier 3 screening techniques—including OLM and PVMRM2 (PVMRM2 is a revised PVMRM formulation in AERMOD and simply replaces it), which provide a more detailed treatment of NOx to NO2 atmospheric chemistry in the presence of ozone—are now regulatory default options in AERMOD. Therefore, applicants intending to use these techniques no longer need to seek case-by-case approval from U.S. EPA, but rather need make only minimal additional approval efforts beyond modeling protocol approval.
This section also confirms that AERMOD should be used for primary PM2.5 modeling; secondary impacts are addressed in later sections.
Mobile Sources, Buoyant Plumes, and Long-Range Transport
Although CALINE3 and related models have been replaced by AERMOD as the preferred model for mobile-source modeling, EPA recognizes that transportation emissions modeling (CO, PM2.5, and PM10) can be accomplished in a number of ways and is thus allowing continued use of the CALINE models over the next three years. The Buoyant Line Plume (BLP) model is integrated into AERMOD (BLP is removed from Appendix A). CALPUFF has been removed as the preferred model for long-range transport and is now classified as a "screening technique." This section allows for the use of other Lagrangian or photochemical models to perform refined long-range transport analyses and discusses use of CALPUFF or other Lagrangian models for long-range screening analysis.
5. Ozone and Secondary Particulate Matter
Section 5 introduces a completely new subject to the Guideline and represents the most substantive revision to the Guideline. This section, in conjunction with EPA-authored technical guidance documents referenced within the Guideline, provides more specific language on how individual facility impacts on atmospheric ozone and secondary particulate formation might be addressed. No specific model or modeling technique is required or recommended in the 2017 Guideline, in order to allow applicants flexibility in how to address these impacts. The section introduces the concept of "Modeling Emission Rate for Precursors" (MERPs) as a way to determine if secondary impacts should be addressed through refined modeling or if they would "screen out," with additional detail provided in a separate technical guidance document entitled Guidance on the Development of Modeled Emission Rates for Precursors (MERPs) as a Tier 1 Demonstration Tool for Ozone and PM2.5 under the PSD Permitting Program. The MERPs technical guidance document provides examples of techniques that can be used to determine emission rates that would indicate negligible impact. The example values provided by EPA are considerably greater than the PSD Significant Emission Rates with MERPs for NOx as a precursor to PM2.5 ranging from 1,155 tons per year (tpy) to 10,037 tpy depending on the region of the U.S. and the PM2.5 standard averaging period considered. Although EPA had previously stated that the MERPs would be proposed in a separate future rulemaking, it subsequently chose to propose a technique to define MERPs on a case-by-case basis rather than to define specific values through a rulemaking. Projected emissions in excess of the MERPs will trigger additional requirements to address secondary formation through a two-tiered approach. That approach consists of a semi-quantitative, first-tier process based on historical photochemical modeling data, monitoring data, and other documented information, and a second tier based on full photochemical modeling analyses.
6. Modeling for Air Quality Related Values and Other Governmental Programs
Section 6 addresses visibility modeling and analysis to determine the impacts on Air Quality Related Values (AQRVs). AQRVs can include nitrate and sulfate deposition on surfaces as well as visibility degradation at Class I areas (national parks, wildlife refuges, monuments, etc.) in the form of plume blight and extinction coefficients. This is the first time that EPA has consolidated this guidance into the Guideline—previously, most of the AQRV guidance has been in the form of reports and guidelines from the Federal Land Managers (e.g., FLAG 2010 guidance) in association with EPA. One significant change related to this section is the removal of CALPUFF as the preferred long-range transport model in Sections 3 and 4; however, EPA notes that the change to the Guideline does not affect CALPUFF's use under the FLM's guidance regarding AQRV assessments (FLAG 2010). Section 6 specifically defers to the most current FLM guidance for AQRV analysis, but goes on to provide guidance on visibility and deposition modeling that emphasizes the use of photochemical models.
7. General Modeling Considerations
Section 7 is designed to capture specific modeling details that are not otherwise included in the remainder of the document. Much of the basic information in this section was included in the previous version of the Guideline and includes topics such as dispersion coefficients, complex winds, gravitational settling and deposition, and plume rise formulations. This section also discusses mobile source modeling, namely how to use AERMOD to appropriately characterize and model roadway emissions. Additionally, EPA removed outdated references and reworded many of the subsections to make the Guideline more of a high-level, general modeling guidance document.
8. Model Input Data
Section 8 expands on the previous discussion regarding use of source input data for the models, including the consideration of actual and allowable emissions, regional inventory requirements, and source parameterization. It also includes an overview of the use of background concentrations, tying in all current guidance to reflect both the old NAAQS forms of the deterministic standards as well as the new probabilistic forms of the new one-hour SO2 and NO2 standards. The discussion on inventory sources and background concentrations reflects EPA's intent to discontinue outdated and generally overly conservative approaches that have been relied upon by applicants in the past. In particular, EPA discusses the use of emission rates that better represent actual operating scenarios for nearby sources in permit application NAAQS modeling analyses, rather than the traditional method of using potential to emit for all sources in a NAAQS assessment. Finally, Section 8 provides the meteorological input requirements for modeling for National Weather Service (NWS) data and onsite data collection. EPA has also developed the Mesoscale Model Interface (MMIF) program that processes prognostic weather data (e.g., MM5 or WRF models) into formats compatible with AERMET and AERMOD, allowing for the use of those forecast model outputs in situations where representative NWS or onsite data are unavailable. The use of the MMIF program is currently required for applications that propose the use of prognostic meteorological data.
9. Regulatory Application of Models
Section 9 remains focused on recommendations for the modeling requirements, development of modeling protocols, use of measured (monitored) air quality data, and emissions limits and their relationship to modeled concentrations. This section also discusses the concept of design concentrations, NAAQS-based limits, and PSD increment-based limits. Previously, Section 9 had also addressed the issues of model accuracy and uncertainties, but EPA deleted those discussions in the 2017 Guideline.
Appendix A – Summaries of Preferred Air Quality Models
Key features of preferred models for regulatory modeling applications are provided in Appendix A, along with information on data inputs and outputs, key details on atmospheric physics treatments, and model availability. The 2017 preferred models, which can be used without a formal demonstration of applicability provided Guideline procedures are followed, are AERMOD, CTDMPLUS, and OCD.
Status of Major Proposed Revisions to AERMOD Formulation
In the preamble to the final rule, EPA requested comments on several proposed revisions to AERMOD. Several of these updates, such as incorporation of a BLP source type and refinements to NOx chemistry screening techniques, are discussed above in the summary of Section 4; the other major proposed revision to AERMOD involves the treatment of turbulence under low wind speed conditions. EPA originally proposed in 2015 that the Beta ADJ_U* option and the Beta LOWWIND3 option be upgraded to default status as part of the proposed rulemaking. In the final 2017 Guideline, however, EPA incorporated a slightly modified version of the ADJ_U* routine as a default option (with lower turbulence). However, based on public comments about possible underprediction of concentrations during certain situations, the LOWWIND3 option remains a Beta option and EPA is recommending further study and evaluation by the modeling community. Additionally, in the 2017 Guideline EPA clarified that the ADJ_U* option cannot be used as a regulatory default option if detailed turbulence data are available from a non-NWS dataset, again due to the possibility raised by commenters that AERMOD could underpredict concentrations in this configuration. A revised version (v16216) of AERMOD and the AERMET meteorological data pre-processor, incorporating these revised algorithms, was released by EPA concurrently with the release of the 2017 Guideline for use by the modeling community. And, just in the past three weeks since its release, EPA has updated the downloadable version on its website correcting a few minor bugs (v16216r). (Warning: Users should be certain to download the most current version from SCRAM.)
In summary, this significant revision to U.S. EPA's primary modeling guidance document is intended to formalize current guidance and provide insight and instruction on leading-edge modeling intricacies. Environmental professionals with a focus on air quality assessment will find the publication, along with the related technical guidance documents, required reading.
of Changes in the 2017 Modeling Guidance
replaced by ARM2 as the regulatory Tier 2 NO
- OLM and
a revised version of PVMRM made regulatory default Tier 3 NO2 screening options
- Low wind
speed turbulence adjustment (ADJ_U*) algorithm
in AERMET made regulatory default option
- Proposed LOWWlND3 option, intended to address
issues with AERMOD overprediction under
low wind speed conditions, not finalized as a
regulatory default option in AERMOD
- Use of
prognostic meteorological data, processed through
EPA's MMIF pre-processor, finalized as an option
for regulatory modeling applications
finalized on methodologies to address the
secondary chemical formation Of ozone and PM2.5 associated with precursor emissions (i.e., NOx and VOC
for ozone and NOx and SO2, for PM2.5) from
removed from list of EPA "preferred" models
for long-range transport assessments, but still
available for use in screening analyses with other Lagrangian models, such as SCICHEM
finalized to recommended modeling procedures
for cumulative impact analyses, including use Of
actual emissions for nearby sources and definition
of background concentrations