Environmental Protection Agency: Vol. 80 Wednesday, No. 145 July 29, 2015

Vol.
80 Wednesday,
No. 145 July 29, 2015
Part III
Environmental Protection Agency

40 CFR Part 51
Revision 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; Proposed Rule
mstockstill on DSK4VPTVN1PROD with PROPOSALS2
VerDate Sep<11>2014 19:11 Jul 28, 2015 Jkt 235001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 E:\FR\FM\29JYP2.SGM 29JYP2
45340 Federal Register / Vol. 80, No. 145 / Wednesday, July 29, 2015 / Proposed Rules
ENVIRONMENTAL PROTECTION OAR–2015–0310, by one of the Docket: All documents in the docket
AGENCY following methods: are listed in the http://
• Federal eRulemaking Portal: http:// www.regulations.gov index. Although
40 CFR Part 51 www.regulations.gov. Follow the online listed in the index, some information is
instructions for submitting comments. not publicly available, e.g., CBI or other
[EPA–HQ–OAR–2015–0310; FRL–9930–11– • Email: A-and-R-Docket@epa.gov. information whose disclosure is
OAR] Include docket ID No. EPA–HQ–OAR– restricted by statute. Certain other
2015–0310 in the subject line of the material, such as copyrighted material,
RIN 2060–AS54 message. will be publicly available only in hard
• Fax: (202) 566–9744. copy. Publicly available docket
Revision to the Guideline on Air • Mail: Environmental Protection
Quality Models: Enhancements to the materials are available either
Agency, Mail code 28221T, Attention electronically in www.regulations.gov or
AERMOD Dispersion Modeling System Docket No. EPA–HQ–OAR–2015–0310,
and Incorporation of Approaches To in hard copy at the Air and Radiation
1200 Pennsylvania Ave. NW., Docket and Information Center, EPA/
Address Ozone and Fine Particulate Washington, DC 20460. Please include a
Matter DC, Room 3334, WJC West Building,
total of two copies. 1301 Constitution Ave. NW.,
AGENCY: Environmental Protection • Hand/Courier Delivery: EPA Docket
Washington, DC. The Public Reading
Agency (EPA). Center, Room 3334, EPA WJC West
Room is open from 8:30 a.m. to 4:30
Building, 1301 Constitution Ave. NW.,
ACTION: Proposed rule; notice of p.m., Monday through Friday, excluding
Washington, DC. Such deliveries are
conference. legal holidays. The telephone number
only accepted during the Docket’s
for the Public Reading Room is (202)
SUMMARY: In this action, the normal hours of operation, and special
566–1744 and the telephone number for
Environmental Protection Agency (EPA) arrangements should be made for
the Air and Radiation Docket and
proposes to revise the Guideline on Air deliveries of boxed information.
Instructions: Direct your comments to Information Center is (202) 566–1742.
Quality Models (‘‘Guideline’’). The Public hearing and conference: The
Docket ID No. EPA–HQ–OAR–2015–
Guideline has been incorporated into public hearing for this action and the
0310. The EPA’s policy is that all
EPA’s regulations, satisfying a Eleventh Conference on Air Quality
comments received will be included in
requirement under the Clean Air Act Modeling will be held in the EPA
the public docket without change and
(CAA) section 165(e)(3) for the EPA to Auditorium, Room C111, 109 T.W.
may be made available online at http://
specify, with reasonable particularity Alexander Drive, Research Triangle
www.regulations.gov, including any
models to be used in the Prevention of Park, NC 27711.
personal information provided, unless
Significant Deterioration (PSD) program. FOR FURTHER INFORMATION CONTACT: Mr.
the comment includes information
It provides EPA-preferred models and claimed to be Confidential Business George M. Bridgers, Air Quality
other recommended techniques, as well Information (CBI) or other information Assessment Division, Office of Air
as guidance for their use in predicting whose disclosure is restricted by statute. Quality Planning and Standards, U.S.
ambient concentrations of air pollutants. Do not submit information that you Environmental Protection Agency, Mail
The proposed revisions to the Guideline consider to be CBI or otherwise code C439–01, Research Triangle Park,
include enhancements to the protected through http:// NC 27711; telephone: (919) 541–5563;
formulation and application of the www.regulations.gov or email. The fax: (919) 541–0044; email:
EPA’s AERMOD near-field dispersion www.regulations.gov Web site is an Bridgers.George@epa.gov.
modeling system and the incorporation ‘‘anonymous access’’ system, which SUPPLEMENTARY INFORMATION:
of a tiered demonstration approach to means the EPA will not know your
address the secondary chemical Table of Contents
identity or contact information unless
formation of ozone and fine particulate you provide it in the body of your The following topics are discussed in
matter (PM2.5) associated with precursor comment. If you send an email this preamble:
emissions from single sources. comment directly to the EPA without I. General Information
Additionally, the EPA proposes various going through http:// A. Does this action apply to me?
editorial changes to update and www.regulations.gov, your email B. What should I consider as I prepare my
reorganize information throughout the address will be automatically captured comments for the EPA?
Guideline to streamline the compliance C. Where can I get a copy of this
and included as part of the comment document?
assessment process. that is placed in the public docket and
Within this action, the EPA is also II. Background
made available on the Internet. If you A. The Guideline on Air Quality Models
announcing the Eleventh Conference on submit an electronic comment, the EPA and EPA Modeling Conferences
Air Quality Modeling and invites the recommends that you include your B. The Tenth Conference on Air Quality
public to participate in the conference. name and other contact information in Modeling
The conference will focus on the the body of your comment and with any III. Public Participation Regarding Revisions
proposed revisions to the Guideline and disk or CD ROM you submit. If the EPA to the Guideline and Notice of Eleventh
part of the conference will also serve as Conference on Air Quality Modeling
cannot read your comment due to IV. Proposed Changes to the Guideline
the public hearing for these revisions. technical difficulties and cannot contact A. Proposed Actions
DATES: Comments must be received on
you for clarification, the EPA may not 1. Clarifications To Distinguish

or before October 27, 2015. be able to consider your comment. Requirements From Recommendations
Public hearing and conference: The Electronic files should avoid the use of 2. Updates to EPA’s AERMOD Modeling
public hearing for this action and the special characters, any form of System
Eleventh Conference on Air Quality encryption, and be free of any defects or 3. Status of AERSCREEN
Modeling will be held August 12–13, 4. Updates to 3-Tiered Demonstration
viruses. For additional information Approach for NO2
2015, from 8:30 a.m. to 5:00 p.m. about the EPA’s public docket, visit the 5. Status of CALINE3 Models
ADDRESSES: Submit your comments, EPA Docket Center homepage at http:// 6. Addressing Single-Source Impacts on
identified by Docket ID No. EPA–HQ– www.epa.gov/epahome/dockets.htm. Ozone and Secondary PM2.5
Federal Register / Vol. 80, No. 145 / Wednesday, July 29, 2015 / Proposed Rules 45341
7. Status of CALPUFF and Assessing Long- electronically within the disk or CD national consistency is maintained in
Range Transport for PSD Increment and ROM the specific information that is air quality analyses for regulatory
Regional Haze claimed as CBI. In addition to one activities under 40 CFR 51.112, 51.117,
8. Role of EPA’s Model Clearinghouse complete version of the comment that 51.150, 51.160, 51.165, 51.166, 52.21,
9. Updates to Modeling Procedures for
Cumulative Impact Analysis
includes information claimed as CBI, a 93.116, 93.123, and 93.150.
10. Updates on Use of Meteorological Input copy of the comment that does not The EPA originally published the
Data for Regulatory Dispersion Modeling contain the information claimed as CBI Guideline in April 1978 (EPA–450/2–
11. Transition Period for Applicability of must be submitted for inclusion in the 78–027), and it was incorporated by
Revisions to the Guideline public docket. Information so marked reference in the regulations for the PSD
B. Proposed Editorial Changes will not be disclosed except in program in June 1978. The EPA revised
V. Statutory and Executive Order Reviews accordance with procedures set forth in the Guideline in 1986 (51 FR 32176),
A. Executive Order 12866: Regulatory 40 CFR part 2. and updated it with supplement A in
Planning and Review and Executive 2. Tips for preparing your comments. 1987 (53 FR 32081), supplement B in
Order 13563: Improving Regulation and July 1993 (58 FR 38816), and
Regulatory Review
When submitting comments, remember
to: supplement C in August 1995 (60 FR
B. Paperwork Reduction Act
C. Regulatory Flexibility Act • Follow directions—The agency may 40465). The EPA published the
D. Unfunded Mandates Reform Act ask you to respond to specific questions Guideline as appendix W to 40 CFR part
E. Executive Order 13132: Federalism or organize comments by referencing a 51 when the EPA issued supplement B.
F. Executive Order 13175: Consultation CFR part or section number. The EPA republished the Guideline in
and Coordination With Indian Tribal • Explain why you agree or disagree, August 1996 (61 FR 41838) to adopt the
Governments suggest alternatives, and substitute CFR system for labeling paragraphs. The
G. Executive Order 13045: Protection of language for your requested changes. publication and incorporation of the
Children From Environmental Health Guideline by reference into the EPA’s
and Safety Risks
• Describe any assumptions and
provide any technical information and/ PSD regulations satisfies the
H. Executive Order 13211: Actions requirement under the CAA section
Concerning Regulations That or data that you used.
Significantly Affect Energy Supply, • If you estimate potential costs or 165(e)(3) for the EPA to promulgate
Distribution, or Use burdens, explain how you arrived at regulations that specify with reasonable
I. National Technology Transfer and your estimate in sufficient detail to particularity models to be used under
Advancement Act allow for it to be reproduced. specified sets of conditions for purposes
J. Executive Order 12898: Federal Actions • Provide specific examples to of the PSD program.
To Address Environmental Justice in illustrate your concerns and suggest To support the process of developing
Minority Populations and Low-Income
alternatives. and revising the Guideline during the
Populations period of 1977–1988, we held the First,
• Explain your views as clearly as
I. General Information possible, avoiding the use of profanity Second, and Third Conferences on Air
or personal threats. Quality Modeling as required by CAA
A. Does this action apply to me? section 320 to help standardize
• Make sure to submit your
This action applies to federal, state, comments by the comment period modeling procedures. These modeling
territorial, and local air quality deadline identified. conferences provided a forum for
management programs that conduct air comments on the Guideline and
quality modeling as part of State C. Where can I get a copy of this associated revisions, thereby helping us
Implementation Plan (SIP) submittals document? introduce improved modeling
and revisions, New Source Review In addition to being available in the techniques into the regulatory process.
(NSR), including new or modifying docket, an electronic copy of this In October 1988, we held the Fourth
industrial sources under Prevention of Conference on Air Quality Modeling to
proposed rule will also be available on
Significant Deterioration (PSD), advise the public on new modeling
the Worldwide Web (WWW) through
Conformity, and other air quality techniques and to solicit comments to
the EPA’s Technology Transfer Network
assessments required under EPA guide our consideration of any
(TTN). Following signature, a copy of
regulation. Categories and entities rulemaking needed to further revise the
this proposed rule will be posted on the
potentially regulated by this action Guideline. We held the Fifth Conference
TTN’s Support Center for Regulatory
include: in March 1991, which also served as a
Atmospheric Modeling (SCRAM) Web
public hearing for the proposed
site at the following address: http://
NAICS a revisions to the Guideline. In August
Category www.epa.gov/ttn/scram. The TTN
Code 1995, we held the Sixth Conference as
provides information and technology
a forum to update our available
Federal/state/territorial/local/ exchange in various areas of air
modeling tools with state-of-the-science
tribal government .................. 924110 pollution control. techniques and for the public to offer
a North American Industry Classification II. Background new ideas.
System. The Seventh Conference was held in
A. The Guideline on Air Quality Models
June 2000, and also served as a public
B. What should I consider as I prepare and EPA Modeling Conferences
hearing for another round of proposed
my comments for the EPA?
The Guideline is used by the EPA, changes to the recommended air quality
1. Submitting CBI. Do not submit this other federal, state, territorial, and local models in the Guideline. These changes
information to the EPA through http:// air quality agencies, and industry to included the CALPUFF modeling
www.regulations.gov or email. Clearly prepare and review new source permits, system, AERMOD modeling system, and
mark any of the information that you source permit modifications, SIP ISC–PRIME model.
claim to be CBI. For CBI information in submittals and revisions, conformity, Subsequently, the EPA revised the
a disk or CD ROM that you mail to the and other air quality assessments Guideline on April 15, 2003 (68 FR
EPA, mark the outside of the disk or CD required under EPA regulation. The 18440), to adopt CALPUFF as the
ROM as CBI and then identify Guideline serves as a means by which preferred model for long-range transport
of emissions from 50 to several hundred public comments received are available and areas for improvement in the model
kilometers and to make various editorial at EPA’s SCRAM Web site at http:// and guidance on their use.
changes to update and reorganize www.epa.gov/ttn/scram/ Several presentations at the Tenth
information and remove obsolete 10thmodconf.htm. Modeling Conference addressed issues
models. and challenges associated with
The EPA promulgated a new 1-hour
We held the Eighth Conference on Air demonstrating compliance with these
Quality Modeling in September 2005. NAAQS for NO2 in January 2010, and a new 1-hour NAAQS for NO2 and SO2.
This conference provided details on new 1-hour NAAQS for SO2 in June This included results from a study
changes to the preferred air quality 2010. Although AERMOD evaluations sponsored by the American Petroleum
models, including available methods for that formed the basis of its promulgation Institute (API) that evaluated AERMOD
model performance evaluation and the as the EPA’s preferred dispersion model model performance under low wind
notice of data availability that the EPA demonstrated that AERMOD provides speed conditions using additional
published in September 2003, related to generally unbiased estimates of ambient National Oceanic and Atmospheric
the incorporation of the PRIME concentrations, the increased stringency Administration (NOAA) field studies at
downwash algorithm in the AERMOD of these new standards resulted in Oak Ridge, TN, and Idaho Falls, ID,
dispersion model (in response to increased scrutiny by the modeling which were not included in the original
comments received from the Seventh community of AERMOD model 17 databases used to support
Conference). Additionally, at the Eighth performance. In response, the EPA AERMOD’s promulgation in 2005. The
Conference, a panel of experts discussed issued several guidance memoranda to API low wind study 5 showed
the use of state-of-the-science prognostic clarify the applicability of the Guideline significant overprediction of observed
meteorological data for informing the and address initial issues with use of concentrations, especially for the Oak
dispersion models. current models and procedures under Ridge study where observed wind
The EPA further revised the Guideline PSD permitting.1 2 3 4 However, the speeds were below 0.5 m/s for 10 of the
on November 9, 2005 (70 FR 68218), to situation also necessitated the EPA and 11 tracer tests, and included wind
adopt AERMOD as the preferred model the modeling community to more speeds as low as 0.15 m/s. The API low
for near-field dispersion of emissions for closely evaluate the science and model wind study also included proposed
distances up to 50 kilometers. formulation of AERMOD to better modifications to the AERMET
The Ninth Conference on Air Quality understand the issues being experienced meteorological processor and AERMOD
Modeling was held in October 2008, and model to address this bias toward
by stakeholders and to address
emphasized the following topics: overprediction under stable/light wind
performance issues in its use for PSD
Reinstituting the Model Clearinghouse, conditions.
permitting under these new standards.
review of non-guideline applications of Prior to the promulgation of the 1-
dispersion models, regulatory status As part of this effort, the EPA
hour NO2 NAAQS, compliance with the
updates of AERMOD and CALPUFF, reconvened the AERMOD
previous annual NO2 NAAQS was
continued discussions on the use of Implementation Workgroup (AIWG) routinely demonstrated based on the
prognostic meteorological data for with state and local agency modelers to Tier 1 assumption of full conversion or
informing dispersion models, and evaluate AERMOD across a variety of a Tier 2 option based on an ambient
presentations reviewing the available hypothetical sources and results from ratio of 75 percent conversion of
model evaluation methods. this assessment were also presented at nitrogen oxides (NOX) to NO2, referred
this conference to inform the modeling to as the Ambient Ratio Method (ARM).
B. The Tenth Conference on Air Quality community of potential implications
Modeling However, compliance with the new 1-
hour NAAQS has typically required a
The most recent EPA modeling 1 U.S. EPA, 2010. Applicability of Appendix W
more refined treatment of NOX
conference was the Tenth Conference on Modeling Guidance for the 1-hour NO2 NAAQS. conversion to NO2. Therefore, several
Air Quality Modeling held in March Tyler Fox Memorandum dated June 28, 2010, Office
of Air Quality Planning & Standards, Research presentations at the Tenth Modeling
2012. This conference covered multiple Triangle Park, North Carolina 27711. http:// Conference focused on issues associated
topics which have been vital in the www.epa.gov/ttn/scram/guidance/clarification/ with demonstrating compliance with
development of the proposed revisions ClarificationMemo_AppendixW_Hourly-NO2- the new 1-hour NO2 NAAQS.
to the Guideline. The conference NAAQS_FINAL_06-28-2010.pdf.
2 U.S. EPA, 2010. Applicability of Appendix W
These presentations included an
addressed updates on the regulatory Modeling Guidance for the 1-hour SO2 NAAQS. overview of an API funded study to
status and future development of Tyler Fox Memorandum dated August 23, 2010, develop a Tier 2 ambient ratio method
AERMOD and CALPUFF, review of the Office of Air Quality Planning & Standards, for the 1-hour NO2 NAAQS, referred to
Mesoscale Model Interface (MMIF) Research Triangle Park, North Carolina 27711. as ARM2. The ARM2 approach was
http://www.epa.gov/ttn/scram/guidance/
prognostic meteorological data clarification/ClarificationMemo_AppendixW_ developed based on an extensive
processing tool for dispersion models, Hourly-SO2-NAAQS_FINAL_08-23-2010.pdf. analysis of ambient ratios of NO2/NOX
draft modeling guidance for compliance 3 U.S. EPA, 2010. Guidance Concerning the that were analyzed by land use (urban
demonstrations of the PM2.5 National Implementation of the 1-hour SO2 NAAQS for the vs. rural) and geographical areas. Based
Ambient Air Quality Standards Prevention of Significant Deterioration Program.
Stephen D. Page, Memorandum dated August 23,
on these analyses of the ambient NO2/
(NAAQS), modeling for compliance 2010, Office of Air Quality Planning & Standards, NOX ratios, an empirical relationship
demonstration of the 1-hour nitrogen Research Triangle Park, North Carolina 27711. between ambient concentrations of NO2
dioxide (NO2) and sulfur dioxide (SO2) http://www.epa.gov/region07/air/nsr/nsrmemos/ and NOX was developed. The EPA
NAAQS, and new and emerging appwso2.pdf.
4 U.S. EPA, 2010. Guidance Concerning the
subsequently reviewed and evaluated
models/techniques for future Implementation of the 1-hour NO2 NAAQS for the this ARM2 approach and then
consideration under the Guideline to Prevention of Significant Deterioration Program.
address single-source modeling for Stephen D. Page, Memorandum dated June 29, 5 AECOM, 2010. AERMOD low wind speed
ozone and secondary PM2.5, as well as 2010, Office of Air Quality Planning & Standards, evaluation study results, http://
Research Triangle Park, North Carolina 27711. mycommittees.api.org/rasa/amp/
long-range transport and chemistry. A http://www.epa.gov/ttn/scram/guidance/ Modeling%20Documents/AECOM%202009%
transcript of the conference proceedings clarification/ClarificationMemo_AppendixW_ 20Low%20Wind%20Speed%20Evaluation%
and a document that summarizes the Hourly-SO2-NAAQS_FINAL_08-23-2010.pdf. 20Study%20Report.pdf.
incorporated this screening technique as consists of representatives from the The public hearing for this action and
a non-Default/Beta option in version EPA, the U.S. Forest Service, the the Eleventh Conference on Air Quality
13350 of AERMOD in December 2013. National Park Service, and the U.S. Fish Modeling will be held August 12–13,
Another issue associated with NO2 and Wildlife Service, was initially 2015, from 8:30 a.m. to 5:00 p.m., in the
NAAQS compliance presented at this formed to support development of EPA Auditorium, Room C111, 109 T.W.
conference focused on the use of technically sound recommendations Alexander Drive, Research Triangle
relative (instantaneous) dispersion regarding assessment of air pollutant Park, NC 27711. On August 12, 2015,
coefficients to define the plume volume source impacts on Federal Class I parks the first half of the conference will
which determines the amount of ozone and wilderness areas. Comments consist of a structured agenda with
available to convert nitrogen (NO) to received from stakeholders at the Tenth presentations. The second half of the
NO2 using the Plume Volume Molar Modeling Conference supported first and all of the second day (August
Ratio Method (PVMRM) option in reinitiating this interagency 13, 2015), is reserved for the public
AERMOD. The relative dispersion collaborative effort (as ‘‘Phase 3’’) to hearing on this proposed rule. Advance
coefficients originally incorporated in provide additional guidance for requests for reserved time to speak
AERMOD for PVMRM are best modeling single-source impacts on during the public hearing should be
representative of daytime convective secondarily formed pollutants (e.g., submitted by August 7, 2015, to Mr.
conditions and may tend to ozone and PM2.5) in the near-field and George M. Bridgers, Air Quality
overestimate plume volumes during for long-range transport. Stakeholder Assessment Division, Office of Air
stable conditions. Such overestimation comments also support the idea of this Quality Planning and Standards, U.S.
of the plume volume will tend to result collaborative effort working in parallel Environmental Protection Agency, Mail
in PVMRM to overestimate with stakeholders to further model code C439–01, Research Triangle Park,
concentrations of NO2. development and evaluation. This NC 27711; telephone: (919) 541–5563;
In addition, modeling of single-source renewed 8 effort included the fax: (919) 541–0044; email:
impacts for ozone and secondarily establishment of two separate working Bridgers.George@epa.gov. The EPA will
formed PM2.5 was a topic of discussion groups, one focused on long-range also provide an opportunity for oral
at the Tenth Modeling Conference. On transport of primary and secondary presentations by individuals that sign
January 4, 2012, the EPA granted a pollutants and the other on near-field up at the public hearing. Information
petition submitted on behalf of the single-source impacts of secondary submitted to the EPA during the
Sierra Club on July 28, 2010 6 and pollutants. The primary objectives of conference will be placed in the docket
committed to engage in rulemaking to this phase of IWAQM include reviewing for this rule proposing revisions to the
evaluate whether updates to the existing approaches for estimating Guideline.
Guideline are warranted and, as single-source secondary pollutant Background information:
appropriate, incorporate new analytical impacts, developing revisions to the Preregistration details, additional
techniques or models for ozone and Guideline, and the development of background information, and a more
secondarily formed PM2.5. As a part of guidance for using technical methods to detailed agenda for the Eleventh
satisfying this commitment, there were estimate downwind secondary pollutant Conference on Air Quality Modeling are
presentations of ongoing research at the impacts. electronically available at http://
Tenth Modeling Conference regarding www.epa.gov/ttn/scram/
III. Public Participation Regarding 11thmodconf.htm. Preregistration for
single-source plume chemistry and Revisions to the Guideline and Notice the conference, while not required, is
photochemical grid modeling of Eleventh Conference on Air Quality strongly recommended due to
techniques, as well as several public Modeling heightened security protocols at the
forums. In addition, the EPA presented Interested persons may provide the EPA–RTP facility.
an overview along with a panel EPA with their views on the proposed REAL ID Act: Because this hearing is
discussion of its Draft Guideline for revisions to the Guideline in several being held at a U.S. government facility,
PM2.5 Permit Modeling that addressed ways. This includes submitting written individuals planning to attend the
the need for consideration of secondary comments to the EPA, participating in hearing should be prepared to show
PM2.5 in demonstrating compliance with the Eleventh Conference on Air Quality valid picture identification to the
the PM2.5 NAAQS.7 Subsequently, Modeling, and speaking at the public security staff in order to gain access to
written comments pertaining to such hearing that will be conducted as part the meeting room. Please note that the
modeling were submitted to the EPA. of the conference. Additional REAL ID Act, passed by Congress in
As introduced at the Tenth Modeling information on how to submit written 2005, established new requirements for
Conference, the Interagency Workgroup comments on the proposed revisions to entering federal facilities. These
on Air Quality Modeling (IWAQM) the Guideline is provided in the requirements took effect July 21, 2014.
process was formally reinitiated in June ADDRESSES section above. If your driver’s license is issued by
2013 to inform the EPA’s process of Alaska, American Samoa, Arizona,
updating the Guideline to address 8 Phase 1 of the IWAQM effort focused on review
Kentucky, Louisiana, Maine,
chemically reactive pollutants in near- of respective regional modeling programs, Massachusetts, Minnesota, Montana,
field and long-range transport development of an organizational framework, and
formulating reasonable objectives and plans that New York, Oklahoma, or the state of
applications. The IWAQM, which were presented to EPA management for support and Washington, you must present an
6 U.S. EPA, 2012. Gina McCarthy Letter to Robert

commitment. Phase 2 of the IWAQM process additional form of identification to enter
continued this work and largely concluded in 1998 the federal buildings where the public
Ukeiley dated January 4, 2012, Washington, DC with the publication of the Interagency Workgroup
20460. http://www.epa.gov/ttn/scram/ on Air Quality Modeling (IWAQM) Phase 2
hearings will be held. Acceptable
10thmodconf/review_material/Sierra_Club_ Summary Report and Recommendations for alternative forms of identification
Petition_OAR-11-002-1093.pdf. Modeling Long Range Transport Impacts (EPA–454/ include: Federal employee badges,
7 U.S. EPA, 2014. Guidance for PM
2.5 Modeling. R–98–019) (USEPA, 1998). This report provided a passports, enhanced driver’s licenses
May 20, 2014, EPA–454/B–14–001. Office of Air series of recommendations concerning the
Quality Planning & Standards, Research Triangle application of the CALPUFF model for use in
and military identification cards. We
Park, NC. http://www.epa.gov/ttn/scram/guidance/ regulatory long-range transport (LRT) modeling that will list any additional acceptable forms
guide/Guidance_for_PM25_Permit_Modeling.pdf. supported the revisions in 2003 to the Guideline. of identification at: http://www.epa.gov/
ttn/scram/11thmodconf.htm. In Any person in attendance wishing to CFR 52.21(l). The applicable models are
addition, you will need to obtain a speak at the public hearing who has not the preferred models listed in appendix
property pass for any personal reserved time prior to the conference A to appendix W to 40 CFR part 51.
belongings you bring with you. Upon may provide oral comments on the However, there has been some
leaving the building, you will be proposed revisions to the Guideline ambiguity in the past with respect to the
required to return this property pass to during time allotted on the last day of ‘‘other requirements’’ specified in the
the security desk. No large signs will be the conference. These parties will need Guideline that must be used in PSD
allowed in the building, cameras may to sign up to speak on the second day permitting analysis and other regulatory
only be used outside of the building and of the hearing and the EPA may need to modeling assessments.
demonstrations will not be allowed on limit the duration of presentations to
Ambiguity can result because the
federal property for security reasons. allow all participants to be heard.
Additional written statements or Guideline generally contains
Conference and Public Hearing: The ‘‘recommendations’’ and these
Eleventh Conference on Air Quality comments on the proposed revisions
should be sent to the OAR Regulatory recommendations are expressed in non-
Modeling will be open to the public. No
Docket (see ADDRESSES section). A mandatory language. For instance, the
registration fee is charged. The
transcript of the conference proceedings Guideline frequently uses ‘‘should’’ and
conference will be conducted informally
and a copy of all written comments will ‘‘may’’ rather than ‘‘shall’’ and ‘‘must.’’
and chaired by an EPA official. As
be maintained in Docket ID No. EPA– This approach is generally preferred
required under CAA section 320, a
HQ–OAR–2015–0310, which will throughout the Guideline because of the
verbatim transcript of the conference
remain open until October 27, 2015, for need to exercise expert judgment in air
proceedings will be produced and
the purpose of receiving additional quality analysis and the reasons
placed in the docket for this proposed
comments after the conference and the discussed in the Guideline that ‘‘dictate
rule.
The Eleventh Conference on Air public hearing on the proposed against a strict modeling ‘cookbook’.’’
Quality Modeling will begin with revisions to the Guideline. (40 CFR part 51, appendix W, section
introductory remarks by the presiding 1.0(c))
IV. Proposed Changes to the Guideline
EPA official. The EPA staff will then Considering the non-mandatory
provide an overview of the revisions to In this action, the EPA is proposing language used throughout the Guideline,
the Guideline as proposed in this two type of revisions to the Guideline. the EPA’s Environmental Appeals Board
document and present on the research The first involve substantive changes to has correctly observed the following:
that supports those revisions and address various topics, including those
presented and discussed at the Tenth ‘‘Although appendix W has been
supports formulation updates to the promulgated as codified regulatory text,
preferred models. The following topics Modeling Conference. These proposed
revisions to the Guideline include appendix W provides permit issuers broad
will be presented: latitude and considerable flexibility in
enhancements to the formulation and
I. Overview of the Eleventh Conference application of air quality modeling.
application of the EPA’s preferred
on Air Quality Modeling; Appendix W is replete with references to
dispersion modeling system, AERMOD,
II. Review of the proposed revisions to ‘‘recommendations,’’ ‘‘guidelines,’’ and
and the incorporation of a tiered
the Guideline; and reviewing authority discretion.’’
III. Review of the proposed revisions to demonstration approach to address the
the preferred air quality models. secondary chemical formation of ozone In Re Prairie State Generating Company,
and PM2.5 associated with precursor 13 E.A.D. 1, 99 (EAB 2005) (internal
At the conclusion of the emissions from single sources. The
presentations, the EPA will convene the citations omitted).
second type of revision involves
public hearing on the proposed editorial changes to update and Although this approach is typical
revisions to the Guideline. The public reorganize information throughout the throughout the Guideline, there are
hearing will span the second half of the Guideline. These revisions are not instances where the EPA does not
first day and throughout the second day intended to meaningfully change the believe permit issues should have broad
of the conference. substance of the Guideline, but rather to latitude. Some principles of air quality
Those wishing to reserve time to make the Guideline easier to use and to modeling described in the Guideline
speak at the public hearing, whether to streamline the compliance assessment must always be applied to produce an
volunteer a presentation on a special process. acceptable analysis. Thus, to promote
topic or to offer general comment on any clarity in the use and interpretation of
of the modeling techniques scheduled A. Proposed Actions the revised Guideline, we have, in these
for presentation, should contact us at This section provides a detailed cases used mandatory language, and
the address given in the FOR FURTHER overview of the substantive proposed made specific reference to
INFORMATION CONTACT section (note the changes to the Guideline that are ‘‘requirements’’ throughout the
cutoff date). Such persons should intended to improve the science of the proposed text where appropriate to
identify the organization (if any) on models and approaches used in distinguish requirements from
whose behalf they are speaking and the regulatory assessments. recommendations in the application of
length of the presentation. If a models for regulatory purposes. We
scheduled presentation is projected to 1. Clarifications To Distinguish solicit comment regarding the
Requirements From Recommendations
be longer than 10 minutes, the presenter appropriateness of these revisions in

should also state why a longer period is The EPA’s PSD permitting regulations providing the necessary clarity on the
needed. Scheduled speakers should specify that ‘‘[a]ll applications of air requirements under the proposed
bring extra copies of their presentation quality modeling involved in this revisions to the Guideline as distinct
for inclusion in the docket and for the subpart shall be based on the applicable from the recommendations in the
convenience of the recorder. Scheduled models, data bases, and other revised text while noting the continued
speakers will also be permitted to enter requirements specified in appendix W flexibilities provided for within the
additional written comments into the of this part (Guideline on Air Quality Guideline including but not limited to
record. Models).’’ 40 CFR 51.166(l); see also 40 use and approval of alternative models.
2. Updates to EPA’s AERMOD Modeling proposed option is selected by the calculating variables needed by the
System model user with the SOURCE type algorithms.
Based on studies presented and ‘‘BOUYLINE’’ to specify the individual We invite comment on incorporation
discussed at the Tenth Modeling buoyant line source locations and of AERSCREEN into the Guideline as
Conference, and additional relevant emissions and the new ‘‘BLAVGVAL’’ the screening model for AERMOD that
research since 2010, the EPA and other keyword to specify average parameters may be applicable in applications in all
researchers have conducted additional for a composite buoyant line. types of terrain and for applications
model evaluations and developed 6. Proposed updates to the NO2 Tier involving building downwash.
changes to the model formulation of the 2 and Tier 3 screening techniques coded 4. Updates to 3-Tiered Demonstration
AERMOD modeling system to improve within AERMOD as described more Approach for NO2
model performance in its regulatory fully later in this preamble section.
Section 5.2.4 of the 2005 Guideline
applications. We propose the following Model performance evaluation and
details a 3-tiered approach for assessing
updates to the AERMOD modeling peer scientific review references for the
NOX sources, which was recommended
system to address a number of technical updated AERMOD modeling system are
to obtain annual average estimates of
concerns expressed by stakeholders: cited, as appropriate. An updated user’s
NO2 from point sources for purposes of
1. A proposed option incorporated in guide and model formulation NSR analysis, including the PSD
AERMET to adjust the surface friction documents for version 15181 have been program and SIP planning purposes.
velocity (u*) to address issues with placed in the docket. We have updated This 3-tiered approach addresses the co-
AERMOD model overprediction under the summary description of the emissions of NO and NO2 and the
stable, low wind speed conditions. This AERMOD modeling system to appendix subsequent conversion of NO to NO2 in
proposed option is selected by the user A of the Guideline to reflect these the atmosphere. The tiered levels
with the METHOD STABLEBL ADJ_U* proposed updates. The essential codes, include: (1) Assuming that all NO is
record in the AERMET Stage 3 input preprocessors, and test cases have been converted to NO2 (full conversion), (2)
file. updated and posted to the EPA’s
2. A proposed low wind option in using the Ambient Ratio Method (ARM),
SCRAM Web site, http://www.epa.gov/ which applies an assumed equilibrium
AERMOD to address issues with model ttn/scram.
overprediction under low wind speed ratio of NO2 to NOX, based on observed
We invite comments on whether we ambient conditions, to the annual
conditions. The low wind option will have reasonably addressed the technical
increase the minimum value of the results from the Tier 1 full conversion,
concerns expressed by the stakeholder and (3) detailed screening options
lateral turbulence intensity (sigma-v) community and are on sound footing to
from 0.2 to 0.3 and adjusts the focused on determining site-specific
recommend these updates to the ratios of NO2 to NOX.
dispersion coefficient to account for the regulatory default version of the In January 2010, a new 1-hour NO2
effects of horizontal plume meander on AERMOD modeling system which standard was promulgated. Prior to the
the plume centerline concentration. It includes its replacement of BLP as an adoption of the 1-hour NO2 standard,
also eliminates upwind dispersion appendix A model for the intended few PSD permit applications required
which is incongruous with a straight- regulatory applications. the use of Tier 3 options and guidance
line, steady-state plume dispersion
3. Status of AERSCREEN available at the time did not fully
model such as AERMOD. The proposed
address the modeling needs for a 1-hour
option is selected by specifying In the preamble of the 2005 Guideline, standard, i.e., tiered approaches for NO2
‘‘LOWWIND3’’ on the CO MODELOPT we stated that a screening version of in the 2005 Guideline specifically
keyword in the AERMOD input file. AERMOD called AERSCREEN was being
3. Modifications to AERMOD targeted an annual standard. As a result,
developed and, in the meantime, several guidance memoranda have been
formulation to address issues with
SCREEN3 may be used until issued by the EPA to further inform
overprediction for applications
AERSCREEN was available. In 2011, the modeling procedures for sources
involving relatively tall stacks located
EPA released AERSCREEN, a program demonstrating compliance with the new
near relatively small urban areas (no
that creates inputs and runs AERMOD 1-hour standard.1 2 3 4. In response to the
user input is required).
4. Proposed regulatory default options in screening mode. AERSCREEN also 1-hour NO2 standard, the EPA is
in AERMOD to address plume rise for interfaces with AERMOD’s terrain proposing several modifications to the
horizontal and capped stacks based on processor, AERMAP, the building Tier 2 and 3 NO2 screening techniques
the July 9, 1993, Model Clearinghouse processor for AERMOD, BPIPPRIME, incorporated into AERMOD.
memorandum,9 with adjustments to and can use AERSURFACE surface For the Tier 2 technique, the EPA is
account for the PRIME algorithm for characteristics in the generation of proposing to replace the existing ARM
sources subject to building downwash. meteorological data for AERMOD via with a revised Ambient Ratio Method 2
These options are selected by the model the MAKEMET utility. In an April 2011 (ARM2). The existing Tier 2 technique,
memorandum, the EPA stated that ARM, was based on a study that focused
user specifying ‘‘POINTCAP’’ or
AERSCREEN was the recommended exclusively on long-term averages.10 A
‘‘POINTHOR’’ for source type on the SO
screening model for simple and recently published study 11 presented a
LOCATION keyword in the AERMOD
complex terrain and replaced SCREEN3. new analysis of national levels of
input file.
5. A proposed buoyant line source Since AERSCREEN invokes AERMOD, ambient ratios of NO2 to NOX based on
option, based on the BLP model, has AERSCREEN represents the state of the
10 Chu, S.H. and E.L. Meyer, 1991. Use of
been incorporated in AERMOD. This science in screening dispersion models.
Ambient Ratios to Estimate Impact of NOX Sources
As part of this proposed update to on Annual NO2 Concentrations. Proceedings, 84th
9 U.S. EPA, 1993. ‘‘Proposal for Calculating Plume AERSCREEN, AERSCREEN now Annual Meeting & Exhibition of the Air & Waste
Rise for Stacks with Horizontal Releases or Rain includes inversion break-up and coastal Management Association, Vancouver, B.C.; 16–21
Caps for Cookson Pigment, Newark, New Jersey’’, fumigation, features that were part of June 1991. (16pp.) (Docket No. A–92–65, II–A–9).
Memorandum from Joseph A. Tikvart, U.S. EPA/ 11 Podrez, M. 2015. An Update to the Ambient
OAQPS, Research Triangle Park, NC. July 9, 1993.

SCREEN3. These fumigation algorithms Ratio Method for 1-h NO2 Air Quality Standards
http://www.epa.gov/ttn/scram/guidance/mch/new_ also take advantage of AERMOD’s Dispersion Modeling. Atmospheric Environment,
mch/R1076_TIKVART_9_JUL_93.pdf. boundary layer parameterizations for 103: 163–170.
hourly data from the EPA’s Air Quality foundation to recommend the updates including PM2.5, PM10, and CO hot-spot
System (AQS). Based on this analysis, a described above. analyses.17
new second tier NO2 screening We solicit comments on our proposal
5. Status of CALINE3 Models
technique, ARM2, has been developed to identify AERMOD as a replacement
and incorporated into AERMOD. The 2005 Guideline identified for CALINE3 as an appendix A model
Because ARM2 is based on hourly CALINE3 14 and its variants (CAL3QHC for its intended regulatory applications.
measurements of the NO2 to NOX ratios and CAL3QHCR) as the preferred model
and provides more detailed estimates of for mobile source modeling for carbon 6. Addressing Single-Source Impacts on
this ratio based on the total NOX monoxide (CO), particulate matter (PM), Ozone and Secondary PM2.5
present, the EPA is proposing to and lead. CALINE3 was developed in On January 4, 2012, the EPA granted
incorporate a modified version of ARM2 the late 1970’s using P–G stability a petition submitted on behalf of the
as the new preferred second tier NOX classes as the basis for the dispersion Sierra Club on July 28, 2010,18 that
modeling approach. algorithms. AERMOD, on the other requested the EPA initiate rulemaking to
For the Tier 3 technique, the EPA hand, uses a planetary boundary layer establish air quality models for ozone
proposes that the existing detailed scaling parameter to characterize and PM2.5 for use by all major sources
screening options of the Ozone Limiting stability and determine dispersion rates, applying for a PSD permit. In granting
Method (OLM) 12 and PVMRM 13 be which has been found to be superior to that petition, the EPA explained that the
formally incorporated into the dispersion parameterizations based on ‘‘complex chemistry of ozone and
regulatory version of AERMOD. Both P–G stability classes.15 In addition, the secondary formation of PM2.5 are well-
OLM and PVMRM have been available LINE and AREA source options in
documented and have historically
as non-regulatory, non-default options AERMOD implement a full numerical
presented significant challenges to the
in AERMOD for many years, but their integration of emissions across the LINE
designation of particular models for
usage in a NAAQS compliance or AREA sources, whereas the CALINE3
assessing the impacts of individual
demonstration required approval by the family of models incorporate a much
stationary sources on the formation of
reviewing authority. Based on the EPA’s less refined approach. Thus, AERMOD
these air pollutants’’ and further
evaluation and external studies provides a more scientifically credible
explained that ‘‘[b]ecause of these
available on their performance, which and accurate characterization of plume
considerations, the EPA’s judgment in
show that OLM and PVMRM are dispersion than CALINE3. Recent model
the past has been that it was not
capable of modeling 1-hour NO2 performance studies 16 have shown that
the CALINE models performed poorly technically sound to designate with
impacts and NO and NO2 speciation
when compared to AERMOD and other particularity specific models that must
with reasonable accuracy when applied
modern dispersion models which also be used to assess the impacts of a single
appropriately, both OLM and PVMRM
employ state-of-the-science dispersion source on ozone concentrations,’’ but
are being proposed as preferred Tier 3
parameters. AERMOD is also able to rather the EPA had established a process
screening methods for NO2 modeling. In
model multiple years in a single model for determining on a case-by-case basis
addition, the EPA is proposing to
run, while the CALINE3 variants are the analytical techniques that should be
incorporate a revised version of the
limited to either a single meteorological used for ozone, as well as for secondary
PVMRM option, referred to as
condition (CALINE3 and CAL3QHC) or formation of PM2.5.
PVMRM2, that utilizes relative
dispersion coefficients to estimate a single year of meteorological data In the petition grant, the EPA
plume volume during convective (CAL3QHCR). Additionally, AERMOD committed to engage in rulemaking to
conditions and total dispersion is able to utilize more recent, and more evaluate whether updates to the
coefficients during stable conditions. representative, meteorological Guideline are warranted and, as
These adjustments to the calculation of observations than are readily available appropriate, incorporate new analytical
plume volume are intended to mitigate for modeling with CAL3QHCR. Based techniques or models for ozone and
potential overprediction of NO2 on the more scientifically sound basis secondarily formed PM2.5. This
conversion in multisource applications, for AERMOD, improved model rulemaking satisfies the EPA’s
especially during stable meteorological performance over CALINE3, and the commitment in the petition grant. As a
conditions. The EPA is proposing to availability of more representative part of this commitment and in
replace the existing PVMRM with the meteorological data, the EPA proposes compliance with CAA section 320, the
new PVMRM2 with both versions being replacing CALINE3 with AERMOD as EPA conducted the Tenth Modeling
made available in the proposed version the preferred appendix A model for Conference in March 2012, where there
of AERMOD to facilitate testing and determining near-field impacts for were presentations of ongoing research
evaluation of the EPA’s proposed primary emissions from mobile sources, of single-source plume chemistry and
replacement of PVMRM option with photochemical grid modeling
new PVMRMR2 option. 14 Benson, Paul E., 1979. CALINE3—A Versatile techniques, as well as several public
We invite comments on whether we Dispersion Model for Predicting Air Pollutant forums, and the EPA subsequently
Levels Near Highways and Arterial Streets. Interim received written comments pertaining to
have reasonably addressed technical Report, Report Number FHWA/CA/TL–79/23.
concerns regarding the 3-tiered Federal Highway Administration, Washington, DC
such modeling.
demonstration approach and specific (NTIS No. PB 80–220841).
15 Cimorelli, A. et al., 2005. AERMOD: A 17 U.S. EPA, 2013, Transportation Conformity
NO2 screening techniques within
Dispersion Model for Industrial Source Guidance for Quantitative Hot-Spot Analyses in
AERMOD and whether we are on sound Applications. Part I: General Model Formulation PM2.5 and PM10 Nonattainment and Maintenance
and Boundary Layer Characterization. Journal of Areas. Publication No. EPA–420–B–13–053, Office
12 Cole, H.S. and J.E. Summerhays, 1979. A Applied Meteorology, 44(5): 682–693. of Transportation and Air Quality, Ann Arbor, MI.
Review of Techniques Available for Estimation of 16 Heist, D., V. Isakov, S. Perry, M. Snyder, A. http://www.epa.gov/otaq/stateresources/transconf/
Short-Term NO2 Concentrations. Journal of the Air Venkatram, C. Hood, J. Stocker, D. Carruthers, S. policy/420b13053-sec.pdf.
Pollution Control Association, 29(8): 812–817. Arunachalam, AND C. Owen. Estimating near-road 18 U.S. EPA, 2012. Gina McCarthy Letter to Robert
13 Hanrahan, P.L., 1999. The Polar Volume Polar pollutant dispersion: a model inter-comparison. Ukeiley dated January 4, 2012, Washington, DC
Ratio Method for Determining NO2/NOX Ratios in Transportation Research Part D: Transport and 20460. http://www.epa.gov/ttn/scram/
Modeling—Part I: Methodology. Journal of the Air Environment. Elsevier BV, AMSTERDAM, 10thmodconf/review_material/Sierra_Club_
& Waste Management Association, 49: 1324–1331. Netherlands, 25:93–105, (2013). Petition_OAR-11-002-1093.pdf.
The EPA initiated Phase 3 of the surrounding chemical environment.20 21 have been used extensively to support
IWAQM process in June 2013 to inform These oxidant and neutralizing agents ozone and PM2.5 SIPs and to explore
this process to update the Guideline to are not routinely measured, but can be relationships between inputs and air
address chemically reactive pollutants generated with a three-dimensional quality impacts in the United States and
for near-field and long-range transport photochemical transport model and elsewhere.26 29 30
applications. Comments received from subsequently input to a Lagrangian For assessing secondary pollutant
stakeholders at the Tenth Modeling modeling system. impacts from single sources, the degree
Conference supported this collaborative In light of these differences between of complexity required to assess
effort to provide additional guidance for photochemical grid models and potential impacts varies depending on
modeling single-source impacts of Lagrangian models that address the nature of the source, its emissions,
secondarily formed pollutants in the chemistry, the EPA believes and the background environment. In
near-field and for long-range transport. photochemical grid models are order to provide the user community
Stakeholder comments also supported generally most appropriate for flexibility in estimating single-source
the idea of this collaborative effort addressing ozone and secondary PM2.5 secondary pollutant impacts and given
occurring in parallel with stakeholders’ because they provide a spatially and the emphasis on the use of
efforts to further model development temporally dynamic realistic chemical photochemical grid models for these
and evaluation. The EPA’s and physical environment for plume purposes, the EPA is proposing a two-
recommended revisions to the Guideline growth and chemical tiered demonstration approach for
are largely based on detailed review and transformation.20 22 Publically available addressing single-source impacts on
assessment of this input. and documented Eulerian ozone and secondary PM2.5. The first
For this proposed revision to the photochemical grid models such as the tier involves use of technically credible
Guideline, the EPA has determined that Comprehensive Air Quality Model with relationships between precursor
advances in photochemical modeling Extensions (CAMx) 23 and the emissions and a source’s impacts that
science indicate it is now reasonable to Community Multiscale Air Quality may be published in the peer-reviewed
provide more specific, generally- (CMAQ) 24 model treat emissions, literature; developed from modeling that
applicable guidance that identifies chemical transformation, transport, and was previously conducted for an area by
particular models or analytical deposition using time and space variant a source, a governmental agency, or
techniques that may be used under meteorology. These modeling systems some other entity and that is deemed
specific circumstances for assessing the include primarily emitted species and sufficient; or generated by a peer-
impacts of an individual source on secondarily formed pollutants such as reviewed reduced form model. The
ozone and secondary PM2.5. ozone and PM2.5.25 26 27 28 These models second tier involves application of more
Quantifying secondary pollutant sophisticated case-specific chemical
formation requires simulating chemical
20 Baker, K.R., Kelly, J.T., 2014. Single source transport models (e.g., photochemical
impacts estimated with photochemical model grid models) to be determined in
reactions and thermodynamic source sensitivity and apportionment approaches.
partitioning in a realistic chemical and Atmospheric Environment, 96: 266–274.
consultation with the EPA Regional
physical environment. Chemical 21 ENVIRON, 2012. Evaluation of chemical Office and conducted consistent with
transport models treat atmospheric dispersion models using atmospheric plume new EPA single-source modeling
measurements from field experiments, EPA guidance.31 The appropriate tier for a
chemical and physical processes such as Contract No: EP–D–07–102. September 2012. 06–
deposition and transport. There are two 20443M6. http://www.epa.gov/ttn/scram/reports/
given application should be selected in
types of chemical transport models, Plume_Eval_Final_Sep_2012v5.pdf. consultation with the appropriate
which are differentiated based on a 22 Zhou, W., Cohan, D.S., Pinder, R.W., Neuman, reviewing authority and be consistent
fixed frame of reference (i.e., Eulerian J.A., Holloway, J.S., Peischl, J., Ryerson, T.B., with EPA guidance.
Nowak, J.B., Flocke, F., Zheng, W.G., 2012. To fully implement these proposed
models, specifically photochemical grid Observation and modeling of the evolution of Texas
models) or a frame of reference that power plant plumes. Atmospheric Chemistry and
changes to the Guideline related to
moves with parcels of air between the Physics, 12: 455–468. addressing ozone and secondary PM2.5
source and receptor point (i.e.,
23 ENVIRON, 2014. User’s Guide Comprehensive impacts, the EPA intends to pursue a
Air Quality Model with Extensions version 6, separate rulemaking to establish a
Lagrangian models).19 http://www.camx.com. ENVIRON International
Comparing these two types of Corporation, Novato.
technical basis and new values for PM2.5
chemical transport models, 24 Byun, D., Schere, K.L., 2006. Review of the Significant Impact Levels (SILs) and to
photochemical grid models are governing equations, computational algorithms, and introduce a new demonstration tool for
integrated, three-dimensional grid-based
other components of the models-3 Community ozone and PM2.5 precursors referred to
Multiscale Air Quality (CMAQ) modeling system. as Model Emissions Rates for Precursors
models that treat chemical and physical Applied Mechanics Reviews, 59: 51–77.
processes in each grid cell and use 25 Chen, J., Lu, J., Avise, J.C., DaMassa, J.A., (MERP). When completed, this rule
Eulerian diffusion and transport Kleeman, M.J., Kaduwela, A.P., 2014. Seasonal
modeling of PM 2.5 in California’s San Joaquin 29 Cai, C., Kelly, J.T., Avise, J.C., Kaduwela, A.P.,
processes to move chemical species to Valley. Atmospheric Environment, 92: 182–190. Stockwell, W.R., 2011. Photochemical modeling in
other grid cells.19 While some 26 Civerolo, K., Hogrefe, C., Zalewsky, E., Hao, W., California with two chemical mechanisms: model
Lagrangian models also treat in-plume Sistla, G., Lynn, B., Rosenzweig, C., Kinney, P.L., intercomparison and response to emission
gas and particulate chemistry, to do so 2010. Evaluation of an 18-year CMAQ simulation: reductions. Journal of the Air & Waste Management
Seasonal variations and long-term temporal changes Association, 61: 559–572.
these models require time and space in sulfate and nitrate. Atmospheric Environment, 30 Hogrefe, C., Hao, W., Zalewsky, E., Ku, J.-Y.,
varying oxidant concentrations, and in
44: 3745–3752. Lynn, B., Rosenzweig, C., Schultz, M., Rast, S.,
the case of PM2.5, neutralizing agents 27 Russell, A.G., 2008. EPA Supersites program- Newchurch, M., Wang, L., 2011. An analysis of
such as ammonia, because important related emissions-based particulate matter long-term regional-scale ozone simulations over the
modeling: initial applications and advances. Northeastern United States: variability and trends.
secondary impacts happen when plume Journal of the Air & Waste Management Atmospheric Chemistry and Physics, 11: 567–582.
edges start to interact with the Association, 58: 289–302. 31 U.S. EPA, 2015. Guidance on the use of models
28 Tesche, T., Morris, R., Tonnesen, G., McNally, for assessing the impacts from single sources on
19 McMurry, P.H., Shepherd, M.F., Vickery, J.S., D., Boylan, J., Brewer, P., 2006. CMAQ/CAMx secondarily formed pollutants ozone and PM2.5.
2004. Particulate matter science for policy makers: annual 2002 performance evaluation over the Publication No. EPA 454/P–15–001. Office of Air
A NARSTO assessment. Cambridge University eastern US. Atmospheric Environment, 40: 4906– Quality Planning & Standards, Research Triangle
Press. 4919. Park, North Carolina 27711.
would differ from the current process approaches similar to those for determined in consultation with the
recommended in the EPA’s Guidance PM2.5.32 33 appropriate EPA Regional Office based
for PM2.5 Permit Modeling.7 A MERP While the development of MERPs for upon new EPA single-source modeling
would neither replace the existing ozone and secondary PM2.5 precursors is guidance.31 Based on several scientific
Significant Emissions Rates (SERs) for expected to address a number of PSD studies, the EPA proposes to determine
these pollutants nor serve as the basis permitting situations, the EPA believes that photochemical grid models are
for the applicability of PSD that most of the remaining situations in appropriate for assessment of near-field
requirements to sources with emissions which a source must demonstrate and regional scale reactive pollutant
above the SER. However, a MERP would compliance under the proposed impacts from specific sources 20 22 34 35 or
represent a level of emissions of Guideline will be addressed sufficiently a group of multiple sources impacting
under the proposed first tier where an area.25 27 28 Even though single-source
precursors that is not expected to
existing technical information could be emissions are injected into a grid
contribute significantly to volume, photochemical transport
used in combination with other
concentrations of ozone or secondarily- models have been shown to adequately
supportive information and analysis for
formed PM2.5. Our present capture single-source impacts when
the purposes of estimating secondary
understanding of the atmospheric impacts from a particular source. The compared with downwind in-plume
science of ozone and secondary PM2.5 existing technical information should measurements.20 22 Where set up
formation indicates that MERP values provide a credible and representative appropriately for the purposes of
will likely be higher than the SERs and estimate of the secondary impacts from assessing the contribution of single
more appropriate for evaluating the the project source. In these situations, a sources to primary and secondarily
impacts of these criteria pollutants as more refined approach for estimating formed pollutants, photochemical grid
precursors to ozone and PM2.5 secondary pollutant impacts from models can be used with a variety of
formation. As part of the separate project sources may not be necessary. approaches to estimate these impacts.
rulemaking, the EPA intends to The EPA has been compiling and These approaches generally fall into the
demonstrate that a source with reviewing screening approaches that are category of source sensitivity (how air
precursor emissions (e.g., NOX and SO2 based on technically credible tools (e.g., quality changes due to changes in
for PM2.5) below the MERP level will photochemical grid models) that relate emissions) and source apportionment
have ambient impacts that will be less source precursor emissions to secondary (what air quality impacts are related to
than the SIL and, thereby, provide a impacts. In review of existing certain emissions). Source
sufficient demonstration that the source approaches detailed in peer reviewed apportionment has been used to
journal articles and non-peer reviewed differentiate the contribution from
will not cause or contribute to a
forms (e.g., technical reports, conference single sources on model predicted
violation of the PM2.5 NAAQS or PSD
presentations), it is not clear that a ozone and PM2.5 concentrations.20 34 The
increments. The EPA’s Guidance for direct decoupled method (DDM) has
PM2.5 Permit Modeling 7 provides for a single approach has been clearly
proposed to and evaluated by the also been used to estimate ozone and
three-tiered approach to address PM2.5 impacts from specific sources 20 35
secondary PM2.5 with (1) a qualitative modeling community for estimating
screening level secondary impacts from as well as the simpler brute-force
assessment; (2) a hybrid qualitative/ sensitivity approach.20 22 35 Limited
quantitative assessment utilizing single sources. Other screening level
alternatives to photochemical grid comparison of single-source impacts
existing technical work; and (3) a full between models 36 and approaches to
model application may include the use
quantitative modeling exercise. The differentiate single-source impacts 20 36
of existing credible photochemical
EPA expects that MERPs as a model impacts for sources deemed to be show generally similar downwind
demonstration tool will replace the first similar in terms of emission rates, spatial gradients and impacts.
tier of a qualitative assessment as Near-source in-plume aircraft based
release parameters, and background
sources that currently would provide a measurement field studies provide an
environment. The EPA will continue to
qualitative assessment are expected to opportunity for evaluating model
engage with the modeling community to
have precursor emissions levels below estimates of (near-source) downwind
identify credible alternative approaches
the MERP. The second and third tier of transport and chemical impacts from
for estimating single-source secondary
assessment will then be consistent with single stationary point sources.21
pollutant impacts which provide
the EPA’s proposed two-tiered Photochemical grid model source
flexibility and are less resource
apportionment and source sensitivity
demonstration approach for PM2.5 intensive for permit demonstration
simulation of a single source downwind
reflected in this proposed revisions to purposes.
impacts compare well against field
the Guideline. To specifically assist the For those situations for which existing
study primary and secondary ambient
public in commenting on this rule modeling or screening estimates are not
measurements made in Tennessee and
within the overall context of the NSR available or appropriate, the second tier
Texas.20 21 This work indicates
program, including PSD, the EPA has proposed by the EPA would apply and
photochemical grid models and source
added two separate memoranda to the involve use of more sophisticated case-
docket of this proposed rule. These specific chemical transport models (e.g., 34 Baker, K.R., Foley, K.M., 2011. A nonlinear
memoranda provide more details on photochemical grid models) to be regression model estimating single source
how this future approach to PSD concentrations of primary and secondarily formed
32 U.S. EPA, 2015. ‘‘Proposed Approach for PM2.5. Atmospheric Environment, 45: 3758–3767.
compliance demonstrations will work Demonstrating PM2.5 PSD Compliance’’, 35 Bergin, M.S., Russell, A.G., Odman, M.T.,
for secondary PM2.5 and also describe Memorandum to Docket No. EPA–HQ–OAR–2015– Cohan, D.S., Chameldes, W.L., 2008. Single-Source
our expectations for how such an 0310 by Tyler J Fox, U.S. EPA/OAQPS, Research Impact Analysis Using Three-Dimensional Air
Triangle Park, NC. June 30, 2015. Quality Models. Journal of the Air & Waste
approach might work for ozone based 33 U.S. EPA, 2015. ‘‘Proposed Approach for Management Association, 58: 1351–1359.
on a future, separate action to similarly Demonstrating Ozone PSD Compliance’’, 36 Baker, K.R., Kelly, J.T., Fox, T., 2013.
establish a SIL and MERPs (for VOC and Memorandum to Docket No. EPA–HQ–OAR–2015– Estimating second pollutant impacts from single
NOX precursors) for ozone using 0310 by Tyler J Fox, U.S. EPA/OAQPS, Research sources (control #27). http://aqmodels.awma.org/
Triangle Park, NC. June 30, 2015. conference-proceedings.
apportionment and source sensitivity EPA recognizes that long-range IWAQM reports 39 40 and published
approaches provide meaningful transport assessments may be necessary literature.
estimates of single-source impacts on in certain limited situations for PSD To inform future consideration of
ozone and secondarily-formed PM2.5. increment. For these situations, the EPA visibility modeling in regulatory
Additional evaluations for longer time is proposing a screening approach applications consistent with proposed
periods and more diverse environments, where CALPUFF along with other changes for addressing chemistry for
both physical and chemical, would be appropriate screening tools and single-source impact on ozone and
valuable to generate broader confidence methods may be used to support long- secondary PM2.5, the final report 40 of
in these approaches for this purpose. range transport PSD increment the IWAQM long-range transport
We invite comments on whether the assessments. subgroup identified that modern CTMs
proposed two-tiered demonstration have evolved sufficiently and provide a
approach and related EPA guidance is To determine if a Class I PSD credible platform for estimating
appropriately based on sound science increment analyses may be necessary potential visibility impacts from a single
and practical application of available beyond 50 km (i.e., long-range transport or small group of emission sources.
models and tools to address single- assessment), the EPA is recommending Chemical transport models are well
source impacts on ozone and secondary a screening approach to determine if a suited for the purpose of estimating
PM2.5. significant impact will occur with long-range impacts of secondary
particular focus on Class I areas that pollutants, such as PM2.5, that
7. Status of CALPUFF and Assessing
may be threatened at such distances. contribute to regional haze and other
Long-Range Transport for PSD
The first step relies upon the near-field secondary pollutants, such as ozone,
Increment and Regional Haze
application of the appropriate screening that contribute to negative impacts on
The 2003 Guideline recommended and/or preferred model to determine the vegetation through deposition
CALPUFF as the preferred model for significance of ambient impact at or processes. These multiple needs require
long-range transport (i.e., source- about 50 km from the new of modifying a full chemistry photochemical model
receptor distances of 50 to several source. If this initial analysis indicates capable of representing both gas,
hundred kilometers) of emissions from there may be significant ambient particle, and aqueous phase chemistry
point, volume, area, and line sources for impacts at that distance, then further for PM2.5, haze, and ozone.
primary criteria pollutants (e.g., PM and analysis is necessary. For assessment of Photochemical transport models are
SO2). Since that time, as discussed Class I ambient impacts, under the suitable for estimating visibility and
previously in this preamble, the EPA proposed Guideline, there will not a deposition since important physical and
has received input from stakeholders preferred model for distances beyond 50 chemical processes related to the
and has worked through the IWAQM km. Typically, a Lagrangian model is formation and transport of PM are
process on analytical techniques to realistically treated. Source sensitivity
the type of model appropriate to use for
address chemically reactive pollutants and apportionment techniques
these screening assessments; however,
for near-field and long-range transport implemented in photochemical grid
applicants should establish approaches
applications. As a result, in order to models have evolved sufficiently and
(models and modeling parameters) on a
provide the user community flexibility provide the opportunity for estimating
case-by-case basis in consultation with potential visibility and deposition
in estimating single-source secondary
pollutant impacts and given the the appropriate reviewing authority, impacts from one or a small group of
availability of more appropriate Regional Office, and the affected Federal emission sources using a full science
modeling techniques, such as Land Manager(s) (FLM(s)). If a photochemical grid model.
photochemical transport models (which cumulative increment analysis is Photochemical grid models using
address limitations of models like necessary, for these limited situations, meteorology output from prognostic
CALPUFF 37), the EPA is proposing that the selection and use of an alternative meteorological models have
the Guideline no longer contain model shall occur in agreement with the demonstrated skill in estimating source-
language that requires the use of appropriate reviewing authority receptor relationships in the near-
CALPUFF or another Lagrangian puff (paragraph 3.0(b)) and approval by the field 20 21 and over long distances.41
model for long-range transport EPA Regional Office based on the It is important that modeling tools
assessments. Additionally, the EPA is requirements of section 3.2.2(e). used for single-source long-range
proposing to remove the CALPUFF As previously noted, Phase 3 of the transport impacts assessments
modeling system as an EPA-preferred IWAQM process was reinitiated in June demonstrate skill in adequately
model for long-range transport due to 2013 to inform the EPA’s commitment replicating source-receptor relationships
concerns about the management and to update the Guideline to address that are not in close proximity. For
maintenance of the model code given chemically reactive pollutants in near-
39 U.S. EPA, 2015. Interagency Workgroup on Air
the frequent change in ownership of the field and long-range transport
Quality Modeling Phase 3 Summary Report: Near-
model code since promulgation in the applications. This Phase 3 effort Field Single Source Secondary Impacts. Publication
previous version of the Guideline.38 The included the establishment of a working No. EPA 454/P–15–002. Office of Air Quality
group composed of EPA and FLM Planning & Standards, Research Triangle Park,
37 U.S. EPA, 2009. Reassessment of the North Carolina 27711.
technical staff focused on long-range
Interagency Workgroup on Air Quality Modeling 40 U.S. EPA, 2015. Interagency Workgroup on Air
transport of primary and secondary
(IWAQM) Phase 2 Summary Report; Revisions to Quality Modeling Phase 3 Summary Report: Long
Phase 2 Recommendations. Draft. Office of Air pollutants with an emphasis on use of Range Transport and Air Quality Related Values.
Quality Planning & Standards, Research Triangle consistent approaches to those being Publication No. EPA 454/P–15–003. Office of Air
Park, North Carolina 27711. http://www.epa.gov/ Quality Planning & Standards, Research Triangle
ttn/scram/guidance/reports/Draft_IWAQM_
developed and applied to meet near- Park, North Carolina 27711.
Reassessment_052709.pdf. field assessment needs for ozone and 41 ENVIRON, 2012. Documentation of the
38 U.S. EPA, 2015. ‘‘Summary of CALPUFF secondarily-formed PM2.5. The EPA Evaluation of CALPUFF and Other Long Range
Ownership Since 2003 Promulgation’’, expects that such approaches will be Transport Models using Tracer Field Experiment
Memorandum to Docket No. EPA–HQ–OAR–2015– Data, EPA Contract No: EP–D–07–102. February
0310 by Tyler J Fox, U.S. EPA/OAQPS, Research
focused on state of the science chemical 2012. 06–20443M4. http://www.epa.gov/ttn/scram/
Triangle Park, NC. June 30, 2015. transport models (CTMs) as detailed in reports/EPA-454_R-12-003.pdf.
source-receptor distances greater than consultation with the appropriate EPA model in this Guideline does not affect
50 km, regional scale photochemical Regional Office. its use under the FLM’s guidance
grid models may be applied for the The current version of the Guideline regarding AQRV assessments (FLAG
assessment of visibility impacts due to does not contain any explicit 2010) nor previous use of this model as
one or a small group of sources. Skill in recommendation regarding the use of part of regulatory modeling applications
estimating source-receptor relationships CALPUFF in the regional haze program, required under the CAA. Similarly, this
on this scale can be illustrated by but in advising states and in making its proposed action does not affect EPA’s
evaluating modeling systems against own BART determinations, the EPA has recommendation that States use
regional scale inert tracer release looked to the Guideline to resolve CALPUFF to determine the applicability
experiments. Historically, several questions regarding the proper and level of BART in regional haze
regional tracer release experiments have application of the model. In particular, implementation plans.43
been used to demonstrate skill in long- the EPA has guided states to use the
applicable regulatory version of 8. Role of EPA’s Model Clearinghouse
range transport of inert pollutants: 1980
Great Plains Mesoscale Tracer Field CALPUFF for such assessments. The EPA’s Model Clearinghouse has
Experiment, the 1983 Cross- Following the EPA’s recommendations, been a fundamental aspect of
Appalachian Tracer Experiment states have used the EPA-preferred communication between the EPA
(CAPTEX), the 1987 Across North version of CALPUFF in hundreds of Region Offices and with the broader
American Tracer Experiment BART determinations since 2005. permitting community on technical
(ANATEX), and 1994 European Tracer Although most assessments of BART are modeling and compliance
Experiment (ETEX).41 42 Photochemical now complete, a handful of BART demonstration issues for almost three
grid models have been shown to determinations remain outstanding. We decades. The Model Clearinghouse
demonstrate similar skill to Lagrangian expect most of the remaining actions serves a critical role in helping resolve
models for pollutant transport when addressing the BART requirements to be issues that arise from unique situations
compared to measurements made from completed within the next two years. that are not specifically addressed in the
multiple mesoscale field experiments.41 The proposed changes to the Guideline or necessitate the
Use of CTMs for Air Quality Related Guideline do not affect the EPA’s consideration of an alternative model or
Values (AQRV) analysis requirements, recommendation in the 2005 BART technique for a specific application or
while not subject to specific EPA model Guidelines to use CALPUFF in the range of applications. The Model
approval requirements outlined in 40 BART determination process. Given that Clearinghouse ensures that fairness,
CFR 51.166(l)(2) and 40 CFR 52.21(l)(2), the overwhelming majority of BART consistency, and transparency in
should be justified for each application determinations have been made using modeling decisions are fostered among
following the general recommendations CALPUFF, we consider it appropriate the Regional Offices and the state, local,
outlined in section 3.2, and concurrence for states (or the EPA) to continue to use and tribal agencies.
sought with the affected FLM(s). this application for the remaining In this action, we are proposing to
In 2005, the EPA issued guidelines for assessments under the current Guideline codify the long-standing process of the
implementation of the best available with approved protocols. This approach Regional Offices consulting and
retrofit technology (BART) requirements assures consistency across and within coordinating with the Model
under the Regional Haze Rule. In these states in the regional haze program. In Clearinghouse on all approvals of
BART Guidelines, the EPA addressed addition, in many instances, the alternative models or techniques. While
the question of how states could best modeling of visibility impacts has the Regional Administrators are the
predict a single source’s contribution to already been completed even though the delegated authority to issue such
visibility impairment.43 At the time, the BART determination process is not yet approvals under section 3.2 of the
EPA recognized that CALPUFF had not done. Allowing states to continue to rely Guideline, all alternative model
yet been fully evaluated for secondary on CALPUFF avoids additional time approvals will only be issued after
pollutant formation, but the EPA still and expense in developing a new consultation with the EPA’s Model
considered CALPUFF to be the best assessment of visibility impacts for a Clearinghouse and formal
application for assessing a single SIP initially due in 2007. We intend to documentation through a concurrence
source’s impact on visibility in a Class continue to advise states with respect to memorandum which demonstrates that
I area for purposes of the regional haze the EPA-preferred version of CALPUFF the requirements within section 3.2 for
program. The EPA took note of the that should be used in specific BART use of an alternative model have been
limitations of CALPUFF for this purpose cases. Consistent with the BART met.
but concluded that CALPUFF was the Guidelines, states may also use We invite comment on our proposal
best modeling application for use in alternative modeling approaches, in to codify existing practice of requiring
evaluating BART, especially given how consultation with the appropriate EPA consultation and coordination between
the modeling results would be used. Regional Office. the EPA Regional Offices and the EPA’s
Based on this assessment, the EPA The EPA is seeking comment on its Model Clearinghouse on all approvals
recommended that the states use proposed screening approach to address under section 3.2 of alternative models
CALPUFF. The EPA also made clear, long-range transport for purposes of or techniques.
however, that states could use other assessing PSD increments; its decision
to remove CALPUFF as a preferred 9. Updates to Modeling Procedures for
alternative approaches, including Cumulative Impact Analysis
photochemical grid models, if done in model in appendix A for such long-

range transport assessments; and its Based on input from the Tenth
42 Hegarty, J., Draxler, R.R., Stein, A.F., Brioude, decision to consider CALPUFF as a Modeling Conference and recent permit
J., Mountain, M., Eluszkiewicz, J., Nehrkorn, T., screening technique along with other modeling experiences under new short-
Ngan, F., Andrews, A., 2013. Evaluation of Lagrangian models to be used in term NAAQS for SO2 and NO2, the EPA
Lagrangian particle dispersion models with consultation with the appropriate is proposing to make modifications to
measurements from controlled tracer releases.
Journal of Applied Meteorology and Climatology, reviewing authority. It is important to section 8 of the Guideline regarding
52: 2623–2637. note that the EPA’s proposed action to model inputs and background
43 See 70 FR 39104, 39122–23 (July 6, 2005). remove CALPUFF as an appendix A concentrations to provide much needed
clarity associated with input and demonstrations for ozone, PM2.5, and establishing the model inputs for
database selection for use in PSD and regional haze (see section 8.2). conducting the regulatory modeling for
SIP modeling. Many of these revisions • Revised recommendations on how PSD and SIP applications.
are based on the EPA clarification to determine background concentrations
10. Updates on Use of Meteorological
memoranda issued since 2010 that were in constructing the design Input Data for Regulatory Dispersion
intended to provide the necessary concentration, or total air quality Modeling
clarification regarding applicability of concentration, as part of a cumulative
the Guideline to PSD modeling for these impact analysis for NAAQS and PSD For near-field dispersion modeling
new standards.1 2 44 45 The EPA has increments. Specific recommendations applications using National Weather
specifically cautioned against the literal are proposed for situations involving Service (NWS) Automated Surface
and uncritical application of very isolated single-source(s) and multi- Observing Stations (ASOS), the EPA
prescriptive procedures for conducting source areas (see section 8.3) with an released a pre-processor to AERMET,
NAAQS and PSD modeling compliance emphasis on how to determine which called AERMINUTE, in 2011 that
demonstrations as described in chapter nearby sources to explicitly model calculates hourly averaged winds from
C of the draft New Source Review based on the concept of significant 2-minute winds reported every minute
Workshop Manual.46 Our main concern concentration gradients and the use of at NWS ASOS sites. AERMET
is that following such procedures in a monitored background to adequately substitutes these hourly averaged winds
literal and uncritical manner has led to represent ‘‘other sources’’ (i.e., that for the standard hourly observations,
practices that are overly conservative portion of the background attributable to thus reducing the number of calms and
and unnecessarily complicate the natural sources, other unidentified missing winds for input into AERMOD.
permitting process. The proposed sources in the vicinity of the project, The presence of calms and missing
changes to section 8 are intended to and regional transport contributions winds were due to the METAR reporting
modify these practices and provide a from more distant sources (domestic methodology of surface observations. In
more appropriate basis for selection and and international)). It is important to March 2013, the EPA released a
use of modeling inputs through the note the interconnectedness of these memorandum regarding the use of
Guideline itself and supporting issues as the question of which nearby ASOS data in AERMOD as well as the
guidance. sources to include in cumulative use of AERMINUTE. When using
modeling is inextricably linked with the meteorological data from ASOS sites for
We have provided a more definitive
question of what ambient monitoring input to AERMOD, hourly averaged
definition of the appropriate modeling
data are available and what these data winds from AERMINUTE should be
domain and how to best characterize the
represent for a specific application. used in most cases.
various contributions to air quality For a near-field dispersion modeling
concentrations within that domain. More specific data requirements and the application where there is no
Specifically, we provide the following format required for the individual representative NWS station, and it is
recommendations: models are described in detail in the prohibitive or not feasible to collect
• Definition and/or factors to users’ guide and/or associated adequately representative site-specific
consider in determining appropriate documentation for each model. data, it may be necessary to use
modeling domain for NAAQS and PSD Given the added complexity of the prognostic meteorological data for the
increment assessments and for SIP technical issues that arise in the context application. The EPA released the
attainment demonstrations (see section of demonstrating compliance with MMIF program that converts the
8.1). NAAQS through dispersion modeling, prognostic meteorological data into a
• Revised requirements on how to we strongly encourage adherence to the format suitable for dispersion modeling
characterize emissions from nearby recommendations in section 9.2.1 of the applications. The most recent 3 years of
sources to be explicitly modeled for proposed Guideline regarding prognostic data are preferred. Use of the
purposes of a cumulative impact development of a modeling protocol, prognostic data is contingent on the
assessment under PSD and new i.e., that ‘‘[e]very effort should be made concurrence of the appropriate
language regarding how to characterize by the Regional Office to meet with all reviewing authorities and collaborating
direct and precursor emissions from parties involved in either a SIP revision agencies that the data are of acceptable
modeled sources for SIP attainment or a PSD permit application prior to the quality and representative of the
start of any work on such a project. modeling application.
44 U.S. EPA, 2011. Additional Clarification During this meeting, a protocol should We solicit comments on our proposed
Regarding Applicability of Appendix W Modeling be established between the preparing updates regarding use of meteorological
Guidance for the 1-hour NO2 NAAQS. Tyler Fox and reviewing parties to define the
Memorandum dated March 1, 2011, Office of Air
input data for regulatory application of
Quality Planning & Standards, Research Triangle
procedures to be followed, the data to be dispersion models.
Park, North Carolina 27711.http://www.epa.gov/ttn/ collected, the model to be used, and the
scram/guidance/clarification/NO2_Clarification_ analysis of the source and concentration 11. Transition Period for Applicability
Memo-20140930.pdf. data.’’ We expect by providing more of Revisions to the Guideline
45 U.S. EPA, 2014. Clarification on the Use of
clarity in the Guideline of the factors to In previous rulemakings to revise the
AERMOD Dispersion Modeling for Demonstrating
Compliance with the NO2 National Ambient Air be considered in the cumulative impact Guideline, we have traditionally
Quality Standard. R. Chris Owen and Roger Brode assessment, permit applicants and communicated that it would be
Memorandum dated September 30, 2014, Office of permitting authorities will be able to appropriate to provide 1 year to
Air Quality Planning & Standards, Research find the proper balance of the transition to the use of new models,
Triangle Park, North Carolina 27711. http://
www.epa.gov/ttn/scram/guidance/clarification/ competing factors that contribute to techniques and procedures in the
NO2_Clarification_Memo-20140930.pdf. these analyses. context of PSD permit applications and
46 U.S. EPA, 1990. New Source Review Workshop We invite comments on whether the other regulatory modeling applications.
Manual: Prevention of Significant Deterioration and updates proposed in section 8 of the We invite comments whether it would
Nonattainment Area Permitting (Draft). Office of Air
Quality Planning & Standards, Research Triangle
Guideline and associated guidance are be appropriate to apply a 1-year
Park, North Carolina 27711. http://www.epa.gov/ appropriate and sufficient to provide the transition after promulgation of the
nsr. necessary clarification in selecting and revised Guideline (i.e., from its effective
date) such that applications conducted additional clarity. Additional criteria pollutants of CO, lead, SO2,
under the current Guideline with information is provided regarding the NO2, and primary PM2.5 and PM10. In
approved protocols would be acceptable importance of CAA section 320 to many respects, the proposed revisions
during that period, but new amendments of the Guideline. to section 4 are a combination of the
requirements and recommendations previous sections 4 and 5, reflecting
3. Section 2
should be used for applications inert criteria pollutants only. The EPA
submitted after that period or protocols The EPA proposes to revise section 2 also proposes to modify section 4 to
approved after that period. to more appropriately discuss the incorporate requirement subsections to
The EPA believes such a transition process by which models are evaluated provide clarity of the various
period is appropriate to avoid the time and considered for use in particular requirements where previously sections
and expense of revisiting modeling that applications. We propose to incorporate 4 and 5 included various requirements
is substantially complete, which would information from the previous section 9 under recommendation subheadings.
cause undue delays to permit pertaining to model accuracy and As proposed, this section provides an
applications that are pending when the uncertainty within this section to clarify in-depth discussion of screening and
proposed revisions to the Guideline are how model performance evaluation is refined models, including the
finalized. The revisions that the EPA is critical in determining the suitability of introduction of AERSCREEN as the
proposing to the Guideline are intended models for particular application. recommended screening model for
as incremental improvements to the We also propose to provide a simple and complex terrain for single
Guideline, and such improvements do discussion in section 2.1 (Model sources and options for multi-source
not necessarily invalidate past practices Accuracy and Uncertainty) of the three screening with AERMOD.47 The EPA
under the previous edition of the types of models historically used for proposes to include a clear discussion of
Guideline. The requirements and regulatory demonstrations. For each each appendix A preferred model in
recommendations in the current (2005) type of model, some strengths and section 4.3 (Refined Models). The EPA
version of the Guideline were weaknesses are listed to assist readers in also proposes to modify the discussion
previously identified as acceptable by the understanding of the particular for each preferred model (i.e., AERMOD
the EPA, and they will continue to be regulatory applications to which they Modeling System, CTDMPLUS, and
acceptable for air quality assessments are most appropriate. OCD) from the previous section 4 with
during the period of transition to the In addition, the EPA proposes appropriate edits and some streamlining
revised version of the Guideline. revisions to section 2.2 with respect to based on information available in the
Where a proposed revision to the the recommended practice of respective model formulation
Guideline does raise questions about the progressing from simplified and documentation and users guides.
acceptability of a requirement or conservative air quality analysis toward The EPA is proposing to add a
recommendation that it replaces, model more complex and refined analysis. In subsection specifically addressing the
users and applicants are encouraged to this section, the EPA proposes to clarify modeling recommendations for SO2
consult with the appropriate reviewing distinctions between various types of where, previously, section 4 of the
authority as soon as possible to assure models that have previously been Guideline was generally understood to
the acceptability of modeling used to described as screening models. In be applicable for SO2. Minor updates are
support permit applications during this addition, this section clarifies proposed with respect to the modeling
period. distinctions between models used for recommendations for each of the other
screening purposes and screening inert criteria pollutants that were
B. Proposed Editorial Changes techniques and demonstration tools that previously found in section 5. For NO2,
The EPA is proposing to make may be acceptable in certain the ARM2 is proposed to be added as a
editorial changes to update and applications. Tier 2 option, and the Tier 3 options of
reorganize information throughout the OLM and PVMRM are proposed to
4. Section 3
Guideline. These revisions are not become part of the regulatory version of
intended to meaningfully change the The EPA proposes minor AERMOD. For any pollutant that had
substance of the Guideline, but rather to modifications to section 3 to more significant emissions from mobile
make the Guideline easier to use. One accurately reflect current EPA practices sources, our previous recommendation
way this is accomplished is by grouping and by moving the discussion of the to use the CALINE3 models is proposed
topics together in a more logical manner EPA’s Model Clearinghouse to a revised to be replaced with AERMOD.
to make related content easier to find. section 3.3 for ease of reference and
prominence within the Guideline. A 6. Section 5
This in turn should streamline the
compliance assessment process. change is proposed to require Regional As already stated, much of the
Editorial changes are described below Office consultation with the Model previous section 5 with respect to the
for each affected section. We invite Clearinghouse on all alternative model inert criteria pollutants is proposed to
comment on any of the changes approvals. Previously, section 3 be incorporated into the revised section
proposed below for the Guideline text. included various requirements under 4. As proposed, the revised section 5 is
recommendation subheading that were now focused only on the modeling
1. Preface not clearly identified as requirements. approaches recommended for ozone and
Only a few minor text revisions are Accordingly, the EPA is proposing to secondary PM2.5.
proposed to this section for consistency modify section 3 with the incorporation Both ozone and secondary PM2.5 are
with the remainder of the Guideline. of requirement subsections to eliminate formed through chemical reactions in
any ambiguity. the atmosphere and are not
2. Section 1
The EPA propose to update the 5. Section 4 47 U.S. EPA, 2015. Technical Support Document
introduction section to reflect the The EPA proposes to significantly (TSD) for Replacement of CALINE3 with AERMOD
for Transportation Related Air Quality Analyses.
reorganized nature of the revised revise section 4 to incorporate the Publication No. EPA–454/B–15–002. Office of Air
Guideline. Minor text revisions are modeling approaches recommended for Quality Planning & Standards, Research Triangle
proposed throughout this section for air quality impact analyses for the Park, NC.
appropriately modeled with traditional 8. Section 7 and other sources. This is intended to
steady-state Gaussian plume models, We propose to revise section 7 to be eliminate confusion of how to identify
such as AERMOD. Chemical transport more streamlined and appropriate to the nearby sources that should be explicitly
models are necessary to appropriately variety of general modeling issues and modeled and all other sources that
assess the single-source air quality considerations that are not already been should be generally represented by air
impacts of precursor pollutants on the covered in sections 4, 5, and 6 of the quality monitoring data. In addition to
formation of ozone or secondary PM2.5. Guideline. The EPA proposes to move air quality monitoring data, a brief
the information concerning design discussion on the use of photochemical
While the proposed revision to
concentrations and receptor sites to grid modeling to appropriately
section 5 do not specify a particular
section 9. The discussion of stability characterize background concentrations
EPA-preferred model or technique for
categories is proposed to be removed has been included in this proposed
use in air quality assessments, a two-
from section 7 since it is specifically section. Updates to Tables 8–1 and 8–
tiered screening approach is proposed
addressed in the model formulation 2 are proposed per changes in the
for ozone and secondary PM2.5 with considerations for nearby sources, as
appropriate references to the EPA’s new documentation and guidance for the
dispersion models that require stability discussed in the Proposed Actions
single-source modeling guidance. The section of this Preamble.
first tier consists of technically credible categories to be defined. As already
stated, the GEP discussion from the The use of prognostic mesoscale
and appropriate relationships between
previous section 6 is proposed to be meteorological models to provide
emissions and the impacts developed meteorological input for regulatory
from existing modeling simulations. If incorporated into this section.
The EPA proposes to expand the dispersion modeling applications is
existing technical information is not proposed to be incorporated throughout
available or appropriate, then a second recommendations for determining rural
or urban dispersion coefficients to the Meteorological Input Data
tier approach would apply, involving subsection, including the introduction
use of sophisticated chemical transport provide more clarity with respect to
appropriate characterization within of the MMIF as a tool to inform
models (e.g., photochemical grid regulatory model applications. Other
AERMOD, including a discussion on the
models) as determined in consultation than additional minor modifications to
existence of highly industrialized areas
with the appropriate EPA Regional the recommendations through this
where population density is low that
Office on a case-by-case basis based subsection based on current EPA
may be best treated with urban rather
upon the EPA’s new single-source practices, the most substantive proposed
than rural dispersion coefficients.
modeling guidance. edits relate to the recommendation to
References to CALPUFF in the Complex
7. Section 6 Winds subsection are proposed to be use the AERMINUTE meteorological
removed due to technical issues data processor to calculate hourly
Revisions to section 6 are proposed to described in the Proposed Actions average wind speed and direction when
more clearly address the modeling section of this preamble. As proposed, processing NWS ASOS data for
recommendations of other federal if necessary for special complex wind developing AERMET meteorological
agencies, such as the FLM(s), that have situations, the setup and application of inputs to the AERMOD dispersion
been developed in response to EPA an alternative model should now be model.
rules or standards. While no attempt is determined in consultation with the 10. Section 9
made to comprehensively discuss each appropriate reviewing authority.
topic, the EPA proposes to provide Finally, the EPA proposes to revise The EPA proposes to move all of the
appropriate references to the respective section 7 to include a new discussion of information previously in section 9
federal agency guidance documents. modeling considerations specific to related to model accuracy and
mobile sources. evaluation into other sections in the
The proposed revision to section 6 revised Guideline (primarily to the
focus primarily on AQRVs, including 9. Section 8 revised section 2 and some to the
near-field and long-range transport revised section 4). This provides for
The EPA propose extensive updates
assessments for visibility impairment greater clarity in those topics as applied
and modifications to section 8 to reflect
and deposition. The interests of the to selection of models under the
current EPA practices, requirements,
Bureau of Ocean Energy and Guideline. However, the EPA proposes
and recommendations for determining
Management for Outer Continental Shelf to remove subsection on the ‘‘Use of
the appropriate modeling domain and
permitting situations and of the Federal Uncertainty in Decision Making.’’. After
model input data from new or
Aviation Administration for airport and removing this content, the EPA
modifying source(s) or sources under
air base permitting situations are proposes to totally revise section 9 to
consideration for a revised permit limit,
represented in proposed section 6.3 focus on the regulatory application of
from background concentrations
(Modeling Guidance for Other models, which would include the
(including air quality monitoring data
Governmental Programs). and nearby and others sources), and majority of the information found
The discussion of Good Engineering from meteorology. As with earlier previously in section 10.
Practices (GEP) for stack height sections, the EPA proposes to modify The EPA proposes to revise the
consideration is proposed to be section 8 to incorporate requirement discussion portion of section 9 to more
modified and moved to section 7. The subsections where previously section 8 clearly summarize the general concepts
EPA proposed to remove the discussion ambiguously included various presented in earlier sections of the
of long-range transport for PSD Class I requirements under recommendation Guideline and to set the stage for the
increment and references to the subheadings. appropriate regulatory application of
previously preferred long-range The Background Concentration models and/or, in rare circumstances,
transport model, CALPUFF, in subsection is proposed to be air quality monitoring data. The
accordance with the more detailed significantly modified from the existing importance of developing and vetting a
discussion in the Proposed Actions Guideline to include a more clear and modeling protocol is more prominently
section of this Preamble. comprehensive discussion of nearby presented in a separate subsection.
The information related to design CALPUFF as refined air quality models techniques) on these entities. The
concentrations is proposed to be preferred for specific regulator proposal features updates to the existing
updated and unified from previous applications. The rational for the EPA-preferred model, AERMOD, that
language found in sections 7 and 10. An removal of these air quality models from serves to increase efficiency and
expanded discussion of receptor sites is the preferred status can be found in the accuracy by changing only
proposed based on language from the Proposed Actions section of this mathematical formulations and specific
previous section 7 and new Preamble. data elements. Also, this proposed
considerations given past practices of action will streamline resources
V. Statutory and Executive Order necessary to conduct necessary
model users tending to define an
Reviews modeling with AERMOD by
excessively large and inappropriate
number of receptors based on vague A. Executive Order 12866: Regulatory incorporating model algorithms from
guidance. Planning and Review and Executive the BLP model and replacing CALINE3
The recommendations for NAAQS Order 13563: Improving Regulation and for mobile source applications.
and PSD increment compliance Regulatory Review Although this proposed action calls for
demonstrations are proposed to be new models and/or techniques for use
This proposed action is not a in addressing ozone and secondary
overhauled to more clearly and
‘‘significant regulatory action’’ under PM2.5, we expect most small entities
accurately reflect the long-standing EPA
the terms of Executive Order 12866 (58 will generally be able to rely on existing
recommendation and practice of
FR 51735, October 4, 1993) and is, modeling simulations; so, we expect
performing a single-source impact
therefore not subject to OMB review minimal burden associated with these
analysis as a first stage of the NAAQS
under Executive Orders 12866 and assessments. Therefore, we do not
and PSD increment compliance
13563 (76 FR 3821, January 21, 2011). believe that that this proposal poses a
demonstration and, as necessary,
conducting a more comprehensive B. Paperwork Reduction Act significant or unreasonable burden on
cumulative impact analysis as the any small entities.
This proposed action does not impose After considering the economic
second stage. The appropriate an information collection burden
considerations and applications of impacts of this rule on small entities, I
subject to OMB review under the certify that this action will not have a
screening and/or refined model are provisions of the Paperwork Reduction
described in each stage. significant economic impact on a
Act, 44 U.S.C. 3501 et seq. substantial number of small entities. We
Finally, the section on Use of
Measured Data in Lieu of Model C. Regulatory Flexibility Act continue to be interested in the
Estimates subsection is proposed to be potential impacts of the proposed rule
The Regulatory Flexibility Act (RFA) on small entities and welcome
revised to provide more details on the generally requires an agency to prepare
process for determining the rare comments on issues related to such
a regulatory flexibility analysis of any impacts.
circumstances in which air quality rule subject to notice and comment
monitoring data may be considered for rulemaking requirements under the D. Unfunded Mandates Reform Act
determining the most appropriate Administrative Procedure Act or any This proposed action contains no
emissions limit for a modification to an other statute unless the agency certifies federal mandates under the provisions
existing source. As with other portions that the rule will not have a significant of Title II of the Unfunded Mandates
of the revised section 9, the language economic impact on a substantial Reform Act of 1995 (UMRA), 2 U.S.C.
throughout this subsection is proposed number of small entities. Small entities 1531–1538 for state, local, or tribal
to be updated to reflect current EPA include small businesses, small governments or the private sector. This
practices, as appropriate. organizations, and small governmental action imposes no enforceable duty on
11. Section 10 jurisdictions. any state, local or tribal governments or
For purposes of assessing the impacts the private sector. Therefore, this action
As discussed, the majority of the
of this rule on small entities, small is not subject to the requirements of
information found previously in section
entity is defined as (1) a small business sections 202 or 205 of the UMRA. This
10 is proposed to be incorporated into
as defined by the Small Business action is also not subject to the
the revised section 9. As proposed,
Administration’s (SBA) regulations at 13 requirements of section 203 of UMRA
section 10 consists of the references that
CFR 121.201; (2) a small governmental because it contains no regulatory
were in the previous section 12. We also
jurisdiction that is a government of a requirements that might significantly or
propose to update each reference, as
city, county, town, school district or uniquely affect small governments.
appropriate, based on the text revisions
special district with a population of less
throughout the Guideline. E. Executive Order 13132: Federalism
than 50,000; and (3) a small
12. Section 11 organization that is any not-for-profit This proposed action does not have
In a streamlining effort, the EPA enterprise which is independently federalism implications. It will not have
proposes to remove this bibliography owned and operated and is not substantial direct effects on the states,
section from the Guideline. dominant in its field. on the relationship between the national
The modeling techniques described in government and the states, or on the
13. Section 12 this proposed action are primarily used distribution of power and
As stated earlier, this references by air agencies and by industries responsibilities among the various
section is now proposed as section 10 owning major sources subject to NSR levels of government, as specified in
with appropriate updates. permitting requirements. To the extent Executive Order 13132. This rule does
that any small entities would have to not create a mandate on state, local or
14. Appendix A to the Guideline conduct air quality assessments, using tribal governments nor does it impose
The EPA proposes to revise appendix the models and/or techniques described any enforceable duties on these entities.
A to the Guideline to remove the in this proposed action are not expect to This action would add better, more
Buoyant Line and Point Source pose any additional burden (compared accurate techniques for conducting air
Dispersion Model (BLP), CALINE3, and to the existing models and/or quality assessments and does not add
any additional requirements for any of because it does not affect the level of Guideline. The third activity is the extensive
the affected parties covered under protection provided to human health or on-going research efforts by the EPA and
Executive Order 13132. Thus, the the environment. others in air quality and meteorological
modeling.
requirements of section 6 of the
List of Subjects in 40 CFR Part 51 c. Based primarily on these three activities,
Executive Order do not apply to this new sections and topics have been included
proposal. In the spirit of Executive Environmental protection, as needed. The EPA does not make changes
Order 13132, and consistent with the Administrative practice and procedure, to the guidance on a predetermined schedule,
EPA policy to promote communications Air pollution control, Carbon monoxide, but rather on an as-needed basis. The EPA
between the EPA and state and local Intergovernmental relations, Nitrogen believes that revisions of the Guideline
governments, the EPA specifically oxides, Ozone, Particulate Matter, should be timely and responsive to user
solicits comment on this proposed rule Reporting and recordkeeping needs and should involve public
requirements, Sulfur oxides. participation to the greatest possible extent.
from state and local officials.
All future changes to the guidance will be
F. Executive Order 13175: Consultation Dated: July 14, 2015. proposed and finalized in the Federal
and Coordination With Indian Tribal Gina McCarthy, Register. Information on the current status of
Governments Administrator. modeling guidance can always be obtained
from EPA’s Regional Offices.
This proposed action does not have For the reasons stated in the
preamble, title 40, chapter I of the Code Table of Contents
tribal implications, as specified in
Executive Order 13175 (65 FR 67249, of Federal Regulations is proposed to be List of Tables
November 9, 2000). This proposed rule amended as follows: 1.0 Introduction
imposes no requirements on tribal 2.0 Overview of Model Use
governments. Accordingly, Executive PART 51—REQUIREMENTS FOR 2.1 Suitability of Models
Order 13175 does not apply to this PREPARATION, ADOPTION, AND 2.1.1 Model Accuracy and Uncertainty
SUBMITTAL OF IMPLEMENTATION 2.2 Levels of Sophistication of Air Quality
action. In the spirit of Executive Order Analyses and Models
13175, the EPA specifically solicits PLANS
2.3 Availability of Models
additional comment on this proposed 3.0 Preferred and Alternative Air Quality
■ 1. The authority citation for part 51
action from tribal officials. Models
continues to read as follows:
3.1 Preferred Models
G. Executive Order 13045: Protection of Authority: 23 U.S.C. 101; 42 U.S.C. 7401– 3.1.1 Discussion
Children From Environmental Health 7671q. 3.1.2 Requirements
and Safety Risks 3.2 Alternative Models
■ 2. Appendix W to part 51 is revised
The EPA interprets Executive Order 3.2.1 Discussion
to read as follows: 3.2.2 Requirements
13045 as applying only to those
APPENDIX W TO PART 51—Guideline 3.3 EPA’s Model Clearinghouse
regulatory actions that concern 4.0 Models for Carbon Monoxide, Lead,
environmental health or safety risks that on Air Quality Models Preface
Sulfur Dioxide, Nitrogen Dioxide and
the EPA has reason to believe may a. Industry and control agencies have long Primary Particulate Matter
disproportionately affect children, per expressed a need for consistency in the 4.1 Discussion
the definition of ‘‘covered regulatory application of air quality models for 4.2 Requirements
action’’ in section 2–202 of the regulatory purposes. In the 1977 Clean Air 4.2.1 Screening Models and Techniques
Executive Order. This action is not Act (CAA), Congress mandated such 4.2.1.1 AERSCREEN
subject to Executive Order 13045 consistency and encouraged the 4.2.1.2 CTSCREEN
standardization of model applications. The 4.2.1.3 Screening in Complex Terrain
because it does not concern an Guideline on Air Quality Models (hereafter, 4.2.2 Refined Models
environmental health risk or safety risk. Guideline) was first published in April 1978 4.2.2.1 AERMOD
H. Executive Order 13211: Actions to satisfy these requirements by specifying 4.2.2.2 CTDMPLUS
Concerning Regulations That models and providing guidance for their use. 4.2.2.3 OCD
The Guideline provides a common basis for 4.2.3 Pollutant Specific Modeling
Significantly Affect Energy Supply, estimating the air quality concentrations of Requirements
Distribution, or Use criteria pollutants used in assessing control 4.2.3.1 Models for Carbon Monoxide
This action is not a ‘‘significant strategies and developing emissions limits. 4.2.3.2 Models for Lead
energy action’’ as defined in Executive b. The continuing development of new air 4.2.3.3 Models for Sulfur Dioxide
Order 13211 (66 FR 28355 (May 22, quality models in response to regulatory 4.2.3.4 Models for Nitrogen Dioxide
requirements and the expanded requirements 4.2.3.5 Models for PM2.5
2001)), because it is not likely to have
for models to cover even more complex 4.2.3.6 Models for PM10
a significant adverse effect on the problems have emphasized the need for 5.0 Models for Ozone and Secondarily
supply, distribution, or use of energy. periodic review and update of guidance on Formed Particulate Matter
I. National Technology Transfer and these techniques. Historically, three primary 5.1 Discussion
activities have provided direct input to 5.2 Recommendations
Advancement Act
revisions of the Guideline. The first is a series 5.3 Recommended Models and Approaches
This rulemaking does not involve of periodic EPA workshops and modeling for Ozone
technical standards. conferences conducted for the purpose of 5.3.1 Models for NAAQS Attainment
ensuring consistency and providing Demonstrations and Multi-Source Air
J. Executive Order 12898: Federal
clarification in the application of models. Quality Assessments

Actions To Address Environmental The second activity was the solicitation and 5.3.2 Models for Single-Source Air
Justice in Minority Populations and review of new models from the technical and Quality Assessments
Low-Income Populations user community. In the March 27, 1980, 5.4 Recommended Models and Approaches
Federal Register, a procedure was outlined for Secondarily Formed PM2.5
The EPA has determined that this for the submittal to the EPA of privately 5.4.1 Models for NAAQS Attainment
proposed rule will not have developed models. After extensive evaluation Demonstrations and Multi-Source Air
disproportionately high and adverse and scientific review, these models, as well Quality Assessments
human health or environmental effects as those made available by the EPA, have 5.4.2 Models for Single-Source Air
on minority or low-income populations been considered for recognition in the Quality Assessments
6.0 Modeling for Air Quality Related Values LIST OF TABLES frequently required. As modeling efforts
and Other Governmental Programs become more complex, it is increasingly
6.1 Discussion Table No. Title important that they be directed by highly
6.2 Air Quality Related Values competent individuals with a broad range of
6.2.1 Visibility 8–1 ......... Point Source Model Emission experience and knowledge in air quality
6.2.1.1 Models for Estimating Near-Field Input for SIP Revisions of Inert meteorology. Further, they should be
Pollutants. coordinated closely with specialists in
Visibility Impairment
8–2 ......... Point Source Model Emission emissions characteristics, air monitoring and
6.2.1.2 Models for Estimating Visibility data processing. The judgment of
Impairment for Long-Range Transport Input for NAAQS Compliance in
PSD Demonstrations. experienced meteorologists, atmospheric
6.2.2 Models for Estimating Deposition scientists, and analysts is essential.
Impacts d. The model that most accurately
6.3 Modeling Guidance for Other 1.0 Introduction estimates concentrations in the area of
Governmental Programs a. The Guideline recommends air quality interest is always sought. However, it is clear
7.0 General Modeling Considerations modeling techniques that should be applied from the needs expressed by the EPA
7.1 Discussion to State Implementation Plan (SIP) submittals Regional Offices, by state, local, and tribal
7.2 Recommendations and revisions, to New Source Review (NSR), agencies, by many industries and trade
7.2.1 All sources including new or modifying sources under associations, and also by the deliberations of
7.2.1.1 Dispersion Coefficients Prevention of Significant Deterioration Congress that consistency in the selection
7.2.1.2 Complex Winds (PSD),1 2 3 conformity analyses,4 and other air and application of models and databases
quality assessments required under EPA should also be sought, even in case-by-case
7.2.1.3 Gravitational Settling and
regulation. Applicable only to criteria air analyses. Consistency ensures that air quality
Deposition control agencies and the general public have
7.2.2 Stationary Sources pollutants, the Guideline is intended for use
by the EPA Regional Offices in judging the a common basis for estimating pollutant
7.2.2.1 Good Engineering Practice Stack concentrations, assessing control strategies,
adequacy of modeling analyses performed by
Height and specifying emissions limits. Such
the EPA, by state, local, and tribal permitting
7.2.2.2 Plume Rise authorities, and by industry. It is appropriate consistency is not, however, promoted at the
7.2.3 Mobile Sources for use by other federal government agencies expense of model and database accuracy. The
8.0 Model Input Data and by state, local, and tribal agencies with Guideline provides a consistent basis for
8.1 Modeling Domain air quality and land management selection of the most accurate models and
8.1.1 Discussion responsibilities. The Guideline serves to databases for use in air quality assessments.
8.1.2 Requirements identify, for all interested parties, those e. Recommendations are made in the
8.2 Source Data Guideline concerning air quality models and
modeling techniques and databases that the
8.2.1 Discussion techniques, model evaluation procedures,
EPA considers acceptable. The Guideline is
and model input databases and related
8.2.2 Requirements not intended to be a compendium of
requirements. The guidance provided here
8.3 Background Concentrations modeling techniques. Rather, it should serve
should be followed in air quality analyses
8.3.1 Discussion as a common measure of acceptable technical
relative to SIPs, NSR, and in supporting
8.3.2 Recommendations for Isolated analysis when supported by sound scientific
analyses required by the EPA and by state,
Single Source judgment.
local, and tribal permitting authorities.
8.3.3 Recommendations for Multi-Source b. Air quality measurements 5 are routinely Specific models are identified for particular
Areas used to characterize ambient concentrations applications. The EPA may approve the use
8.4 Meteorological Input Data of criteria pollutants throughout the nation of an alternative model or technique that can
but are rarely sufficient for characterizing the be demonstrated to be more appropriate than
8.4.1 Discussion
ambient impacts of individual sources or those recommended in the Guideline. In all
8.4.2 Recommendations and
demonstrating adequacy of emissions limits cases, the model or technique applied to a
Requirements for an existing source due to limitations in
8.4.3 National Weather Service Data given situation should be the one that
spatial and temporal coverage of ambient provides the most accurate representation of
8.4.3.1 Discussion monitoring networks. The impacts of new
8.4.3.2 Recommendations atmospheric transport, dispersion, and
sources that do not yet exist and chemical transformations in the area of
8.4.4 Site-specific data modifications to existing sources that have interest. However, to ensure consistency,
8.4.4.1 Discussion yet to be implemented can only be deviations from the Guideline should be
8.4.4.2 Recommendations determined through modeling. Thus, models carefully documented as part of the public
8.4.5 Prognostic meteorological data have become a primary analytical tool in record and fully supported by the
8.4.5.1 Discussion most air quality assessments. Air quality appropriate reviewing authority, as discussed
8.4.5.2 Recommendations measurements can be used in a later.
8.4.6 Treatment of Near-Calms and Calms complementary manner to air quality models, f. From time to time, situations arise
8.4.6.1 Discussion with due regard for the strengths and requiring clarification of the intent of the
8.4.6.2 Recommendations weaknesses of both analysis techniques, and guidance on a specific topic. Periodic
9.0 Regulatory Application of Models are particularly useful in assessing the workshops are held with EPA headquarters,
9.1 Discussion accuracy of model estimates. EPA Regional Office, and state, local, and
9.2 Recommendations c. It would be advantageous to categorize tribal agency modeling representatives to
9.2.1 Modeling Protocol the various regulatory programs and to apply ensure consistency in modeling guidance and
a designated model to each proposed source to promote the use of more accurate air
9.2.2 Design Concentration and Receptor
needing analysis under a given program. quality models, techniques, and databases.
Sites
However, the diversity of the nation’s The workshops serve to provide further
9.2.3 NAAQS and PSD Increments topography and climate, and variations in explanations of Guideline requirements to
Compliance Demonstrations for New or
source configurations and operating the EPA Regional Offices and workshop
Modified Sources characteristics dictate against a strict materials are issued with this clarifying
9.2.3.1 Considerations in Developing modeling ‘‘cookbook.’’ There is no one model information. In addition, findings from
Emissions Limits capable of properly addressing all ongoing research programs, new model
9.2.4 Use of Measured Data in Lieu of conceivable situations even within a broad development, or results from model
Model Estimates category such as point sources. evaluations and applications are
10.0 References Meteorological phenomena associated with continuously evaluated. Based on this
Appendix A to Appendix W of Part 51— threats to air quality standards are rarely information, changes in the applicable
Summaries of Preferred Air Quality amenable to a single mathematical treatment; guidance may be indicated and appropriate
Models thus, case-by-case analysis and judgment are revisions to the Guideline may be considered.
g. All changes to the Guideline must follow the detail and accuracy of the input model input conditions are changing over the
rulemaking requirements since the Guideline databases, i.e., emissions inventory, model domain and model time step.
is codified in appendix W to 40 Code of meteorological data, and air quality data; (3) Lagrangian models can also be used to
Federal Regulations (CFR) part 51. The EPA the manner in which complexities of determine near and far-field impacts from a
will promulgate proposed and final rules in atmospheric processes are handled in the limited number of sources at a high
the Federal Register to amend this appendix. model; (4) the technical competence of those resolution. Traditionally, Lagrangian models
The EPA utilizes the existing procedures undertaking such simulation modeling; and have been used for relatively inert pollutants,
under CAA section 320 that requires EPA to (5) the resources available to apply the with slightly more complex considerations of
conduct a Conference on Air Quality model. Any of these factors can have a removal than Gaussian models. Some
Modeling at least every 3 years. These significant influence on the overall model Lagrangian models treat in-plume gas and
modeling conferences are intended to performance, which must be thoroughly particulate chemistry. However, these models
develop standardized air quality modeling evaluated to determine the suitability of an require time and space varying concentration
procedures and form the basis for associated air quality model to a particular application fields of oxidants and, in the case of fine
revisions to this Guideline in support of the or range of applications. particulate matter (PM2.5), neutralizing
EPA’s continuing effort to prescribe with b. Air quality models are most accurate and agents, such as ammonia. Reliable
‘‘reasonable particularity’’ air quality models reliable in areas that have gradual transitions background fields are critical for applications
and meteorological and emission databases of land use and topography. Meteorological involving secondary pollutant formation
suitable for modeling National Ambient Air conditions in these areas are spatially because secondary impacts generally occur
Quality Standards (NAAQS) 6 and PSD uniform such that observations are broadly when in-plume precursors mix and react
increments (CAA 320, 42 U.S.C. 7620). representative and air quality model with species in the background
Ample opportunity for public comment will projections are not further complicated by a atmosphere.7 8 These oxidant and
be provided for each proposed change and heterogeneous environment. Areas subject to neutralizing agents are not routinely
public hearings scheduled. major topographic influences experience measured, but can be generated with a three-
h. A wide range of topics on modeling and meteorological complexities that are often dimensional photochemical grid model.
databases are discussed in the Guideline. difficult to measure and simulate. Models iii. Photochemical grid models are three-
Section 2 gives an overview of models and with adequate performance are available for dimensional Eulerian grid-based models that
their suitability for use in regulatory increasingly complex environments. treat chemical and physical processes in each
applications. Section 3 provides specific However, they are resource intensive and grid cell and use diffusion and transport
guidance on the determination of preferred frequently require site-specific observations processes to move chemical species between
air quality models and on the selection of and formulations. Such complexities and the grid cells.9 Eulerian models assume that
alternative models or techniques. Sections 4 related challenges for the air quality emissions are spread evenly throughout each
through 6 provide recommendations on simulation should be considered when model grid cell. Typically, Eulerian models
modeling techniques for assessing criteria selecting the most appropriate air quality have difficulty with fine scale resolution of
pollutant impacts from single and multiple model for an application. individual plumes. However, these types of
sources with specific modeling requirements c. Appropriate model input data should be models can be appropriately applied for
available before an attempt is made to assessment of near-field and regional scale
for selected regulatory applications. Section
evaluate or apply an air quality model. reactive pollutant impacts from specific
7 discusses general considerations common
Assuming the data are adequate, the greater sources 7 10 11 12 or all sources.13 14 15
to many modeling analyses for stationary and
the detail with which a model considers the Photochemical gird models simulate a more
mobile sources. Section 8 makes
spatial and temporal variations in realistic environment for chemical
recommendations for data inputs to models
meteorological conditions and permit- transformation,7 12 but simulations can be
including source, background air quality, and
enforceable emissions, the greater the ability more resource intensive than Lagrangian or
meteorological data. Section 9 summarizes
to evaluate the source impact and to Gaussian plume models.
how estimates and measurements of air
distinguish the effects of various control e. Competent and experienced
quality are used in assessing source impact
strategies. meteorologists, atmospheric scientists, and
and in evaluating control strategies. analysts are an essential prerequisite to the
i. Appendix W to 40 CFR part 51 contains d. There are three types of models that
have historically been used in the regulatory successful application of air quality models.
an appendix: Appendix A. Thus, when The need for such specialists is critical when
reference is made to ‘‘appendix A’’ in this demonstrations applicable in the Guideline,
each having strengths and weaknesses that the more sophisticated models are used or
document, it refers to appendix A to the area being investigated has complicated
appendix W to 40 CFR part 51. Appendix A lend themselves to particular regulatory
applications. meteorological or topographic features. It is
contains summaries of refined air quality important to note that a model applied
models that are ‘‘preferred’’ for particular i. Gaussian plume models use a ‘‘steady-
state’’ approximation, which assumes that improperly or with inappropriate data can
applications; both EPA models and models lead to serious misjudgments regarding the
developed by others are included. over the model time step, the emissions,
meteorology and other model inputs, are source impact or the effectiveness of a
2.0 Overview of Model Use constant throughout the model domain, control strategy.
f. The resource demands generated by use
a. Increasing reliance has been placed on resulting in a resolved plume with the
of air quality models vary widely depending
concentration estimates from air quality emissions distributed throughout the plume
on the specific application. The resources
models as the primary basis for regulatory according to a Gaussian distribution. This
required may be important factors in the
decisions concerning source permits and formulation allows Gaussian models to
selection and use of a model or technique for
emission control requirements. In many estimate near-field impacts of a limited
a specific analysis. These resources depend
situations, such as review of a proposed new number of sources at a relatively high
on the nature of the model and its
source, no practical alternative exists. Before resolution, with temporal scales of an hour
complexity, the detail of the databases, the
attempting to implement the guidance and spatial scales of meters. However, this
difficulty of the application, the amount and
contained in this document, the reader formulation allows for only relatively inert level of expertise required, and the costs of
should be aware of certain general pollutants, with very limited considerations manpower and computational facilities.
information concerning air quality models of transformation and removal (e.g.,

and their evaluation and use. Such deposition), and further limits the domain for 2.1.1 Model Accuracy and Uncertainty
information is provided in this section. which the model may be used. Thus, a. The formulation and application of air
Gaussian models may not be appropriate if quality models are accompanied by several
2.1 Suitability of Models model inputs are changing sharply over the sources of uncertainty. ‘‘Irreducible’’
a. The extent to which a specific air quality model time step or within the desired model uncertainty stems from the ‘‘unknown’’
model is suitable for the assessment of source domain or if more advanced considerations conditions, which may not be explicitly
impacts depends upon several factors. These of chemistry are needed. accounted for in the model (e.g., the
include: (1) The topographic and ii. Lagrangian puff models, on the other turbulent velocity field). Thus, there are
meteorological complexities of the area; (2) hand, are non-steady-state, and assume that likely to be deviations from the observed
concentrations in individual events due to requirements by eliminating the need of more establish a demonstration tool or method as
variations in the unknown conditions. detailed modeling when it is not necessary in a sufficient means for a user or applicant to
‘‘Reducible’’ uncertainties 16 are caused by: a specific regulatory application. For make a demonstration required by regulation,
(1) Uncertainties in the ‘‘known’’ input example, in the context of a PSD permit either by itself or as part of a modeling
conditions (e.g., emission characteristics and application, a simplified or conservative demonstration. To be used for such
meteorological data); (2) errors in the analysis may be sufficient where it shows the regulatory purposes, such a tool or method
measured concentrations; and (3) inadequate proposed construction clearly will not cause must be reflected in a codified regulation or
model physics and formulation. or contribute to ambient concentrations in have a well-documented technical basis and
b. Evaluations of model accuracy should excess of either the NAAQS or the PSD reasoning that is contained or incorporated in
focus on the reducible uncertainty associated increments.2 3 the record of the regulatory decision in
with physics and the formulation of the b. There are two general levels of which it is applied.
model. The accuracy of the model is sophistication of air quality models. The first
normally determined by an evaluation level consists of screening models that 2.3 Availability of Models
procedure which involves the comparison of provide conservative modeled estimates of a. For most of the screening and refined
model concentration estimates with the air quality impact of a specific source or models discussed in the Guideline, codes,
measured air quality data.17 The statement of source category based on simplified associated documentation and other useful
model accuracy is based on statistical tests or assumptions of the model inputs (e.g., preset, information are publicly available for
performance measures such as bias, noise, worst-case meteorological conditions). In the download from the EPA’s Support Center for
correlation, etc.18 19 case of a PSD assessment, if a screening Regulatory Atmospheric Modeling (SCRAM)
c. Since the 1980’s, the EPA has worked model indicates that the concentration Web site at http://www.epa.gov/ttn/scram.
with the modeling community to encourage contributed by the source could cause or This is a Web site with which air quality
development of standardized model contribute to a violation of any NAAQS or modelers should become familiar and
evaluation methods and the development of PSD increment, then the second level of more regularly visit for important model updates
continually improved methods for the sophisticated models should be applied. and additional clarifications and revisions to
characterization of model c. The second level consists of refined modeling guidance documents that are
performance.16 18 20 21 22 There is general models that provide more detailed treatment applicable to EPA programs and regulations.
consensus on what should be considered in of physical and chemical atmospheric
Codes and documentation may also available
the evaluation of air quality models; namely, processes, require more detailed and precise
from the National Technical Information
quality assurance planning, documentation input data, and provide spatially and
Service (NTIS), http://www.ntis.gov, and,
and scrutiny should be consistent with the temporally resolved concentration estimates.
when available, is referenced with the
intended use and should include: As a result they provide a more sophisticated
and, at least theoretically, a more accurate appropriate NTIS accession number.
• Scientific peer review;
• Supportive analyses (diagnostic estimate of source impact and the 3.0 Preferred and Alternative Air Quality
evaluations, code verification, sensitivity effectiveness of control strategies. Models
analyses); d. There are situations where a screening
a. This section specifies the approach to be
• Diagnostic and performance evaluations model or a refined model is not available
such that screening and refined modeling are taken in determining preferred models for
with data obtained in trial locations; and use in regulatory air quality programs. The
not viable options to determine source-
• Statistical performance evaluations in status of models developed by the EPA, as
specific air quality impacts. In such
the circumstances of the intended well as those submitted to the EPA for review
situations, a screening technique or reduced-
applications. and possible inclusion in this Guideline, is
form model may be viable options for
Performance evaluations and diagnostic estimating source impacts. discussed in this section. The section also
evaluations assess different qualities of how i. Screening techniques are differentiated provides the criteria and process for
well a model is performing, and both are from a screening model in that screening obtaining EPA approval for use of alternative
needed to establish credibility within the techniques are approaches that make models for individual cases in situations
client and scientific community. simplified and conservative assumptions where the preferred models are not
d. Performance evaluations allow the EPA about the physical and chemical atmospheric applicable or available. Additional sources of
and model users to determine the relative processes important to determining source relevant modeling information are the EPA’s
performance of a model in comparison with impacts while screening models make Model Clearinghouse 23 (section 3.3), EPA
alternative modeling systems. Diagnostic assumptions about conservative inputs to a modeling conferences, periodic Regional,
evaluations allow determination of a model specific model. The complexity of screening State, and Local Modelers’ Workshops, and
capability to simulate individual processes techniques ranges from simplified the EPA’s SCRAM Web site (section 2.3).
that affect the results, and usually employ assumptions of chemistry applied to refined b. When approval is required for a specific
smaller spatial/temporal scale date sets (e.g., or screening model output to sophisticated modeling technique or analytical procedure
field studies). Diagnostic evaluations enable approximations of the chemistry applied in this Guideline, we refer to the
the EPA and model users to build confidence within a refined model. ‘‘appropriate reviewing authority.’’ Many
that model predictions are accurate for the ii. Reduced-form models are states and some local agencies administer
right reasons. However, the objective computationally efficient simulation tools for NSR and PSD permitting under programs
comparison of modeled concentrations with characterizing the pollutant response to approved into SIPs. In some EPA regions,
observed field data provides only a partial specific types of emission reductions for a federal authority to administer NSR and PSD
means for assessing model performance. Due particular geographic area or background permitting and related activities has been
to the limited supply of evaluation datasets, environmental conditions that reflect delegated to state or local agencies. In these
there are practical limits in assessing model underlying atmospheric science of a refined cases, such agencies ‘‘stand in the shoes’’ of
performance. For this reason, the conclusions model but reduce the computational the respective EPA regions. Therefore,
reached in the science peer reviews and the resources of running a complex, numerical depending on the circumstances, the
supportive analyses have particular relevance air quality model such as a photochemical appropriate reviewing authority may be an
in deciding whether a model will be useful grid model. EPA Regional Office, a state, local, or tribal
for its intended purposes. In such situations, an attempt should be agency, or perhaps the Federal Land Manager
made to acquire or improve the necessary (FLM). In some cases, the Guideline requires
2.2 Levels of Sophistication of Air Quality databases and to develop appropriate review and approval of the use of an
Analyses and Models analytical techniques, but the screening alternative model by the EPA Regional Office
a. It is desirable to begin an air quality technique or reduced-form model may be (sometimes stated as ‘‘Regional
analysis by using simplified or conservative sufficient in conducting regulatory modeling Administrator’’). For all approvals of
methods (or both) followed, as appropriate, applications when applied in consultation alternative models or techniques, the EPA
by more complex and refined methods. The with the EPA Regional Office. Regional Office will coordinate and shall
purpose of this approach is to streamline the e. Consistent with the general principle seek concurrence with the EPA’s Model
process and sufficiently address regulatory described in paragraph 2.2(a), the EPA may Clearinghouse. If there is any question as to
the appropriate reviewing authority, you i. The model must be written in a common model inputs and/or analyzing the model
should contact the EPA Regional Office programming language, and the executable(s) results without otherwise altering the model
modeling contact (http://www.epa.gov/ttn/ must run on a common computer platform. kernel is another example of a modification
scram/guidance_cont_regions.htm), whose ii. The model must be documented in a that does not affect concentrations. However,
jurisdiction generally includes the physical user’s guide or model formulation report when any changes are made, the Regional
location of the source in question and its which identifies the mathematics of the Administrator must require a test case
expected impacts. model, data requirements and program example to demonstrate that the modeled
c. In all regulatory analyses, early operating characteristics at a level of detail concentration are not affected.
discussions among the EPA Regional Office comparable to that available for other c. A preferred model must be operated
staff, state, local, and tribal agency staff, recommended models in appendix A. with the options listed in appendix A for its
industry representatives, and where iii. The model must be accompanied by a intended regulatory application. If other
appropriate, the FLM, are invaluable and are complete test dataset including input options are exercised, the model is no longer
strongly encouraged. Prior to the actual parameters and output results. The test data ‘‘preferred.’’ Any other modification to a
analyses, agreement on the databases to be must be packaged with the model in preferred model that would result in a
used, modeling techniques to be applied, and computer-readable form. change in the concentration estimates
the overall technical approach helps avoid iv. The model must be useful to typical likewise alters its status so that it is no longer
misunderstandings concerning the final users, e.g., state air agencies, for specific air a preferred model. Use of the modified model
results and may reduce the later need for quality control problems. Such users should must then be justified as an alternative model
be able to operate the computer program(s) on a case-by-case basis to the appropriate
additional analyses. The preparation of a
from available documentation. reviewing authority and approved by the
written modeling protocol that is vetted with
v. The model documentation must include Regional Administrator.
the appropriate reviewing authority helps to
a robust comparison with air quality data d. Where the EPA has not identified a
keep misunderstandings and resource
(and/or tracer measurements) or with other preferred model for a particular pollutant or
expenditures at a minimum.
well- established analytical techniques. situation, the EPA may establish a multi-
d. The identification of preferred models in vi. The developer must be willing to make tiered approach for making a demonstration
this Guideline should not be construed as a the model and source code available to users required under PSD or another CAA program.
determination that the preferred models at reasonable cost or make them available for The initial tier or tiers may involve use of
identified here are to be permanently used to public access through the Internet or demonstration tools, screening models,
the exclusion of all others or that they are the National Technical Information Service. The screening techniques, or reduced-form
only models available for relating emissions model and its code cannot be proprietary. models; while the last tier may involve the
to air quality. The model that most accurately d. The EPA’s process of establishing a use of demonstration tools, refinded models
estimates concentrations in the area of preferred model includes a determination of or techniques, or alternative models
interest is always sought. However, technical merit, in accordance with the above approved under section 3.2.
designation of specific preferred models is six items including the practicality of the
needed to promote consistency in model model for use in ongoing regulatory 3.2 Alternative Models
selection and application. programs. Each model will also be subjected 3.2.1 Discussion
3.1 Preferred Models to a performance evaluation for an a. Selection of the best model or techniques
appropriate database and to a peer scientific for each individual air quality analysis is
3.1.1 Discussion review. Models for wide use (not just an always encouraged, but the selection should
a. The EPA has developed some models isolated case) that are found to perform better be done in a consistent manner. A simple
suitable for regulatory application, while will be proposed for inclusion as preferred listing of models in this Guideline cannot
other models have been submitted by private models in future Guideline revisions. alone achieve that consistency nor can it
developers for possible inclusion in the e. No further evaluation of a preferred necessarily provide the best model for all
Guideline. Refined models that are preferred model is required for a particular application possible situations. As discussed in section
and required by the EPA for particular if the EPA requirements for regulatory use 3.1.1, the EPA has determined and applied a
applications have undergone the necessary specified for the model in the Guideline are specific evaluation protocol that provides a
peer scientific reviews 24 25 and model followed. Alternative models to those listed statistical technique for evaluating model
performance evaluation exercises 26 27 that in appendix A should generally be compared performance for predicting peak
include statistical measures of model with measured air quality data when they are concentration values, as might be observed at
performance in comparison with measured used for regulatory applications consistent individual monitoring locations.29 This
air quality data as described in section 2.1.1. with recommendations in section 3.2. protocol is available to assist in developing
b. An American Society for Testing and 3.1.2 Requirements a consistent approach when justifying the use
Materials (ASTM) reference 28 provides a a. Appendix A identifies refined models of other-than-preferred models recommended
general philosophy for developing and that are preferred for use in regulatory in the Guideline (i.e., alternative models).
implementing advanced statistical applications. If a model is required for a The procedures in this protocol provide a
evaluations of atmospheric dispersion particular application, the user must select a general framework for objective decision-
models, and provides an example statistical model from appendix A or follow procedures making on the acceptability of an alternative
technique to illustrate the application of this in section 3.2.2 for use of an alternative model for a given regulatory application.
philosophy. Consistent with this approach, model or technique. Preferred models may be These objective procedures may be used for
the EPA has determined and applied a used without a formal demonstration of conducting both the technical evaluation of
specific evaluation protocol that provides a applicability as long as they are used as the model and the field test or performance
statistical technique for evaluating model indicated in each model summary in evaluation.
performance for predicting peak appendix A. Further recommendations for b. This subsection discusses the use of
concentration values, as might be observed at the application of preferred models to alternate models and defines three situations
individual monitoring locations.29 specific source applications are found in when alternative models may be used. This
c. When a single model is found to perform subsequent sections of the Guideline. subsection also provides a procedure for
better than others, it is recommended for b. If changes are made to a preferred model implementing 40 CFR 51.166(l)(2) in PSD
application as a preferred model and listed without affecting the modeled permitting. This provision requires written
in appendix A. If no one model is found to concentrations, the preferred status of the approval of the Administrator for any
clearly perform better through the evaluation model is unchanged. Examples of modification or substitution of an applicable
exercise, then the preferred model listed in modifications that do not affect model. An applicable model for purposes of
appendix A may be selected on the basis of concentrations are those made to enable use 40 CFR 51.166(l) is a preferred model in
other factors such as past use, public of a different computer platform or those that appendix A to the Guideline. Approval to use
familiarity, resource requirements, and only affect the format or averaging time of the an alternative model under section 3.2 of the
availability. Accordingly, the models listed model results. The integration of a graphical Guideline qualifies as approval for the
in appendix A meet these conditions: user interface (GUI) to facilitate setting up the modification or substitution of a model under
40 CFR 51.166(l)(2). The Regional be followed, as appropriate. Preparation and 4.0 Models for Carbon Monoxide, Lead,
Administrators are delegated authority to implementation of an evaluation protocol Sulfur Dioxide, Nitrogen Dioxide and
issue such approvals under section 3.2 of the which is acceptable to both control agencies Primary Particulate Matter
Guideline, provided that such approval is and regulated industry is an important
issued after consultation with EPA’s Model element in such an evaluation. 4.1 Discussion
Clearinghouse and formally documented in a e. Finally, for condition (3) in paragraph (b) a. This section identifies modeling
concurrence memorandum from EPA’s of this subsection, an alternative model or approaches generally used in the air quality
Model Clearinghouse which demonstrates technique may be approved for use provided impact analysis of sources that emit the
that the requirements within section 3.2 for that: criteria pollutants carbon monoxide (CO),
use of an alternative model have been met. i. The model or technique has received a lead, sulfur dioxide (SO2), nitrogen dioxide
3.2.2 Requirements scientific peer review; (NO2), and primary particulates (PM2.5 and
ii. The model or technique can be PM10).
a. Determination of acceptability of an demonstrated to be applicable to the problem b. The guidance in this section is specific
alternative model is an EPA Regional Office on a theoretical basis; to the application of the Gaussian plume
responsibility in consultation with EPA’s iii. The databases which are necessary to models identified in appendix A. Gaussian
Model Clearinghouse as discussed in perform the analysis are available and plume models assume that emissions and
paragraphs 3.0(b) and 3.2.1(b). Where the adequate; meteorology are in a steady-state, which is
Regional Administrator finds that an iv. Appropriate performance evaluations of typically based on an hourly time step. This
alternative model is more appropriate than a the model or technique have shown that the approach results in a plume that has an
preferred model, that model may be used model or technique is not inappropriately hourly-averaged distribution of emission
subject to the approval of the EPA Regional biased for regulatory application; a and mass according to a Gaussian curve through
Office based on the requirements of this v. A protocol on methods and procedures the plume. Though Gaussian steady-state
subsection. This finding will normally result to be followed has been established. models conserve the mass of the primary
from a determination that (1) a preferred air f. To formally document that the pollutant throughout the plume, they can
quality model is not appropriate for the requirements of section 3.2 for use of an still take into account a limited consideration
particular application; or (2) a more alternative model are satisfied for a particular of first-order removal processes (e.g., wet and
appropriate model or technique is available application or range of applications, a dry deposition) and limited chemical
and applicable. memorandum will be prepared by the EPA’s conversion (e.g., OH oxidation).
b. An alternative model shall be evaluated Model Clearinghouse through a consultative c. Due to the steady-state assumption,
from both a theoretical and a performance process with the Region Office. Gaussian plume models are generally
perspective before it is selected for use. There considered applicable to distances less than
are three separate conditions under which 3.3 EPA’s Model Clearinghouse
50 km, beyond which, modeled predictions
such a model may be approved for use: a. The Regional Administrator has the of plume impact are likely conservative. The
1. If a demonstration can be made that the authority to select models that are locations of these impacts are expected to be
model produces concentration estimates appropriate for use in a given situation. unreliable due to changes in meteorology that
equivalent to the estimates obtained using a However, there is a need for assistance and are likely to occur during the travel time.
preferred model; guidance in the selection process so that d. The applicability of Gaussian plume
2. If a statistical performance evaluation fairness, consistency, and transparency in models may vary depending on the
has been conducted using measured air modeling decisions are fostered among the topography of the modeling domain, i.e.,
quality data and the results of that evaluation EPA Regional Offices and the state, local, and simple or complex. Simple terrain, as used
indicate the alternative model performs tribal agencies. To satisfy that need, the EPA here, is considered to be an area where
better for the given application than a established the Model Clearinghouse 23 to terrain features are all lower in elevation than
comparable model in appendix A; or serve a central role of coordination and the top of the stack of the source(s) in
3. If there is no preferred model. collaboration between EPA headquarters and question. Complex terrain is defined as
Any one of these three separate conditions the EPA Regional Offices. Additionally, the terrain exceeding the height of the stack
may justify use of an alternative model. Some EPA holds periodic workshops with EPA being modeled.
known alternative models that are applicable headquarters, EPA Regional Office, and state, e. Gaussian models determine source
for selected situations are listed on the EPA’s local, and tribal agency modeling impacts at discrete locations (receptors) for
SCRAM Web site (section 2.3). However, representatives. each meteorological and emission scenario,
inclusion there does not confer any unique b. The EPA Regional Office should always and generally attempt to estimate
status relative to other alternative models be consulted for information and guidance concentrations at specific sites that represent
concerning modeling methods and
that are being or will be developed in the an ensemble average of numerous repetitions
interpretations of modeling guidance, and to
future. of the same ‘‘event.’’ Uncertainties in model
ensure that the air quality model user has
c. Equivalency, condition (1) in paragraph estimates are driven by this formulation, and
available the latest most up-to-date policy
(b) of this subsection, is established by as noted in section 2.1.1, evaluations of
and procedures. As appropriate, the EPA
demonstrating that the maximum or highest, model accuracy should focus on the
Regional Office may also request assistance
second highest concentrations are within reducible uncertainty associated with
from the EPA’s Model Clearinghouse on
+/¥ 2 percent of the estimates obtained from physics and the formulation of the model.
other applications of models, analytical
the preferred model. The option to show The ‘‘irreducible’’ uncertainty associated
techniques, or databases or to clarify
equivalency is intended as a simple with Gaussian plume models may be
interpretation of the Guideline or related
demonstration of acceptability for an responsible for variation in concentrations of
modeling guidance.
alternative model that is so nearly identical as much as +/¥ 50 percent.30 ‘‘Reducible’’
c. The EPA Regional Office will coordinate
(or contains options that can make it uncertainties 16 can be on a similar scale. For
with the EPA’s Model Clearinghouse after an
identical) to a preferred model that it can be example, Pasquill 31 estimates that, apart
initial evaluation and decision has been
treated for practical purposes as the preferred from data input errors, maximum ground-
developed concerning the application of an
model. However, notwithstanding this level concentrations at a given hour for a
alternative model. The acceptability and
demonstration, models that are not point source in flat terrain could be in error
formal approval process for an alternative
equivalent may be used when one of the two by 50 percent due to these uncertainties.
model is described in section 3.2.
other conditions described in paragraphs (d) Errors of 5 to 10 degrees in the measured
and (e) of this subsection are satisfied. wind direction can result in concentration
a For PSD and other applications that use the
d. For condition (2) in paragraph (b) of this errors of 20 to 70 percent for a particular time
model results in an absolute sense, the model
subsection, established statistical should not be biased toward underestimates.
and location, depending on stability and
performance evaluation procedures and Alternatively, for ozone and PM2.5 SIP attainment station location. Such uncertainties do not
techniques 28 29 for determining the demonstrations and other applications that use the indicate that an estimated concentration does
acceptability of a model for an individual model results in a relative sense, the model should not occur, only that the precise time and
case based on superior performance should note be biased toward overestimates. locations are in doubt. Composite errors in
highest estimated concentrations of 10 to 40 used for this second level assessment, but appropriate reviewing authority (paragraph
percent are found to be typical.32 33 However, applicants shall reach agreed upon 3.0(b)) on the choice of the screening model
estimates of concentrations paired in time approaches (models and modeling or technique for each analysis, on the input
and space with observed concentrations are parameters) on a case-by-case basis. When data and model settings, and the appropriate
less certain. Lagrangian models are used in this manner, metric for satisfying regulatory requirements.
f. Model evaluations and inter-comparisons they shall not include plume-depleting 4.2.1.1 AERSCREEN
should take these aspects of uncertainty into reactions, such that model estimates are
account. For a regulatory application of a considered conservative, as is generally a. Released in 2011, AERSCREEN is the
model, the emphasis of model evaluations is appropriate for screening assessments. EPA’s recommended screening model for
generally placed on the highest modeled d. In those limited situations where a simple and complex terrain for single sources
impacts. Thus, the Cox-Tikvart model cumulative increment analysis beyond 50 km including point sources, area sources,
evaluation approach, which compares the is necessary, the selection and use of an horizontal stacks, capped stacks, and flares.
highest modeled impacts on several alternative model shall occur in agreement AERSCREEN runs AERMOD in a screening
timescales, is recommended for comparisons with the appropriate reviewing authority mode and consists of two main components:
of models and measurements and model (paragraph 3.0(b)) and approval by the EPA (1) The MAKEMET program which generates
inter-comparisons. The approach includes Regional Office based on the requirements of a site-specific matrix of meteorological
bootstrap techniques to determine the paragraph 3.2.2(e). conditions for input into the AERMOD
significance of various modeled predictions model; and (2) the AERSCREEN command-
4.2.1 Screening Models and Techniques prompt interface.
and increases the robustness of such
a. Where a preliminary or conservative b. The MAKEMET program generates a
comparisons when the number of available
estimate is desired, point source screening matrix of meteorological conditions, in the
measurements are limited.34 35 Because of the
techniques are an acceptable approach to air form of AERMOD-ready surface and profile
uncertainty in paired modeled and observed
quality analyses. files, based on user-specified surface
concentrations, any attempts at calibration of
b. As discussed in paragraph 2.2(a), characteristics, ambient temperatures,
models based on these comparisons is of
screening models or techniques are designed minimum wind speed, and anemometer
questionable benefit and shall not be done.
to provide a conservative estimate of height. The meteorological matrix is
4.2 Requirements concentrations. The screening models used generated based on looping through a range
in most applications are the screening of wind speeds, cloud covers, ambient
a. For NAAQS compliance demonstrations
versions of the preferred models for refined temperatures, solar elevation angles, and
under PSD, use of the screening and
applications. The two screening models, convective velocity scales (w*, for convective
preferred models for the pollutants listed in
AERSCREEN 37 38 and CTSCREEN, are conditions only) based on user-specified
this subsection shall be limited to the near-
screening versions of AERMOD (American surface characteristics (Zo, Bo, r). For
field at a nominal distance of 50 km or less.
Meteorological Society (AMS)/EPA unstable cases, the convective mixing height
Near-field application is consistent with
Regulatory Model) and CTDMPLUS (Zic) is calculated based on w*, and the
capabilities of Gaussian plume models and,
(Complex Terrain Dispersion Model Plus mechanical mixing height (Zim) is calculated
based on the EPA’s assessment, is sufficient
Algorithms for Unstable Situations), for unstable and stable conditions based on
to address whether a source will cause or respectively. AERSCREEN is the preferred the friction velocity, u*.
contribution to ambient concentrations in screening model for most applications in all c. For applications involving simple or
excess to a NAAQS. In most cases, maximum types of terrain and for applications complex terrain, AERSCREEN interfaces with
source impacts of inert pollutant are involving building downwash. For those AERMAP. AERSCREEN also interfaces with
anticipated to occur within 10 to 20 km from applications in complex terrain where the BIPPRM to provide the necessary building
the source. Therefore, the EPA does not application involves a well-defined hill or parameters for applications involving
consider a long-range transport assessment ridge, CTSCREEN 39 can be used. building downwash using the PRIME
beyond 50 km necessary for these c. Although AERSCREEN and CTSCREEN downwash algorithm. AERSCREEN generates
pollutants.36 are designed to address a single-source inputs to AERMOD via MAKEMET,
b. For assessment of PSD increments scenario, there are approaches that can be AERMAP, and BPIPPRM and invokes
within the near-field nominal distance of 50 used on a case-by-case basis to address multi- AERMOD in a screening mode. The screening
km or less, use of the screening and preferred source situations using screening mode of AERMOD forces the AERMOD
models for the pollutants listed in this meteorology or other conservative model model calculations to represent values for the
subsection shall be limited to the same assumptions. However, the appropriate plume centerline, regardless of the source-
screening and preferred models approved for reviewing authority (paragraph 3.0(b)) shall receptor-wind direction orientation. The
NAAQS compliance demonstrations. be consulted, and concurrence obtained, on maximum concentration output from
c. To determine if a Class I PSD increment the protocol for modeling multiple sources AERSCREEN represents a worst-case 1-hour
analyses may be necessary beyond 50 km with AERSCREEN or CTSCREEN to ensure concentration. Averaging-time scaling factors
(i.e., long-range transport assessment), the that the worst case is identified and assessed. of 0.9 for 3-hour, 0.7 for 8-hour, 0.40 for 24-
following screening approach shall be used d. As discussed in section 4.2.3.4, there are hour, and 0.08 for annual concentration
to determine if a significant impact will also screening techniques built into averages are applied internally by
occur with particular focus on Class I areas AERMOD that use simplified or limited AERSCREEN to the highest 1-hour
that may be threatened at such distances. chemistry assumptions for determining the concentration calculated by the model for
i. Based on application in the near-field of partitioning of NO and NO2 for NO2 non-area type sources. For area type source
the appropriate screening and/or preferred modeling. These screening techniques are concentrations for averaging times greater
model, determine the significance of the part of the EPA’s preferred modeling than one hour, the concentrations are equal
ambient impacts at or about 50 km from the approach for NO2 and do not need to be to the 1-hour estimates.37 40
new or modifying source. If this initial step approved as an alternative model. However,
indicates there may be significant ambient 4.2.1.2 CTSCREEN
as with other screening models and
impacts at that distance or such near-field techniques, their usage shall occur in a. CTSCREEN 39 41 can be used to obtain
assessment is not available, then further agreement with the appropriate reviewing conservative, yet realistic, worst-case
assessment is necessary. authority (paragraph 3.0(b)). estimates for receptors located on terrain
ii. For assessment of Class I significance of e. All screening models and techniques above stack height. CTSCREEN accounts for
ambient impacts and cumulative increment shall be configured to appropriately address the three-dimensional nature of plume and
analyses, there is not a preferred model or the site and problem at hand. Close attention terrain interaction and requires detailed
screening approach for distances beyond 50 must be paid to whether the area should be terrain data representative of the modeling
km. Thus, the EPA Regional Office shall be classified urban or rural in accordance with domain. The terrain data must be digitized in
consulted in determining the appropriate and section 7.2.1.1. The climatology of the area the same manner as for CTDMPLUS and a
agreed upon modeling approach to conduct must be studied to help define the worst-case terrain processor is available.42 CTSCREEN is
the second level assessment. Typically a meteorological conditions. Agreement shall designed to execute a fixed matrix of
Lagrangian model may be the type of model be reached between the model user and the meteorological values for wind speed (u),
standard deviation of horizontal and vertical 4.2.2.1 AERMOD 4.2.2.3 OCD

wind speeds (sv, sw), vertical potential a. For a wide range of regulatory a. If the modeling application involves
temperature gradient (dq/dz), friction applications in all types of terrain, and for determining the impact of offshore emissions
velocity (u*), Monin-Obukhov length (L), from point, area, or line sources on the air
aerodynamic building downwash, the
mixing height (zi) as a function of terrain quality of coastal regions, the recommended
recommended model is AERMOD.44 45 The
height, and wind directions for both neutral/ model is the OCD (Offshore and Coastal
stable conditions and unstable convective AERMOD regulatory modeling system
consists of the AERMOD dispersion model, Dispersion) Model. OCD is a straight-line
conditions. The maximum concentration Gaussian model that incorporates overwater
output from CTSCREEN represents a worst- the AERMET meteorological processor, and
the AERMAP terrain processor. AERMOD is plume transport and dispersion as well as
case 1-hour concentration. Time-scaling changes that occur as the plume crosses the
factors of 0.7 for 3-hour, 0.15 for 24-hour and a steady-state Gaussian plume model
shoreline. OCD is also applicable for
0.03 for annual concentration averages are applicable to directly emitted air pollutants
situations that involve platform building
applied internally by CTSCREEN to the that employs best state-of-practice
downwash.
highest 1-hour concentration calculated by parameterizations for characterizing the
the model. meteorological influences and dispersion. 4.2.3 Pollutant Specific Modeling
Differentiation of simple versus complex Requirements
4.2.1.3 Screening in Complex Terrain
terrain is unnecessary with AERMOD. In 4.2.3.1 Models for Carbon Monoxide
a. For applications utilizing AERSCREEN, complex terrain, AERMOD employs the well-
AERSCREEN automatically generates a polar- a. Models for assessing the impact of CO
known dividing-streamline concept in a
grid receptor network with spacing emissions are needed to meet NSR
simplified simulation of the effects of plume- requirements, including PSD, to address
determined by the maximum distance to
terrain interactions. compliance with the CO NAAQS and to
model. If the application warrants a different
b. The AERMOD modeling system has been determine localized impacts from
receptor network than that generated by
AERSCREEN, it may be necessary to run extensively evaluated across a wide range of transportations projects. Examples include
AERMOD in screening mode with a user- scenarios based on numerous field studies, evaluating effects of point sources, congested
defined network. For CTSCREEN including tall stacks in flat and complex roadway intersections, and highways, as well
applications or AERMOD in screening mode terrain settings, sources subject to building as the cumulative effect of numerous sources
outside of AERSCREEN, placement of downwash influences, and low-level non- of CO in an urban area.
receptors requires very careful attention buoyant sources.27 These evaluations b. The general modeling recommendations
when modeling in complex terrain. Often the included several long-term field studies and requirements for screening models in
highest concentrations are predicted to occur associated with operating plants as well as section 4.2.1 and refined models in section
under very stable conditions, when the several intensive tracer studies. Based on 4.2.2 shall be applied for CO modeling. Given
plume is near, or impinges on, the terrain. these evaluations, AERMOD has shown the relatively low CO background
The plume under such conditions may be consistently good performance, with ‘‘errors’’ concentrations, screening techniques are
quite narrow in the vertical, so that even in predicted vs. observed peak likely to be adequate in most cases. However,
relatively small changes in a receptor’s concentrations, based on the Robust Highest since the screening model specified in
location may substantially affect the Concentration (RHC) metric, consistently section 4.2.1 (AERSCREEN) can only handle
predicted concentration. Receptors within within the range of 10 to 40 percent cited in one source at a time, a section 4.2.2 model
about a kilometer of the source may be even paragraph 4.1(g). may be used with screening meteorology
more sensitive to location. Thus, a dense c. AERMOD incorporates the Plume Rise (e.g., generated with MAKEMET) to conduct
array of receptors may be required in some Model Enhancements (PRIME) algorithm to screening assessments of CO projects
cases. account for enhanced plume growth and involving more than one source (e.g.,
b. For applications involving AERSCREEN, restricted plume rise for plumes affected by roadway hotspot assessments).47
AERSCREEN interfaces with AERMAP to building wake effects.46 The PRIME 4.2.3.2 Models for Lead
generate the receptor elevations. For algorithm accounts for entrainment of plume
applications involving CTSCREEN, digitized a. In January 1999 (40 CFR part 58,
mass into the cavity recirculation region, appendix D), the EPA gave notice that
contour data must be preprocessed 42 to including re-entrainment of plume mass into
provide hill shape parameters in suitable concern about ambient lead impacts was
the wake region beyond the cavity. being shifted away from roadways and
input format. The user then supplies d. AERMOD incorporates the Buoyant Line
receptors either through an interactive toward a focus on stationary point sources.
and Point Source (BLP) Dispersion model to Thus, models for assessing the impact of lead
program that is part of the model or directly,
account for buoyant plume rise from line emissions are needed to meet NSR
by using a text editor; using both methods to
sources. The BLP option within AERMOD requirements, including PSD, to address
select receptors will generally be necessary to
utilizes the standard meteorological inputs compliance with the lead NAAQS and for
assure that the maximum concentrations are
provided by the AERMET meteorological SIP attainment demonstrations. The EPA has
estimated by either model. In cases where a
terrain feature may ‘‘appear to the plume’’ as processor. also issued guidance on siting ambient
smaller, multiple hills, it may be necessary e. The state-of-the-science for modeling monitors in the vicinity of stationary point
to model the terrain both as a single feature atmospheric deposition is evolving and new sources.48 For lead, the SIP should contain an
and as multiple hills to determine design modeling techniques are continually being air quality analysis to determine the
concentrations. assessed and their results are being compared maximum rolling 3-month average lead
c. Other screening techniques may be with observations. Consequently, while concentration resulting from major lead point
acceptable for complex terrain cases where deposition treatment is available in sources, such as smelters, gasoline additive
established procedures 43 are used. The user AERMOD, the approach taken for any plants, etc. The EPA has developed a post-
is encouraged to confer with the appropriate purpose shall be coordinated with the processor to calculate rolling 3-month
reviewing authority (paragraph 3.0(b)) if any appropriate reviewing authority (paragraph average concentrations from model output.49
unresolvable problems are encountered, e.g., 3.0(b)). General guidance for lead SIP development is
applicability, meteorological data, receptor also available.50
4.2.2.2 CTDMPLUS
siting, or terrain contour processing issues. b. For major lead point sources, such as
a. If the modeling application involves an smelters, which contribute fugitive emissions
4.2.2 Refined Models elevated point source with a well-defined hill and for which deposition is important,
a. A brief description of each preferred or ridge and a detailed dispersion analysis of professional judgment should be used, and
model for refined applications is found in the spatial pattern of plume impacts is of there shall be coordination with the
appendix A. Also listed in that appendix are interest, CTDMPLUS is available. appropriate reviewing authority (paragraph
availability, the model input requirements, CTDMPLUS provides greater resolution of 3.0(b)). For most applications, the general
the standard options that shall be selected concentrations about the contour of the hill requirements for screening and refined
when running the program, and output feature than does AERMOD through a models of section 4.2.1 and 4.2.2 are
options. different plume-terrain interaction algorithm. applicable to lead modeling.
4.2.3.3 Models for Sulfur Dioxide authority (paragraph 3.0(b)). Additionally, to justify a source’s anticipated NO2/NOX in-
a. Models for SO2 are needed to meet NSR since screening techniques are conservative stack ratios, such as manufacturer test data,
requirements, including PSD, to address by their nature, there are limitations to how state or local agency guidance, peer-reviewed
compliance with the SO2 NAAQS and PSD these options can be used. Specifically, literature, the EPA’s NO2/NOX ratio database.
increments, for SIP attainment negative emissions should not be modeled e. For Tier 3, a detailed screening
demonstrations,51 and for characterizing because decreases in concentrations would technique shall be applied on a case-by-case
current air quality via modeling.52 SO2 is one be overestimated. Each tiered approach (see
basis. Because of the additional input data
of a group of highly reactive gasses known as Figure 4–1) accounts for increasing complex
‘‘oxides of sulfur’’ with largest emissions considerations of NO2 chemistry and is requirements and complexities associated
sources being fossil fuel combustion at power described in paragraphs b through d of this with the Tier 3 options, their usage shall
plants and other industrial facilities. subsection. The tiers of NO2 modeling occur in consultation with the EPA Regional
b. Given the relatively inert nature of SO2 include: Office in addition to the appropriate
on the short-term time scales of interest (i.e., i. A first-tier (most conservative) ‘‘full’’ reviewing authority. The Ozone Limiting
1-hour) and the sources of SO2 (i.e., conversion approach; Method (OLM) 56 and the Plume Volume
stationary point sources), the general ii. A second-tier approach that assumes Molar Ratio Method (PVMRM) 57 are two
modeling requirements for screening models ambient equilibrium between NO and NO2; detailed screening techniques that may be
in section 4.2.1 and refined models in section and used for most sources. These two techniques
4.2.2 are applicable for SO2 modeling iii. A third-tier consisting of several use an appropriate section 4.2.2 model to
applications. For urban areas, AERMOD detailed screening techniques that account estimate NOX concentrations and then
automatically invokes a half-life of 4 hours 53 for ambient ozone and the relative amount of estimate the conversion of primary NO
to SO2. Therefore, care must be taken when NO and NO2 emitted from a source. emissions to NO2 based on the ambient levels
determining whether a source is urban or c. For Tier 1, use an appropriate section of ozone and the plume characteristics. OLM
rural (see section 7.2.1.1 for urban/rural 4.2.2 refined model to estimate nitrogen
only accounts for NO2 formation based on the
determination methodology). oxides (NOX) concentrations and assume a
4.2.3.4 Models for Nitrogen Dioxide total conversion of NO to NO2. If the ambient levels of ozone while PVMRM also
a. Models for assessing the impact of resulting design concentrations exceed the accommodates distance-dependent
sources on ambient NO2 concentrations are NAAQS or PSD increments for NO2, proceed conversion ratios based on ambient ozone.
needed to meet NSR requirements, including to Tier 2. Both PVMRM and OLM require that ambient
PSD, to address compliance with the NO2 d. For Tier 2, multiply the Tier 1 result(s) ozone concentrations be provided on an
NAAQS and PSD increments. Impact of an by the Ambient Ratio Method 2 (ARM2), hourly basis and explicit specification of the
individual source on ambient NO2 depends, which provides estimates of representative speciation of the NO2/NOX in-stack ratios.
in part, on the chemical environment into equilibrium ratios of NO2/NOX value based PVMRM works best for relatively isolated
which the source’s plume is to be emitted. ambient levels of NO2 and NOX derived from and elevated point source modeling while
This is due to the fact that NO2 sources co- national data from the EPA’s Air Quality OLM works best for large groups of sources,
emit NO along with NO2 and any emitted NO System (AQS).55 The national default for area sources, and near-surface releases,
may react with ambient ozone to convert to ARM2 will include a minimum NO2/NOX including road-way sources.
additional NO2 downwind. Thus, ratio of 0.5 and a maximum ratio of 0.9. The f. Alternative models or techniques may be
comprehensive modeling of NO2 would need reviewing agency may establish alternative
considered on a case-by-case basis and their
to consider the ratio of emitted NO and NO2, default minimum NO2/NOX values based on
usage shall be approved by the EPA Regional
the ambient levels of ozone and subsequent the source’s in-stack emissions ratios, with
reactions between ozone and NO, and the alternative minimum values reflecting the Office (section 3.2). Such techniques should
photolysis of NO2 to NO. source’s in-stack NO2/NOX ratios. Preferably, consider individual quantities of NO and
b. Due to the complexity of NO2 modeling, alternative default NO2/NOX values should NO2 emissions, atmospheric transport and
a multi-tiered approach is required to obtain be based on source-specific data which dispersion, and atmospheric transformation
hourly and annual average estimates of satisfies all quality assurance procedures that of NO to NO2. Dispersion models that
NO2.54 Since these methods are considered ensure data accuracy for both NO2 and NOX account for more explicit photochemistry
screening, their usage shall occur in within the typical range of measured values. may also be applied to estimate ambient
agreement with the appropriate reviewing However, alternate information may be used impacts of NOX sources.
4.2.3.5 Models for PM2.5 PM10 in a document published on June 3, PM10. In these circumstances, the alternative
a. The PM2.5 NAAQS, promulgated on July 1993. Models for PM10 are needed to meet modeling approach shall be approved by the
18, 1997, includes particles with an NSR requirements, including PSD, to address EPA Regional Office (section 3.2).
aerodynamic diameter nominally less than or compliance with the PM10 NAAQS and PSD e. The general modeling requirements for
equal to 2.5 micrometers. PM2.5 is a mixture increments and for SIP attainment the refined models discussed in section 4.2.2
consisting of several diverse components58. demonstrations. should be applied for PM10 hot-spot
Ambient PM2.5 generally consists of two b. For most sources, the general modeling modeling for mobile sources. Specific
components, the primary component, requirements for screening models in section guidance is available for analyzing direct
emitted directly from a source, and the 4.2.1 and refined models in section 4.2.2 PM10 impacts from highways, terminals, and
secondary component, which is formed in shall be applied for PM10 modeling. In cases other projects.61
the atmosphere from other pollutants emitted where the particle size and its effect on
5.0 Models for Ozone and Secondarily
from the source. Models for PM2.5 are needed ambient concentrations need to be
Formed Particulate Matter
to meet NSR requirements, including PSD, to considered, particle deposition may be used
address compliance with the PM2.5 NAAQS in on a case-by-case basis and their usage 5.1 Discussion
and PSD increments and for SIP attainment shall be approved by the EPA Regional Office a. Air pollutants formed through chemical
demonstrations. (section 3.2). A SIP development guide 62 is reactions in the atmosphere are referred to as
b. For NSR, including PSD, modeling also available to assist in PM10 analyses and secondary pollutants. For example, ground-
assessments, the refined methods in section control strategy development. level ozone and a portion of particulate
4.2.2 are required for modeling the primary c. Fugitive dust usually refers to dust put matter with aerodynamic diameter less than
component of PM2.5, while the methods in into the atmosphere by the wind blowing 2.5 m m (PM2.5 or fine PM) are secondary
section 5.4 are recommended for addressing over plowed fields, dirt roads or desert or pollutants formed through photochemical
the secondary component of PM2.5. Guidance sandy areas with little or no vegetation. reactions. Ozone and secondarily formed
for PSD assessments is available for Fugitive emissions include the emissions particulate matter are closely related to each
determining the best approach to handling resulting from the industrial process that are other in that they share common sources of
sources of primary and secondary PM2.5.59 not captured and vented through a stack but emissions or are formed in the atmosphere
c. For SIP attainment demonstrations and may be released from various locations from chemical reactions with similar
regional haze reasonable progress goal within the complex. In some unique cases, a precursors.
analyses, effects of a control strategy on PM2.5 model developed specifically for the b. Ozone formation is driven by emissions
are estimated from the sum of the effects on situation may be needed. Due to the difficult of NOX and volatile organic compounds
the primary and secondary components nature of characterizing and modeling (VOCs). Ozone formation is a complicated
composing PM2.5. Model users should refer to fugitive dust and fugitive emissions, the nonlinear process that requires favorable
section 5.4.1 and associated SIP modeling proposed procedure shall be determined in meteorological conditions in addition to VOC
guidance 60 for further details concerning consultation with the appropriate reviewing and NOX emissions. Sometimes complex
appropriate modeling approaches. authority (paragraph 3.0(b)) for each specific terrain features also contribute to the build-
d. The general modeling requirements for situation before the modeling exercise is up of precursors and subsequent ozone
the refined models discussed in section 4.2.2 begun. Re-entrained dust is created by formation or destruction.
should be applied for PM2.5 hot-spot vehicles driving over dirt roads (e.g., haul c. PM2.5 can be either primary (i.e., emitted
modeling for mobile sources. Specific roads) and dust-covered roads typically directly from sources) or secondary in nature.
guidance is available for analyzing direct found in arid areas. Such sources can be The fraction of PM2.5 which is primary versus
PM2.5 impacts from highways, terminals, and characterized as line, area or volume secondary varies by location and season. In
other projects.61 sources.61 63 Emission rates may be based on the United States, PM2.5 is dominated by a
site-specific data or values from the general variety of chemical species or components of
4.2.3.6 Models for PM10 literature. atmospheric particles, such as ammonium
a. The NAAQS for PM10 was promulgated d. Under certain conditions, recommended sulfate, ammonium nitrate, organic carbon
on July 1, 1987. The EPA promulgated dispersion models may not be suitable to (OC) mass, elemental carbon (EC), and other
EP29JY15.001</GPH>
regulations for PSD increment measured as appropriately address the nature of ambient soil compounds and oxidized metals. PM2.5
sulfate, nitrate, and ammonium ions are or PM2.5, model results should be compared agencies with jurisdiction over areas with
predominantly the result of chemical to observation data to generate confidence ozone problems should use photochemical
reactions of the oxidized products of sulfur that the modeling system is representative of grid models to evaluate the relationship
dioxide (SO2) and NOX emissions with direct the local and regional air quality. For ozone between precursor species and ozone. Use of
ammonia (NH3) emissions.64 related projects, model estimates of ozone photochemical grid models is the
d. Modeled strategies designed to reduce should be compared with observations in recommended means for identifying control
ozone or PM2.5 levels typically need to both time and space. For PM2.5, model strategies needed to address high ozone
consider the chemical coupling between estimates of speciated PM2.5 components concentrations in such areas. Judgment on
these pollutants. Control measures reducing (such as sulfate ion, nitrate ion, etc.) should the suitability of a model for a given
ozone and PM2.5 precursor emissions may be compared with observations in both time application should consider factors that
not lead to proportional reductions in ozone and space.65 include use of the model in an attainment
and PM2.5. This coupling is important in d. Model performance metrics comparing test, development of emissions and
understanding processes that control the observations and predictions are often used meteorological inputs to the model, and
levels of both pollutants. Thus, when to summarize model performance. These choice of episodes to model. Guidance on the
feasible, it is important to use models that metrics include mean bias, mean error, use of models and other analyses for
take into account the chemical coupling fractional bias, fractional error, and demonstrating attainment of the air quality
between ozone and PM2.5. In addition, using correlation coefficient.65 There are no goals for ozone is available.60 Users should
such a multi-pollutant modeling system can specific levels of any model performance consult with the appropriate reviewing
reduce the resource burden associated with metric that indicate ‘‘acceptable’’ model authority (paragraph 3.0(b)) to ensure the
applying and evaluating separate models for performance. The EPA’s preferred approach most current modeling guidance is applied.
each pollutant and promotes consistency for providing context about model 5.3.2 Models for Single-Source Air Quality
among the strategies themselves. performance is to compare model Assessments
e. PM2.5 is a mixture consisting of several performance metrics with similar
diverse chemical species or components of contemporary applications.60 65 Because a. Depending on the magnitude of
atmospheric particles. Because chemical and model application purpose and scope vary, emissions, estimating the impact of an
physical properties and origins of each model users should consult with the individual source’s emissions of NOX and
component differ, it may be appropriate to appropriate reviewing authority (paragraph VOC on ozone concentrations is necessary for
use either a single model capable of 3.0(b)) to determine what model performance obtaining a permit. The simulation of ozone
addressing several of the important elements should be emphasized and formation and transport requires realistic
components or to model primary and presented to provide confidence in the treatment of atmospheric chemistry and
secondary components using different regulatory model application. deposition. Models should be applied which
models. Effects of a control strategy on PM2.5 e. There is no preferred modeling system integrate chemical and physical processes
is estimated from the sum of the effects on or technique for estimating ozone or important in the formation, decay, and
the specific components composing PM2.5. secondary PM2.5 for specific source impacts transport of ozone and important precursor
or to assess impacts from multiple sources. species (e.g., Lagrangian and photochemical
5.2 Recommendations grid models). Photochemical grid models are
For assessing secondary pollutant impacts
a. Chemical transformations can play an from single sources, the degree of complexity primarily designed to characterize precursor
important role in defining the concentrations required to assess potential impacts varies emissions and impacts from a wide variety of
and properties of certain air pollutants. depending on the nature of the source, its sources over a large geographic area but can
Models that take into account chemical emissions, and the background environment. also be used to assess the impacts from
reactions and physical processes of various The EPA recommends a two-tiered approach specific sources.7 11 12
pollutants (including precursors) are needed where the first tier consists of using existing b. The first tier of assessment for ozone
for determining the current state of air technically credible and appropriate impacts involves those situations where
quality, as well as predicting and projecting relationships between emissions and impacts existing technical information is available
the future evolution of these pollutants. It is developed from previous modeling that is (e.g., results from existing photochemical
important that a modeling system provide a deemed sufficient for evaluating a source’s grid modeling, published empirical estimates
realistic representation of chemical and impacts. The second tier consists of more of source specific impacts, or reduced-form
physical processes leading to secondary sophisticated case-specific modeling models) in combination with other
pollutant formation and removal from the analyses. The appropriate tier for a given supportive information and analysis for the
atmosphere. application should be selected in purposes of estimating secondary impacts
b. Chemical transport models treat consultation with the appropriate reviewing from a particular source. The existing
atmospheric chemical and physical processes authority (paragraph 3.0(b)) and be consistent technical information should provide a
such as deposition and motion. There are two with EPA guidance.66 credible and representative estimate of the
types of chemical transport models, Eulerian secondary impacts from the project source.
(grid based) and Lagrangian. These types of 5.3 Recommended Models and Approaches The appropriate reviewing authority
models are differentiated from each other by for Ozone (paragraph 3.0(b)) and appropriate EPA
their frame of reference. Eulerian models are a. Models that estimate ozone guidance 66 should be consulted to determine
based on a fixed frame of reference and concentrations are needed to guide the what types of assessments may be
Lagrangian models use a frame of reference choice of strategies for the purposes of a appropriate on a case-by-case basis.
that moves with parcels of air between the nonattainment area demonstrating future c. The second tier of assessment for ozone
source and receptor point.9 Photochemical year attainment of the ozone NAAQS. impacts involves those situations where
grid models are three-dimensional Eulerian Additionally, models that estimate ozone existing technical information is not
grid-based models that treat chemical and concentrations are needed to assess impacts available such that chemical transport
physical processes in each grid cell and use from specific sources or source complexes to models (e.g., photochemical grid models)
diffusion and transport processes to move satisfy requirements for NSR, including PSD, should be used to address single-source
chemical species between grid cells. These and other regulatory programs. Other impacts. Special considerations are needed
types of models are appropriate for purposes for ozone modeling include when using these models to evaluate the
assessment of near-field and regional scale estimating the impacts of specific events on ozone impact from an individual source.
reactive pollutant impacts from specific air quality, ozone deposition impacts, and Guidance on the use of models and other
sources7 10 11 12 or all sources.13 14 15 In some planning for areas that may be attaining the analyses for demonstrating the impacts of
limited cases, the secondary processes can be ozone NAAQS. single sources for ozone is available.66 This
treated with a box model, potentially in document provides a more detailed
combination with a number of other 5.3.1 Models for NAAQS Attainment discussion of the appropriate approaches to
modeling techniques and/or analyses to treat Demonstrations and Multi-Source Air obtaining estimates of ozone impacts from a
individual source sectors. Quality Assessments single source. Model users should use the
c. Regardless of the modeling system used a. Simulation of ozone formation and latest version of this guidance in consultation
to estimate secondary impacts of ozone and/ transport is a complex exercise. Control with the appropriate reviewing authority
(paragraph 3.0(b)) to determine the most available (e.g., results from existing 6.2 Air Quality Related Values
suitable single-source ozone modeling photochemical grid modeling, published a. The 1997 CAA Amendments give FLMs
approach on a case-by-case basis. empirical estimates of source specific an ‘‘affirmative responsibility’’ to protect the
impacts, or reduced-form models) in natural and cultural resources of Class I areas
5.4 Recommended Models and Approaches combination with other supportive
for Secondarily Formed PM 2.5 from the adverse impacts of air pollution and
information and analysis for the purposes of to provide the appropriate procedures and
a. Models are needed to guide the choice estimating secondary impacts from a analysis techniques. The Act identifies the
of strategies to address an observed PM2.5 particular source. The existing technical FLM as the Secretary of the department, or
problem in an area not attaining the PM2.5 information should provide a credible and their designee, with authority over these
NAAQS. Additionally, models are needed to representative estimate of the secondary lands. Mandatory Federal Class I areas are
assess PM2.5 impacts from specific sources or impacts from the project source. The defined in the CAA as international parks,
industrial source complexes to satisfy appropriate reviewing authority (paragraph national parks over 6,000 acres and
requirements for NSR, including PSD, and 3.0(b)) and appropriate EPA guidance 66 wilderness areas and memorial parks over
other regulatory programs. Other purposes should be consulted to determine what types 5,000 acres, established as of 1977. The FLMs
for PM2.5 modeling include estimating the of assessments may be appropriate on a case- are also concerned with the protection of
impacts of specific events on air quality, by-case basis. resources in federally managed Class II areas
visibility, deposition impacts, and planning c. The second tier of assessment for because of other statutory mandates to
for areas that may be attaining the PM2.5 secondary PM2.5 impacts involves those protect these areas.
NAAQS. situations where existing technical b. The FLM agency responsibilities include
5.4.1 Models for NAAQS Attainment information is not available such that the review of air quality permit applications
Demonstrations and Multi-Source Air chemical transport models (e.g., from proposed new or modified major
Quality Assessments photochemical grid models) should be used pollution sources that may affect these Class
for assessments of single-source impacts. I areas to determine if emissions from a
a. Models for PM2.5 are needed to assess the Special considerations are needed when
adequacy of a proposed strategy for meeting proposed or modified source will cause or
using these models to evaluate the secondary contribute to adverse impacts on air quality
the annual and/or 24-hour PM2.5 NAAQS. particulate matter impact from an individual
Modeling primary and secondary PM2.5 can related values (AQRVs) of a Class I area and
source. Guidance on the use of models and making recommendations to the FLM.
be a multi-faceted and complex problem, other analyses for demonstrating the impacts
especially for secondary components of PM2.5 AQRVs are resources identified by the FLM
of single sources for secondary PM2.5 is agencies, which have the potential to be
such as sulfates and nitrates. Control available.66 This document provides a more
agencies with jurisdiction over areas with affected by air pollution. These resources
detailed discussion of the appropriate may include visibility, scenic, cultural,
secondary PM2.5 problems should use models approaches to obtaining estimates of
which integrate chemical and physical physical, or ecological resources for a
secondary particulate matter concentrations
processes important in the formation, decay, particular area. The FLM agencies take into
from a single source. Model users should use
and transport of these species (e.g., account the particular resources and AQRVs
the latest version of this guidance in
photochemical grid models). Suitability of a that would be affected; the frequency and
consultation with the appropriate reviewing
modeling approach or mix of modeling magnitude of any potential impacts; and the
authority (paragraph 3.0(b)) to determine the
approaches for a given application requires direct, indirect, and cumulative effects of any
most suitable single-source modeling
technical judgment as well as professional potential impacts in making their
approach for secondary PM2.5 on a case-by-
experience in choice of models, use of the recommendations.
case basis.
model(s) in an attainment test, development c. While the AQRV notification and impact
of emissions and meteorological inputs to the 6.0 Modeling for Air Quality Related analysis requirements are outlined in the
model, and selection of days to model. Values and Other Governmental Programs PSD regulations at 40 CFR 51.166(p) and 40
Guidance on the use of models and other CFR 52.21(p), determination of appropriate
6.1 Discussion analytical methods and metrics for AQRV’s
analyses for demonstrating attainment of the
air quality goals for PM2.5 is available.59 60 a. Other federal agencies have also are determined by the FLM agencies and are
Users should consult with the appropriate developed specific modeling approaches for published in guidance external to the general
reviewing authority (paragraph 3.0(b)) to their own regulatory or other requirements. recommendations of this paragraph.
ensure the most current modeling guidance Although such regulatory requirements and d. To develop greater consistency in the
is applied. guidance have come about because of EPA application of air quality models to assess
rules or standards, the implementation of potential AQRV impacts in both Class I areas
5.4.2 Models for Single-Source Air Quality
such regulations and the use of the modeling and protected Class II areas, the FLM
Assessments
techniques is under the jurisdiction of the agencies have developed the Federal Land
a. Depending on the magnitude of agency issuing the guidance or directive. Managers’ Air Quality Related Values Work
emissions, estimating the impact of an This section covers such situations with Group Phase I Report (FLAG) 67. FLAG
individual source’s emissions on secondary reference to those guidance documents, when focuses upon specific technical and policy
particulate matter concentrations is necessary they are available. issues associated with visibility impairment,
for obtaining a permit. Primary PM2.5 b. When using the model recommended or effects of pollutant deposition on soils and
components shall be simulated using discussed in the Guideline in support of surface waters, and ozone effects on
AERMOD (see section 4.2.2). The simulation programmatic requirements not specifically vegetation. Model users should consult the
of secondary particulate matter formation covered by EPA regulations, the model user latest version of the FLAG report for current
and transport is a complex exercise requiring should consult the appropriate federal or modeling guidance and with affected FLM
realistic treatment of atmospheric chemistry state agency to ensure the proper application agency representatives for any application
and deposition. Models should be applied and use of the models and/or techniques. specific guidance which is beyond the scope
which integrate chemical and physical Other federal agencies have developed of the Guideline.
processes important in the formation, decay, specific modeling approaches for their own
and transport of these species (e.g., regulatory or other requirements. Most of the 6.2.1 Visibility
Lagrangian and photochemical grid models). programs have, or will have when fully a. Visibility in important natural areas (e.g.,
Photochemical grid models are primarily developed, separate guidance documents that Federal Class I areas) is protected under a
designed to characterize precursor emissions cover the program and a discussion of the number of provisions of the CAA, including
and impacts from a wide variety of sources tools that are needed. The following sections 169A and 169B (addressing impacts
over a large geographic area and can also be paragraphs reference those guidance primarily from existing sources) and section
used to assess the impacts from specific documents, when they are available. No 165 (new source review). Visibility
sources.7 10 attempt has been made to provide a impairment is caused by light scattering and
b. The first tier of assessment for secondary comprehensive discussion of each topic since light absorption associated with particles and
PM2.5 impacts involves those situations the reference documents were designed to do gases in the atmosphere. In most areas of the
where existing technical information is that. country, light scattering by PM2.5 is the most
significant component of visibility physical processes leading to secondary be applied for an assessment of deposition
impairment. The key components of PM2.5 pollutant formation and removal from the impacts due to one or a small group of
contributing to visibility impairment include atmosphere. sources. Over these distances chemical and
sulfates, nitrates, organic carbon, elemental b. Chemical transport models treat physical transformations can change
carbon, and crustal material.67 atmospheric chemical and physical processes atmospheric residence time due to different
b. Visibility regulations (40 CFR 51.300 such as deposition and motion. There are two propensity for deposition to the surface of
through 51.309) require state, local, and tribal types of chemical transport models, Eulerian different forms of nitrate and sulfate. Users
agencies to mitigate current and prevent (grid based) and Lagrangian. These types of should consult the latest version of the FLAG
future visibility impairment in any of the 156 models are differentiated from each other by report 67 and relevant FLM representatives for
mandatory Federal Class I areas where their frame of reference. Eulerian models are guidance on the use of models for deposition.
visibility is considered an important based on a fixed frame of reference and Where source and receptors are in close
attribute. In 1999, the EPA issued revisions Lagrangian models use a frame of reference proximity, users should contact the
to the regulations to address visibility that moves with parcels of air between the appropriate FLM for application specific
impairment in the form of regional haze, source and receptor point.9 Photochemical guidance.
which is caused by numerous, diverse grid models are three-dimensional Eulerian
sources (e.g., stationary, mobile, and area grid-based models that treat chemical and 6.3 Modeling Guidance for Other
sources) located across a broad region (40 physical processes in each grid cell and use Governmental Programs
CFR 51.308 through 51.309). The state of diffusion and transport processes to move a. Dispersion and photochemical grid
relevant scientific knowledge has expanded chemical species between grid cells.9 These modeling need to be conducted to ensure that
significantly since the 1997 CAA types of models are appropriate for individual and cumulative offshore oil and
Amendments. A number of studies and assessment of near-field and regional scale gas exploration, development, and
reports 68 69 have concluded that long-range reactive pollutant impacts from specific production plans and activities do not
transport (e.g., up to hundreds of kilometers) sources 7 10 11 12 or all sources.13 14 15 significantly affect the air quality of any state
of fine particulate matter plays a significant c. Development of the requisite as required under the Outer Continental
role in visibility impairment across the meteorological and emissions databases
Shelf Lands Act (OCSLA). Air quality
country. CAA section 169A requires states to necessary for use of photochemical grid
modeling requires various input datasets,
develop SIPs containing long-term strategies models to estimate AQRVs should conform to
including emissions sources, meteorology,
for remedying existing and preventing future recommendations in section 8 and those
and pre-existing pollutant concentrations.
visibility impairment in the 156 mandatory outlined in the EPA’s Modeling Guidance for
Demonstrating Attainment of Air Quality For sources under the reviewing authority of
Class I Federal areas, where visibility is
Goals for Ozone, PM2.5, and Regional Haze.60 the Department of Interior, Bureau of Ocean
considered an important attribute. In order to
Demonstration of the adequacy of prognostic Energy Management (BOEM), guidance for
develop long-term strategies to address
regional haze, many state, local, and tribal meteorological fields can be established the development of all necessary Outer
agencies will need to conduct regional-scale through appropriate diagnostic and statistical Continental Shelf (OCS) air quality modeling
modeling of fine particulate concentrations performance evaluations consistent with inputs and appropriate model selection and
and associated visibility impairment. recommendations provided in the application is available from the BOEMS’s
c. The FLAG visibility modeling appropriate guidance.60 Model users should Web site: http://www.boem.gov/
recommendations are divided into two consult the latest version of this guidance Environmental-Stewardship/Environmental-
distinct sections to address different and with the appropriate reviewing authority Studies/Gulf-of-Mexico-Region/Approved-
requirements for (1) near field modeling (paragraph 3.0(b)) for any application specific Air-Quality-Models-for-the-GOMR.aspx.
where plumes or layers are compared against guidance which is beyond the scope of this b. The Federal Aviation Administration
a viewing background and (2) distant/multi- subsection. (FAA) is the appropriate reviewing authority
source modeling for plumes and aggregations for air quality assessments of primary
6.2.2 Models for Estimating Deposition
of plumes that affect the general appearance pollutant impacts at airports and air bases.
Impacts
of a scene.67 The recommendations Air quality application for this purpose is
separately address visibility assessments for a. For many Class I areas, AQRVs have intended for estimating the collective impact
sources proposing to locate relatively near been identified that are sensitive to of changes in aircraft operations, point
and at farther distances from these areas.67 atmospheric deposition of air pollutants. source, and mobile source emissions at
Emissions of NOX, sulfur oxides, NH3, airports on pollutant concentrations. The
6.2.1.1 Models for Estimating Near-Field mercury, and secondary pollutants such as latest version of the Aviation Environmental
Visibility Impairment ozone and particulate matter affect Design Tool (AEDT), is developed and is
a. To calculate the potential impact of a components of ecosystems. In sensitive supported by the FAA, and is appropriate for
plume of specified emissions for specific ecosystems, these compounds can acidify air quality assessment of primary pollutant
transport and dispersion conditions (‘‘plume soils and surface waters, add nutrients that impacts at airports or air bases. AEDT has
blight’’) for source-receptor distances less change biodiversity, and affect the ecosystem adopted AERMOD for treating dispersion.
than 50 km, a screening model and guidance services provided by forests and natural Application of AEDT is intended for
are available.67 70 If a more comprehensive areas.67 To address the relationship between estimating the collective impact of changes in
analysis is necessary, a refined model should deposition and ecosystem effects the FLM aircraft operations, point source, and mobile
be selected. The model selection, procedures, agencies have developed estimates of critical source emissions on pollutant
and analyses should be determined in loads. A critical load is defined as ‘‘A concentrations. It is not intended for PSD,
consultation with the appropriate reviewing quantitative estimate of an exposure to one SIP, or other regulatory air quality analyses
authority (paragraph 3.0(b)) and the affected or more pollutants below which significant of point or mobile sources at or peripheral to
FLM(s). harmful effects on specified sensitive airport property that are unrelated to airport
elements of the environment do not occur operations. The latest version of AEDT may
6.2.1.2 Models for Estimating Visibility
according to present knowledge.’’ 71 be obtained from FAA at its Web site:
Impairment for Long-Range Transport
b. The FLM deposition modeling https://aedt.faa.gov.
a. Chemical transformations can play an recommendations are divided into two
important role in defining the concentrations distinct sections to address different 7.0 General Modeling Considerations
and properties of certain air pollutants. requirements for (1) near field modeling, and
Models that take into account chemical (2) distant/multi-source modeling for 7.1 Discussion
reactions and physical processes of various cumulative effects. The recommendations a. This section contains recommendations
pollutants (including precursors) are needed separately address deposition assessments for concerning a number of different issues not
for determining the current state of air sources proposing to locate relatively near explicitly covered in other sections of the
quality, as well as predicting and projecting and at farther distances from these areas.67 Guideline. The topics covered here are not
the future evolution of these pollutants. It is Where the source and receptors are not in specific to any one program or modeling area
important that a modeling system provide a close proximity, chemical transport (e.g., but are common to dispersion modeling
realistic representation of chemical and photochemical grid) models generally should analyses for criteria pollutants.
7.2 Recommendations (paragraph 3.0(b)) and the latest version of using the results obtained from the screening
7.2.1 All Sources the AERMOD Implementation Guide.76 techniques.
e. It should be noted that AERMOD allows ii. Shoreline fumigation. Fumigation can be
7.2.1.1 Dispersion Coefficients for modeling rural and urban sources in a an important phenomenon on and near the
a. For any dispersion modeling exercise, single model run. For analyses of whole shoreline of bodies of water. This can affect
the urban or rural determination of a source urban complexes, the entire area should be both individual plumes and area-wide
is critical in determining the boundary layer modeled as an urban region if most of the emissions. When fumigation conditions are
characteristics that affect the model’s sources are located in areas classified as expected to occur from a source or sources
prediction of downwind concentrations. urban. For tall stacks located within or with tall stacks located on or just inland of
Historically, steady-state Gaussian plume adjacent to small or moderate sized urban a shoreline, this should be addressed in the
models used in most applications have areas, the stack height or effective plume air quality modeling analysis. EPA has
employed dispersion coefficients based on height may extend above the urban boundary evaluated several coastal fumigation models,
Pasquill-Gifford 72 in rural areas and layer and, therefore, may be more and the evaluation results of these models are
McElroy- Pooler 73 in urban areas. These appropriately modeled using rural available for their possible application on a
coefficients are still incorporated in the BLP coefficients. Model users should consult with case-by-case basis when air quality estimates
and OCD models. However, the AERMOD the appropriate reviewing authority under shoreline fumigation conditions are
model incorporates a more up-to-date (paragraph 3.0(b)) when evaluating this needed.79 Selection of the appropriate model
characterization of the atmospheric boundary situation and the latest version of the for applications where shoreline fumigation
layer using continuous functions of AERMOD Implementation Guide.76 is of concern should be determined in
parameterized horizontal and vertical f. Buoyancy-induced dispersion (BID), as consultation with the appropriate reviewing
turbulence based on Monin-Obukhov identified by Pasquill,77 is included in the authority (paragraph 3.0(b)).
similarity (scaling) relationships.44 Another preferred models and should be used where iii. Stagnation. Stagnation conditions are
key feature of AERMOD’s formulation is the buoyant sources, e.g., those involving fuel characterized by calm or very low wind
option to use directly observed variables of combustion, are involved. speeds, and variable wind directions. These
the boundary layer to parameterize stagnant meteorological conditions may
7.2.1.2 Complex Winds
dispersion.44 45 persist for several hours to several days.
b. The selection of rural or urban a. Inhomogeneous local winds. In many During stagnation conditions, the dispersion
dispersion coefficients in a specific parts of the United States, the ground is of air pollutants, especially those from low-
application should follow one of the neither flat nor is the ground cover (or land level emissions sources, tends to be
procedures suggested by Irwin 74 to use) uniform. These geographical variations minimized, potentially leading to relatively
determine whether the character of an area is can generate local winds and circulations, high ground-level concentrations. If point
primarily urban or rural: and modify the prevailing ambient winds sources are of interest, users should note the
i. Land Use Procedure: (1) Classify the land and circulations. Geographic effects are most guidance provided in paragraph (a) of this
use within the total area, Ao, circumscribed apparent when the ambient winds are light subsection. Selection of the appropriate
by a 3km radius circle about the source using or calm.78 In general these geographically model for applications where stagnation is of
the meteorological land use typing scheme induced wind circulation effects are named concern should be determined in
proposed by Auer; 75 (2) if land use types I1, after the source location of the winds, e.g., consultation with the appropriate reviewing
I2, C1, R2, and R3 account for 50 percent or lake and sea breezes, and mountain and authority (paragraph 3.0(b)).
more of Ao, use urban dispersion coefficients; valley winds. In very rugged hilly or
7.2.1.3 Gravitational Settling and
otherwise, use appropriate rural dispersion mountainous terrain, along coastlines, or
Deposition
coefficients. near large land use variations, the
ii. Population Density Procedure: (1) characterization of the winds is a balance of a. Gravitational settling and deposition
Compute the average population density, p̄ various forces, such that the assumptions of may be directly included in a model if either
per square kilometer with Ao as defined steady-state straight-line transport both in is a significant factor. When particulate
above; (2) If p̄ is greater than 750 people/km2, time and space are inappropriate. In such matter sources can be quantified and settling
use urban dispersion coefficients; otherwise cases, a model should be chosen to fully treat and dry deposition are problems,
use appropriate rural dispersion coefficients. the time and space variations of meteorology professional judgment should be used, and
(Of the two methods, the land use procedure effects on transport and dispersion. The there should be coordination with the
is considered more definitive.) setup and application of such a model should appropriate reviewing authority (paragraph
c. Population density should be used with be determined in consultation with the 3.0(b)). AERMOD contains algorithms for dry
caution and generally not be applied to appropriate reviewing authority (paragraph and wet deposition of gases and particles.80
highly industrialized areas where the 3.0(b)) consistent with limitations of For other Gaussian plume models, an
population density may be low and thus a paragraph 3.2.2(e). The meteorological input ‘‘infinite half-life’’ may be used for estimates
rural classification would be indicated. data requirements for developing the time of particle concentrations when only
However, the area is likely to be sufficiently and space varying three-dimensional winds exponential decay terms are used for treating
built-up so that the urban land use criteria and dispersion meteorology for these settling and deposition. Lagrangian models
would be satisfied. Therefore, in this case, situations are discussed in paragraph have varying degrees of complexity for
the classification should be ‘‘urban’’ and 8.4.1.2(c). Examples of inhomogeneous dealing with settling and deposition and the
urban dispersion parameters should be used. winds include, but are not limited to, selection of a parameterization for such
d. For applications of AERMOD in urban situations described in the following should be included in the approval process
areas, under either the Land Use Procedure paragraphs: for selecting a Lagrangian model. Eulerian
or the Population Density Procedure, the user i. Inversion breakup fumigation. Inversion grid models tend to have explicit
needs to estimate the population of the urban breakup fumigation occurs when a plume (or parameterizations for gravitational settling
area affecting the modeling domain because multiple plumes) is emitted into a stable and deposition as well as wet deposition
the urban influence in AERMOD is scaled layer of air and that layer is subsequently parameters already included as part of the
based on a user-specified population. For mixed to the ground through convective chemistry scheme.
non-population oriented urban areas, or areas transfer of heat from the surface or because 7.2.2 Stationary Sources
influenced by both population and industrial of advection to less stable surroundings.
activity, the user will need to estimate an Fumigation may cause excessively high 7.2.2.1 Good Engineering Practice Stack
equivalent population to adequately account concentrations but is usually rather short- Height
for the combined effects of industrialized lived at a given receptor. There are no a. The use of stack height credit in excess
areas and populated areas within the recommended refined techniques to model of Good Engineering Practice (GEP) stack
modeling domain. Selection of the this phenomenon. There are, however, height or credit resulting from any other
appropriate population for these applications screening procedures 40 that may be used to dispersion technique is prohibited in the
should be determined in consultation with approximate the concentrations. development of emissions limits by 40 CFR
the appropriate reviewing authority Considerable care should be exercised in 51.118 and 40 CFR 51.164. The definition of
GEP stack height and dispersion technique c. Stack tip downwash generally occurs appropriate input data available should
are contained in 40 CFR 51.100. Methods and with poorly constructed stacks and when the always be selected for use in modeling
procedures for making the appropriate stack ratio of the stack exit velocity to wind speed analyses. Modeled concentrations can vary
height calculations, determining stack height is small. An algorithm developed by Briggs 86 widely depending on the source data or
credits and an example of applying those is the recommended technique for this meteorological data used. This section
techniques are found in several situation and is used in preferred models for attempts to minimize the uncertainty
references,81 82 83 84 which provide a great point sources. associated with database selection and use by
deal of additional information for evaluating 7.2.3 Mobile Sources identifying requirements for input data used
and describing building cavity and wake in modeling. More specific data requirements
a. Emissions of primary pollutants from and the format required for the individual
effects.
mobile sources can be modeled with an models are described in detail in the users’
b. If stacks for new or existing major appropriate model identified in section 4.2.
sources are found to be less than the height guide and/or associated documentation for
Screening of mobile sources can be each model.
defined by the EPA’s refined formula for accomplished by using screening
determining GEP height, then air quality meteorology, such as that generated by the 8.1 Modeling Domain
impacts associated with cavity or wake MAKEMET component of AERSCREEN,
effects due to the nearby building structures 8.1.1 Discussion
which can generate a range of meteorological
should be determined. The EPA refined scenarios using site-specific characteristics, a. The modeling domain is the geographic
formula height is defined as H + 1.5L.83 Since such as albedo, Bowen ratio, and surface area for which the required air quality
the definition of GEP stack height defines roughness. Maximum hourly concentrations analyses for the NAAQS and PSD increments
excessive concentrations as a maximum computed from screening runs can be are conducted.
ground-level concentration due in whole or converted to longer averaging periods using 8.1.2 Requirements
in part to downwash of at least 40 percent the scaling ratios specific in the AERSCREEN a. For a NAAQS or PSD increment
in excess of the maximum concentration User’s Guide.37 assessment, the modeling domain or project’s
without downwash, the potential air quality b. Mobile sources can be modeled in impact area shall include all locations where
impacts associated with cavity and wake AERMOD as either line (i.e., elongated area) the emissions of a pollutant from the new or
effects should also be considered for stacks sources or as a series of volume sources. modifying source(s) may cause a significant
that equal or exceed the EPA formula height However, since mobile source modeling ambient impact. This impact area is defined
for GEP. The AERSCREEN model can be used usually includes an analysis of very near- as an area with a radius extending from the
to obtain screening estimates of potential source impacts (e.g., hot-spot modeling, new or modifying source to: (1) The most
downwash influences, based on the PRIME which can include receptors within 5–10 distant point source where air quality
downwash algorithm incorporated in the meters of the roadway), the results can be modeling predicts a significant ambient
AERMOD model. If more refined highly sensitive to the characterization of the impact will occur, or (2) the nominal 50 km
concentration estimates are required, the mobile emissions. When modeling roadway distance considered applicable for Gaussian
recommended steady-state plume dispersion links, such as highway and arterial links, the dispersion models, whichever is less. The
model in section 4.2.2, AERMOD, should be EPA recommends that line/area sources required air quality analysis shall be carried
used. instead of volume sources be used whenever out within this geographical area with
7.2.2.2 Plume Rise possible, as it is easier to characterize them characterization of source impacts, nearby
correctly. Important characteristics for both source impacts, and background
a. The plume rise methods of Briggs 85 86
line/area and volume sources include the concentrations, as recommended later in this
are incorporated in many of the preferred
plume release height, source width, and section.
models and are recommended for use in initial dispersion characteristics, which
many modeling applications. In b. For SIP attainment demonstrations for
should also take into account the impact of ozone and PM2.5, or regional haze reasonable
AERMOD,44 45 for the stable boundary layer, traffic-induced turbulence, which can cause
plume rise is estimated using an iterative progress goal analyses, the modeling domain
roadway sources to have larger initial is determined by the nature of the problem
approach, similar to that in the CTDMPLUS dimensions than might normally be used for
model. In the convective boundary layer, being modeled and the spatial scale of the
representing line sources. emissions which impact the nonattainment
plume rise is superposed on the c. The EPA’s quantitative PM hot-spot
displacements by random convective or Class I area(s). The modeling domain shall
guidance 61 and Haul Road Workgroup Final be designed so that all major upwind source
velocities.87 In AERMOD, plume rise is Report 63 provide guidance on the
computed using the methods of Briggs except areas that influence the downwind
appropriate characterization of mobile nonattainment area are included in addition
cases involving building downwash, in sources as a function of the roadway and
which a numerical solution of the mass, to all monitor locations that are currently or
vehicle characteristics. The EPA’s
energy, and momentum conservation laws is recently violating the NAAQS or close to
quantitative PM hot-spot guidance includes
performed.88 No explicit provisions in these violating the NAAQS in the nonattainment
important considerations and should be
models are made for multistack plume rise area. Similarly, all Class I areas to be
consulted when modeling roadway links.
enhancement or the handling of such special evaluated in a regional haze modeling
Line or area sources are recommended for
plumes as flares; these problems should be application shall be included and sufficiently
mobile sources. However, if volume sources
considered on a case-by-case basis. distant from the edge of the modeling
are used, it is particularly important to insure
b. Gradual plume rise is generally domain. Guidance on the determination of
that roadway emissions are appropriately
recommended where its use is appropriate: the appropriate modeling domain for
spaced when using volume source so that the
(1) In AERMOD; (2) in complex terrain emissions field is uniform across the photochemical grid models in demonstrating
screening procedures to determine close-in roadway. Additionally, receptor placement is attainment of these air quality goals is
impacts and (3) when calculating the effects particularly important for volume sources, available.60 Users should consult the latest
of building wakes. The building wake which have ‘‘exclusion zones’’, where version of this guidance for the most current
algorithm in AERMOD incorporates and concentrations are not calculated for modeling guidance and with the appropriate
exercises the thermodynamically based receptors located ‘‘within’’ the volume reviewing authority (paragraph 3.0(b)) for any
application specific guidance which is
gradual plume rise calculations as described sources, i.e., less than 2.15 times the initial
in paragraph (a) of this subsection. If the lateral dispersion coefficient from the center beyond the scope of this section.
building wake is calculated to affect the of the volume.61 Placing receptors in these 8.2 Source Data
plume for any hour, gradual plume rise is ‘‘exclusion zones’’ will result in
also used in downwind dispersion underestimates of roadway impacts. 8.2.1 Discussion
calculations to the distance of final plume a. Sources of pollutants can be classified as
rise, after which final plume rise is used. 8.0 Model Input Data point, line, area, and volume sources. Point
Plumes captured by the near wake are re- a. Databases and related procedures for sources are defined in terms of size and may
emitted to the far wake as a ground-level estimating input parameters are an integral vary between regulatory programs. The line
volume source. part of the modeling process. The most sources most frequently considered are
roadways and streets along which there are 8.2.2 Requirements the establishment of the appropriate
well-defined movements of motor vehicles. a. For SIP attainment demonstrations for emissions inputs for regulatory modeling
They may also be lines of roof vents or the purpose of projecting future year NAAQS applications with respect to PSD assessments
stacks, such as in aluminum refineries. Area attainment for ozone, PM2.5, and regional for a proposed new or modifying source.
and volume sources are often collections of haze reasonable progress goal analyses, d. For stationary source applications,
a multitude of minor sources with emissions which reflect actual emissions changes in operating conditions that affect
individually small emissions that are during the base modeling year time period the physical emission parameters (e.g.,
impractical to consider as separate point or should be input to models for base year release height, initial plume volume, and exit
line sources. Large area sources are typically modeling. Emissions projections to future velocity) shall be considered to ensure that
treated as a grid network of square areas, years should account for key variables such maximum potential impacts are
with pollutant emissions distributed as growth due to increased or decreased appropriately determined in the assessment.
uniformly within each grid square. Generally, activity, expected emissions controls due to For example, the load or operating condition
input data requirements for air quality regulations, settlement agreements or consent for point sources that causes maximum
models necessitate the use of metric units. As decrees, fuel switches, and any other relevant ground-level concentrations shall be
necessary, any English units common to information. Guidance on emissions established. As a minimum, the source
engineering applications should be estimation techniques (including future year should be modeled using the design capacity
appropriately converted to metric. projections) for SIP attainment (100 percent load). If a source operates at
b. For point sources, there are many source demonstrations is available.60 90 greater than design capacity for periods that
characteristics and operating conditions that b. For the purpose of SIP revisions for could result in violations of the NAAQS or
may be needed to appropriately model the stationary point sources, the regulatory PSD increment, this load should be modeled.
facility. For example, the plant layout (e.g., modeling of inert pollutants shall use the Where the source operates at substantially
location of stacks and buildings), stack emissions input data shown in Table 8–1 for less than design capacity, and the changes in
parameters (e.g., height and diameter), boiler short-term and long-term NAAQS. To the stack parameters associated with the
size and type, potential operating conditions, demonstrate compliance and/or establish the operating conditions could lead to higher
and pollution control equipment parameters. appropriate SIP emissions limits, Table 8–1 ground level concentrations, loads such as 50
generally provides for the use of ‘‘allowable’’ percent and 75 percent of capacity should
Such details are required inputs to air quality
emissions in the regulatory dispersion also be modeled. Malfunctions which may
models and are needed to determine
modeling of the stationary point source(s) of result in excess emissions are not considered
maximum potential impacts.
interest. In such modeling, these source(s) to be a normal operating condition. They
c. Modeling mobile emissions from streets
should be modeled sequentially with these generally should not be considered in
and highways requires data on the road
loads for every hour of the year. As part of determining allowable emissions. However,
layout, including the width of each traveled a cumulative impact analysis, Table 8–1 if the excess emissions are the result of poor
lane, the number of lanes, and the width of allows for the model user to account for maintenance, careless operation, or other
the median strip. Additionally, traffic actual operations in developing the preventable conditions, it may be necessary
patterns should be taken into account (e.g., emissions inputs for dispersion modeling of to consider them in determining source
daily cycles of rush hour, differences in nearby sources, while other sources are best impact. A range of operating conditions
weekday and weekend traffic volumes, and represented by air quality monitoring data. should be considered in screening analyses;
changes in the distribution of heavy-duty Consultation with the appropriate reviewing the load causing the highest concentration, in
trucks and light-duty passenger vehicles), as authority (paragraph 3.0(b)) is advisable on addition to the design load, should be
these patterns will affect the types and the establishment of the appropriate included in refined modeling.
amounts of pollutant emissions allocated to emissions inputs for regulatory modeling e. Emissions from mobile sources also have
each lane, and the height of emissions. applications with respect to SIP revisions for physical and temporal characteristics that
d. Emission factors can be determined stationary point sources. should be appropriately accounted for. For
through source specific testing and c. For the purposes of demonstrating example, an appropriate emissions model
measurements (e.g., stack test data) from NAAQS compliance in a PSD assessment, the shall be used to determine emissions profiles.
existing sources or provided from a regulatory modeling of inert pollutants shall Such emissions should include speciation
manufacturing association or vendor. use the emissions input data shown in Table specific for the vehicle types used on the
Additionally, emissions factors for a variety 8–2 for short and long-term NAAQS. The roadway (e.g., light duty and heavy duty
of source types are compiled in an EPA new or modifying stationary point source trucks) and subsequent parameterizations of
publication commonly known as AP–42.89 shall be modeled with ‘‘allowable’’ emission the physical emissions characteristics (e.g.,
AP–42 also provides an indication of the in the regulatory dispersion modeling. As release height) should reflect those emissions
quality and amount of data on which many part of a cumulative impact analysis, Table sources. For long-term standards, annual
of the factors are based. Other information 8–2 allows for the model user to account for average emissions may be appropriate, but
concerning emissions is available in EPA actual operations in developing the for short-term standards, discrete temporal
publications relating to specific source emissions inputs for dispersion modeling of representation of emissions should be used
categories. The appropriate reviewing nearby sources, while other sources are best (e.g., variations in weekday and weekend
authority (paragraph 3.0(b)) should be represented by air quality monitoring data. traffic or the diurnal rush-hour profile typical
consulted to determine appropriate source For purposes of situations involving of many cities). Detailed information and
definitions and for guidance concerning the emissions trading refer to current EPA policy data requirements for modeling mobile
determination of emissions from and and guidance to establish input data. sources of pollution are provided in the
techniques for modeling the various source Consultation with the appropriate reviewing user’s manuals for each of the models
types. authority (paragraph 3.0(b)) is advisable on applicable to mobile sources.61 63
TABLE 8–1—POINT SOURCE MODEL EMISSION INPUT FOR SIP REVISIONS OF INERT POLLUTANTS 1
Emissions limit Operating level Operating factor
× ×
Averaging time (lb/MMBtu) 2 (lb/MMBtu) 2 (e.g., hr/yr. hr/day)
Stationary Point Source(s) Subject to SIP Emissions Limit(s) Evaluation for Compliance With Ambient Standards
(Including Areawide Demonstrations)
Annual & quarterly ..................... Maximum allowable emission Actual or design capacity Actual operating factor aver-
limit or federal enforceable (whichever is greater), or fed- aged over the most recent 2
permit limit. erally permit enforceable per- years.3
mit condition.
TABLE 8–1—POINT SOURCE MODEL EMISSION INPUT FOR SIP REVISIONS OF INERT POLLUTANTS 1—Continued
Emissions limit × Operating level × Operating factor
Averaging time (lb/MMBtu) 2 (lb/MMBtu) 2 (e.g., hr/yr. hr/day)
Short term (≤24 hours) .............. Maximum allowable emission Actual or design capacity Continuous operation, i.e., all
limit or federally enforceable (whichever is greater), or fed- hours of each time period
permit limit. erally enforceable permit con- under consideration (for all
dition.4 hours of the meteorological
database).5
Nearby Source(s).6
Annual & quarterly ..................... Maximum allowable emission Annual level when actually op- Actual operating factor aver-
limit or federal enforceable erating, averaged over the aged over the most recent 2
permit limit.5 most recent 2 years.3 years.3 8
Short term (≤24 hours) .............. Maximum allowable emission Temporally representative level Continuous operation, i.e., all
limit or federal enforceable when actually operating, re- hours of each time period
permit limit.6 flective of the most recent 2 under consideration (for all
years.3 7 hours of the meteorological
database).5
Other Source(s) 8 9
The ambient impacts from Non-nearby or Other Sources (e.g., natural sources, minor sources and, distant major source and unidentified
sources) can be represented by air quality monitoring data unless adequate data do not exist.
1 For purposes of emissions trading, NSR, or PSD, other model input criteria may apply. See Section 8.2 for more information regarding attain-
ment demonstrations of primary PM2.5.
2 Terminology applicable to fuel burning sources; analogous terminology (e.g., lb/throughput) may be used for other types of sources.
3 Unless it is determined that this period is not representative.
4 Operating levels such as 50 percent and 75 percent of capacity should also be modeled to determine the load causing the highest concentra-
tion.
5 If operation does not occur for all hours of the time period of consideration (e.g., 3 or 24–hours) and the source operation is constrained by a
federally enforceable permit condition, an appropriate adjustment to the modeled emission rate may be made (e.g., if operation is only 8 a.m. to
4 p.m. each day, only these hours will be modeled with emissions from the source. Modeled emissions should not be averaged across non-oper-
ating
6 See Section 8.3.3.
7 Temporally representative operating level could be based on Continuous Emissions Monitoring (CEM) data or other information and should
be determined through consultation with the appropriate reviewing authority (Paragraph 3.0(b)).
8 For those permitted sources not in operation or that have not established an appropriate factor, continuous operation (i.e., 8760) should be
used.
TABLE 8–2—POINT SOURCE MODEL EMISSION INPUT FOR NAAQS COMPLIANCE IN PSD DEMONSTRATIONS
Emissions limit Operating level Operating factor
Averaging time × ×
(lb/MMBtu) 1 (lb/MMBtu) 2 (e.g., hr/yr. hr/day)
Proposed Major New or Modified Source
Annual & quarterly ..................... Maximum allowable emission Design capacity or federally en- Continuous operation (i.e., 8760
limit or federal enforceable forceable permit condition. hours).2
permit limit.
Short term (≤24 hours) .............. Maximum allowable emission Design capacity or federally en- Continuous operation, i.e., all
limit or federal enforceable forceable permit condition.3 hours of each time period
permit limit. under consideration (for all
hours of the meteorological
database).2
Nearby Source(s) 4 5
Annual & quarterly ..................... Maximum allowable emission Annual level when actually op- Actual operating factor aver-
limit or federal enforceable erating, averaged over the aged over the most recent 2
permit limit.5 most recent 2 years.6 years.6 8
Short term (≤24 hours) .............. Maximum allowable emission Annual level when actually op- Continuous operation, i.e., all
limit or federal enforceable erating, averaged over the hours of each time period
permit limit.5 most recent 2 years.6 7 under consideration (for all

hours of the meteorological
database).2
Other Source(s) 5 9
The ambient impacts from Non-nearby or Other Sources (e.g., natural sources, minor sources and,distant major sources, and unidentified
sources) can be represented by air quality monitoring data unless adequate data do not exist.
1 Terminology applicable to fuel burning sources; analogous terminology (e.g., lb/throughput) may be used for other types of sources.
2 If operation does not occur for all hours of the time period of consideration (e.g., 3 or 24–hours) and the source operation is constrained by a
federally enforceable permit condition, an appropriate adjustment to the modeled emission rate may be made (e.g., if operation is only 8 a.m. to
4 p.m. each day, only these hours will be modeled with emissions from the source. Modeled emissions should not be averaged across non-oper-
ating
3 Operating levels such as 50 percent and 75 percent of capacity should also be modeled to determine the load causing the highest concentra-
tion.
4 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 for major modification.
6 Unless it is determined that this period is not representative.
7 Temporally representative operating level could be based on Continuous Emissions Monitoring (CEM) data or other information and should
be determined through consultation with the appropriate reviewing authority (Paragraph 3.0(b)).
8 For those permitted sources not in operation or that have not established an appropriate factor, continuous operation (i.e., 8760) should be
used.
8.3 Background Concentrations modeling approach does provide for an values when the source in question is
8.3.1 Discussion appropriate atmospheric environment to impacting the monitor. In such cases,
assess single-sources impacts for ozone and monitoring sites inside a 90° sector
a. Background concentrations are essential secondary PM2.5. downwind of the source may be used to
in constructing the design concentration, or d. For PSD assessments in general and SIP determine the area of impact.
total air quality concentration, as part of a attainment demonstrations for inert ii. There may be other circumstances
cumulative impact analysis for NAAQS and pollutants, the development of the which would necessitate modifications to the
PSD increments (section 9.2.4). Background appropriate background concentration for a ambient data record. Such cases could
air quality should not include the ambient cumulative impact analysis involves proper include removal of data from specific days or
impacts of the project source under accounting of each contribution to the design hours when a monitor is being impacted
consideration. Instead, it should include: concentration and will depend upon whether activities that are not typical or expected to
i. Nearby sources: These are individual the project area’s situation consists of either occur again in the future (e.g., construction,
sources in the vicinity of the source(s) under an isolated single source(s) or a multitude of roadway repairs, forest fires, or unusual
consideration for emissions limits that are sources. agricultural activities). There may also be
not adequately represented by ambient cases where scaling (multiplying the
monitoring data. Typically, sources that 8.3.2 Recommendations for Isolated Single
Source monitored concentrations with a scaling
cause a significant concentration gradient in factor) or adjusting (adding or subtracting a
the vicinity of the source(s) under a. In areas with an isolated source(s), constant value the monitored concentrations)
consideration for emissions limits are not determining the appropriate background of data from specific days or hours. Such
adequately represented by background concentration should focus on adjustments would make the monitored
ambient monitoring. The ambient characterization of contributions from all background concentrations more temporally
contributions from these nearby sources are other sources through adequately and/or spatially representative of area around
thereby accounted for by explicitly modeling representative ambient monitoring data. the new or modifying source for the purposes
their emissions (section 8.2). b. The EPA recommends use of the most of the regulator assessment.
ii. Other sources: That portion of the recent quality assured air quality monitoring iii. For short-term standards, the diurnal or
background attributable to natural sources, data collected in the vicinity of the source to seasonal patterns of the air quality
other unidentified sources in the vicinity of determine the background concentration for monitoring data may differ significantly from
the project, and regional transport the averaging times of concern. In most cases, the patterns associated with the modeled
contributions from more distant sources the EPA recommends using data from the concentrations. When this occurs, it may be
(domestic and international). The ambient monitor closest to and upwind of the project appropriate to pair the air quality monitoring
contributions from these sources are typically area. If several monitors are available, data in a temporal manner that reflects these
accounted for through use of ambient preference should be given to the monitor patterns (e.g., pairing by season and/or hour
monitoring data or, in some cases, regional- with the most similar characteristics as the of day).92
scale photochemical grid modeling results. project area. If there are no monitors located iv. For situations where monitored air
b. The monitoring network used for in the vicinity of the new or modify source, quality concentrations vary across the
developing background concentrations is a ‘‘regional site’’ may be used to determine modeling domain, it may be appropriate to
expected to conform to the same quality background concentrations. A regional site is consider air quality monitoring data from
assurance and other requirements as those one that is located away from the area of multiple monitors within the project area.
networks established for PSD purposes.91 interest but is impacted by similar or d. Determination of the appropriate
Accordingly, the air quality monitoring data adequately representative sources. background concentrations should be
should be of sufficient completeness and c. Many of the challenges related to consistent with appropriate EPA modeling
follow appropriate data validation cumulative impact analyses arise in the guidance 59 92 and justified in the modeling
procedures. These data should be adequately context of defining the appropriate metric to protocol that is vetted with the appropriate
representative of the area to inform characterize background concentrations from reviewing authority (paragraph 3.0(b)).
calculation of the design concentration for ambient monitoring data and determining the e. Considering the spatial and temporal
comparison to the applicable NAAQS appropriate method for combining this variability throughout a typical modeling
(section 9.2.2) monitor-based background contribution to domain on an hourly basis and the
c. For photochemical grid modeling the modeled impact of the project and other complexities and limitations of hourly
conducted in SIP attainment demonstrations nearby sources. For many cases, the best observations from the ambient monitoring
for ozone, PM2.5 and regional haze, the starting point would be use of the current network, the EPA does not recommend
emissions from nearby and other sources are design value for the applicable NAAQS as a hourly or daily pairing of monitored
included as model inputs and fully uniform monitored background contribution background and modeled concentrations
accounted for in the modeling application across the project area. However, there are except in rare cases of relatively isolated
and predicted concentrations. The concept of cases in which the current design value may sources where the available monitor can be
adding individual components to develop a not be appropriate. Such cases include but shown to be representative of the ambient
design concentration, therefore, do not apply are not limited to: concentration levels in the areas of maximum
in these SIP applications. However, such i. For situations involving a modifying impact from the proposed new source. The
modeling results may then be appropriate for source where the existing facility is implicit assumption underlying hourly
consideration in characterizing background determined to impact the ambient monitor, pairing is that the background monitored
concentrations for other regulatory the background concentration at each levels for each hour are spatially uniform and
applications. Also, as noted in section 5, this monitor can be determined by excluding that the monitored values are fully
representative of background levels at each of a concentration gradient will be greatest in (e.g., natural sources, minor and distance
receptor for each hour. Such an assumption the proximity of the source and will major sources) should be accounted for
clearly ignores the many factors that generally not be significant at distances through use of ambient monitoring data and
contribute to the temporal and spatial greater than 10 times the height of the determined by the procedures found in
variability of ambient concentrations across a stack(s) at that source without consideration section 8.3.2 in keeping with eliminating or
typical modeling domain on an hourly basis. of terrain influences. reducing the source-oriented impacts from
In most cases, the seasonal (or quarterly) iii. The number of nearby sources to be nearby sources to avoid potential double-
pairing of monitored and modeled explicitly modeled in the air quality analysis counting of modeled and monitored
concentrations should sufficiently address is expected to be few except in unusual contributions.
situations to which the impacts from situations. In most cases, the few nearby
modeled emissions are not temporally sources will be located within 10 to 20 km 8.4 Meteorological Input Data
correlated with background monitored levels. from the source(s) under consideration. 8.4.1 Discussion
f. In those cases where adequately Owing to both the uniqueness of each a. This subsection covers meteorological
representative monitoring data to modeling situation and the large number of input data for use in dispersion modeling for
characterize background concentrations are variables involved in identifying nearby regulatory applications and is separate from
not available, it may be appropriate to use sources, no attempt is made here to recommendations made for photochemical
results from a regional-scale photochemical comprehensively define a ‘‘significant grid modeling. Recommendations for
grid model or other representative model concentration gradient.’’ Rather, meteorological data for photochemical grid
application as background concentrations identification of nearby sources calls for the modeling applications are outlined in the
consistent with the considerations discussed exercise of professional judgement by the latest version of EPA’s Guidance on the Use
above and in consultation with the appropriate reviewing authority (paragraph of Models and Other Analyses for
appropriate reviewing authority (paragraph 3.0(b)). This guidance is not intended to alter Demonstrating Attainment of Air Quality
3.0(b)). the exercise of that judgement or to Goals for Ozone, PM2.5, and Regional Haze 93.
8.3.3 Recommendations for Multi-Source comprehensively prescribe which sources In cases where Lagrangian models are
Areas should be included as nearby sources. applied for regulatory purposes, appropriate
c. For cumulative impact analyses of short- meteorological inputs should be determined
a. In multi-source areas, determining the term and annual ambient standards, the
appropriate background concentration in consultation with the appropriate
nearby sources as well as the project reviewing authority (paragraph 3.0(b)).
involves: (1) identification and source(s) must be evaluated using an
characterization of contributions from nearby b. The meteorological data used as input to
appropriate appendix A model or approved a dispersion model should be selected on the
sources through explicit modeling, and (2)
alternative model with the emission input basis of spatial and climatological (temporal)
characterization of contributions from other
data shown in Table 8–1 or 8–2. representativeness as well as the ability of
sources through adequately representative
i. When modeling a nearby source that the individual parameters selected to
ambient monitoring data. A key point here is
does not have a permit and the emissions characterize the transport and dispersion
the interconnectedness of each component in
limits contained in the SIP for a particular conditions in the area of concern. The
that the question of which nearby sources to
source category is greater than the emissions representativeness of the measured data is
include in the cumulative modeling is
possible given the source’s maximum dependent on numerous factors including
inextricably linked to the question of what
physical capacity to emit, the ‘‘maximum but not limited to: (1) The proximity of the
the ambient monitoring data represents
within the project area. allowable emissions limit’’ for such a nearby meteorological monitoring site to the area
b. Nearby sources: All sources in the source may be calculated as the emissions under consideration; (2) The complexity of
vicinity of the source(s) under consideration rate representative of the nearby source’s the terrain; (3) The exposure of the
for emissions limits that are not adequately maximum physical capacity to emit, meteorological monitoring site; and (4) The
represented by ambient monitoring data considering its design specifications and period of time during which data are
should be explicitly modeled. Since an allowable fuels and process materials. collected. The spatial representativeness of
ambient monitor is limited to characterizing However, the burden is on the permit the data can be adversely affected by large
air quality at a fixed location, sources that applicant to sufficiently document what the distances between the source and receptors
causes a significant concentration gradient in maximum physical capacity to emit is for of interest and the complex topographic
the vicinity of the source(s) under such a nearby source. characteristics of the area. Temporal
consideration for emissions limits are not ii. It is appropriate to model nearby sources representativeness is a function of the year-
likely to be adequately characterized by the only during those times when they, by their to-year variations in weather conditions.
monitored data due to the high degree of nature, operate at the same time as the Where appropriate, data representativeness
variability of the source’s impact. primary source(s). Accordingly, it is not should be viewed in terms of the
i. The pattern of concentration gradients necessary to model impacts of a nearby appropriateness of the data for constructing
can vary significantly based on the averaging source that does not, by its nature, operate at realistic boundary layer profiles and, where
period being assessed. In general, the same time as the primary source, applicable, three-dimensional meteorological
concentration gradients will be smaller and regardless of an identified significant fields, as described in paragraphs (c) and (d)
more spatially uniform for annual averages concentration gradient from the nearby of this subsection.
than for short-term averages, especially for source. The burden is on the permit c. The meteorological data should be
hourly averages. The spatial distribution of applicant to adequately justify the exclusion adequately representative and may be site-
annual impacts around a source will often of nearby sources to the satisfaction of the specific data, data from a nearby National
have a single peak downwind of the source appropriate reviewing authority (paragraph Weather Service (NWS) or comparable
based on the prevailing wind direction, 3.0(b)). The following examples illustrate two station, or prognostic meteorological data.
except in cases where terrain or other cases in which a nearby source may be The implementation of ASOS (automated
geographic effects are important. By contrast, shown not to operate at the same time as the surface observing stations) in recent years
the spatial distribution of peak short-term primary source(s) being modeled: (1) should not preclude the use of NWS–ASOS
impacts will typically show several localized Seasonal sources (only used during certain data if such a station is determined to be
concentration peaks with more significant seasons of the year). Such sources would not representative of the modeled area.94
gradient. be modeled as nearby sources during times d. Model input data are normally obtained
ii. Concentration gradients associated with in which they do not operate; and (2) either from the NWS or as part of a site-
a particular source will generally be largest Emergency backup generators, to the extent specific measurement program. State
between that source’s location and the that they do not operate simultaneously with climatology offices, local universities, FAA,
distance to the maximum ground-level the sources that they back up. Such military stations, industry and pollution
concentrations from that source. Beyond the emergency equipment would not be modeled control agencies may also be sources of such
maximum impact distance, concentration as nearby sources. data. In specific cases, prognostic
gradients will generally be much smaller and d. Other sources. That portion of the meteorological data may be appropriate for
more spatially uniform. Thus, the magnitude background attributable to all other sources use and obtained from similar sources. Some
recommendations and requirements for the height to be adequately representative, data to the needed model input. Site-specific
use of each type of data are included in this especially for sources located in complex measurements of model input parameters
subsection. terrain. Whereas, for a variable such as have been made for many modeling studies,
8.4.2 Recommendations and Requirements temperature, data from a station several and those methods and techniques are
kilometers away from the source may be becoming more widely applied, especially in
a. AERMET 95 shall be used to preprocess considered to be adequately representative. situations such as complex terrain
all meteorological data, be it observed or More information about meteorological data, applications, where available NWS data are
prognostic, for use with AERMOD in representativeness, and surface not adequately representative. However,
regulatory applications. The AERMINUTE 96 characteristics can be found in the AERMOD there are many modeling applications where
processor, in most cases, should be used to Implementation Guide 76. NWS data are adequately representative, and
process 1-minute ASOS wind data for input c. Regulatory application of CTDMPLUS the applications still rely heavily on the NWS
into AERMET when processing NWS ASOS requires the input of multi-level data.
sites in AERMET. When processing measurements of wind speed, direction, b. Many models use the standard hourly
prognostic meteorological data for AERMOD, temperature, and turbulence from an weather observations available from the
the Mesoscale Model Interface Program appropriately sited meteorological tower. The National Centers for Environmental
(MMIF) 93 should be used to process data for measurements should be obtained up to the Information (NCEI) b. These observations are
input into AERMET. Other methods of representative plume height(s) of interest.
processing prognostic meteorological data for then preprocessed before they can be used in
Plume heights of interest can be determined the models. Prior to the advent of ASOS in
input into AERMET should be approved by by use of screening procedures such as
the appropriate reviewing authority. the early 1990’s, the ‘‘hourly’’ weather
CTSCREEN. observation was a human observer-based
Additionally, the following meteorological d. Regulatory application of OCD requires
preprocessors are recommended by the EPA: observation reflecting a single 2-minute
meteorological data over land and over water. average generally taken about 10 minutes
PCRAMMET 97, MPRM 98, and METPRO 99. The over land or surface data processed
PCRAMMET is the recommended before the hour. However, beginning with
through PCRAMMET 97 which provides January 2000 for first-order stations and
meteorological data preprocessor for use in hourly stability class, wind direction and
applications of OCD employing hourly NWS March 2005 for all stations, NCEI has
speed, ambient temperature, and mixing archived the rolling 2-minute average winds
data. MPRM is the recommended height are required. Data over water requires
meteorological data preprocessor for at every minute for ASOS sites. The
hourly mixing height, relative humidity, air AERMINUTE processor 96 was developed to
applications of OCD employing site-specific temperature, and water surface temperature.
meteorological data. METPRO is the reduce calm and missing hours by taking
Missing winds are substituted with the advantage of the availability of the 1-minute
recommended meteorological data surface winds. Vertical wind direction shear,
preprocessor for use with CTDMPLUS.100 ASOS wind data to calculate full hourly
vertical temperature gradient, and turbulence average winds to replace standard hourly
b. Regulatory application of AERMOD intensities are optional.
necessitates careful consideration of the observations and reduce the number of calm
e. The model user should acquire enough and missing winds in AERMET processing.
meteorological data for input to AERMET. meteorological data to ensure that worst-case
Data representativeness, in the case of meteorological conditions are adequately 8.4.3.2 Recommendations
AERMOD, means utilizing data of an represented in the model results. The use of a. The preferred models listed in appendix
appropriate type for constructing realistic 5 years of adequately representative NWS A all accept as input the NWS meteorological
boundary layer profiles. Of particular meteorological data, at least 1 year of site- data preprocessed into model compatible
importance is the requirement that all specific, or at least 3 years of prognostic form. If NWS data are judged to be
meteorological data used as input to meteorological data are required. If 1 year or adequately representative for a specific
AERMOD should be adequately more, up to 5 years, of site-specific data is modeling application, they may be used.
representative of the transport and dispersion available, these data are preferred for use in NEIS makes available surface 105 106 and
within the analysis domain. Where surface air quality analyses. Such data should have upper air 107 meteorological data online and
conditions vary significantly over the been subjected to quality assurance in CD–ROM format. Upper air data are also
analysis domain, the emphasis in assessing procedures as described in section 8.4.4.2.
representativeness should be given to available at the Earth System Research
f. Objective analysis in meteorological Laboratory Global Systems Divisions Web
adequate characterization of transport and
modeling is to improve meteorological site (http://esrl.noaa.gov/gsd).
dispersion between the source(s) of concern
analyses (the ‘‘first guess field’’) used as b. Although most NWS wind
and areas where maximum design
initial conditions for prognostic measurements are made at a standard height
concentrations are anticipated to occur. The
meteorological models by incorporating of 10 meters, the actual anemometer height
EPA recommends that the surface
information from meteorological should be used as input to the preferred
characteristics input to AERMET should be
observations. Direct and indirect (using meteorological processor and model.
representative of the land cover in the
remote sensing techniques) observations of c. Standard hourly NWS wind directions
vicinity of the meteorological data, i.e., the
temperature, humidity, and wind from are reported to the nearest 10 degrees. A
location of the meteorological tower for
measured data or the representative grid cell surface and radiosonde reports are commonly specific set of randomly generated numbers
for prognostic data. Therefore, the model user employed to improve these analysis fields. has been developed for use with the
should apply the latest version For LRT applications, it is recommended that preferred EPA models and should be used
AERSURFACE 101 102, where applicable, for objective analysis procedures using direct with standard NWS data to ensure a lack of
determining surface characteristics when and indirect meteorological observations be bias in wind direction assignments within
processing measured meteorological data employed in preparing input fields to the models.
through AERMET. In areas where it is not produce prognostic meteorological datasets. d. Beginning with year 2000, NCDC began
possible to use AERSURFACE output, surface The length of record of observations should archiving 2-minute winds, reported every
characteristics can determined using conform to recommendations outlined in minute for NWS ASOS sites. The
techniques that apply the same analysis as paragraph 8.4.2(e) for prognostic AERMINUTE processor was developed to
AERSURFACE. In the case of prognostic meteorological model datasets. read those winds and calculate hourly
meteorological data, the surface 8.4.3 National Weather Service Data average winds for input into AERMET. When
characteristics associated with the prognostic such data are available for the NWS ASOS
meteorological model output for the 8.4.3.1 Discussion site being processed, the AERMINUTE
representative grid cell should be used.103 104 a. The NWS meteorological data are processor should be used in most cases to
Furthermore, since the spatial scope of each routinely available and familiar to most calculate hourly average wind speed and
variable could be different, model users. Although the NWS does not direction when processing NWS ASOS data
representativeness should be judged for each provide direct measurements of all the for input to AERMOD.94
variable separately. For example, for a needed dispersion model input variables,
variable such as wind direction, the data methods have been developed and b Formerly the National Climatic Data Center
should ideally be collected near plume successfully used to translate the basic NWS (NCDC).
e. Data from universities, FAA, military to provide an adequately representative have been obtained traditionally using
stations, industry and pollution control characterization of pollutant transport meteorological sensors mounted on tall
agencies may be used if such data are between sources and receptors of interest. towers. A feasible alternative to tall towers is
equivalent in accuracy and detail (e.g., siting The appropriate reviewing authority the use of meteorological remote sensing
criteria, frequency of observations, data (paragraph 3.0(b)) is available to help instruments (e.g., acoustic sounders or radar
completeness, etc.) to the NWS data, they are determine the appropriateness of the wind profilers) to provide winds aloft,
judged to be adequately representative for the measurement locations. coupled with 10-meter towers to provide the
particular application and have undergone b. All processed site-specific data should near-surface winds. Note that when site-
quality assurance checks. be in the form of hourly averages for input specific wind measurements are used,
f. After valid data retrieval requirements into the dispersion model. These data AERMOD, at a minimum, requires wind
have been met,108 large number of hours in include surface wind speed, transport observations at a height above ground
the record having missing data should be direction, dilution wind speed, and between seven times the local surface
treated according to an established data turbulence measurements sA and sE (for use roughness height and 100 meters. (For
substitution protocol provided that in stability determinations and direct input additional requirements for AERMOD and
adequately representative alternative data are into the dispersion model). The hourly CTDMPLUS, see appendix A.) Specifications
available. Data substitution guidance is average turbulence measurements should be for wind measuring instruments and systems
provided in section 5.3 of reference 108. If no the square root of the arithmetic average of are contained in reference 108.
representative alternative data are available the 15-minute average variances (square of sA h. Turbulence. There are several dispersion
for substitution, the absent data should be or sE). models that are capable of using direct
coded as missing using missing data codes c. Missing data substitution. After valid measurements of turbulence (wind
appropriate to the applicable meteorological data retrieval requirements have been met,108 fluctuations) in the characterization of the
pre-processor. Appropriate model options for hours in the record having missing data vertical and lateral dispersion (e.g.,
treating missing data, if available in the should be treated according to an established CTDMPLUS, AERMOD). For specific
model, should be employed. data substitution protocol provided that requirements for CTDMPLUS, AERMOD, see
adequately representative alternative data are appendix A. For technical guidance on
8.4.4 Site-Specific data
available. Such protocols are usually part of measurement and processing of turbulence
8.4.4.1 Discussion the approved monitoring program plan. Data parameters, see reference 108. When
a. Spatial or geographical substitution guidance is provided in section turbulence data are used in this manner to
representativeness is best achieved by 5.3 of reference 108. If no representative directly characterize the vertical and lateral
collection of all of the needed model input alternative data are available for substitution, dispersion, the averaging time for the
data in close proximity to the actual site of the absent data should be coded as missing turbulence measurements should be 1 hour.
the source(s). Site-specific measured data are using missing data codes appropriate to the However, since AERMOD incorporates an
therefore preferred as model input, provided applicable meteorological pre-processor. algorithm to account for horizontal plume
that appropriate instrumentation and quality Appropriate model options for treating meander under low wind conditions, the
assurance procedures are followed and that missing data, if available in the model, methodology outlined in paragraph 8.4.4.2(b)
the data collected are adequately should be employed. should be used to calculate hourly averages
representative (free from inappropriate local d. Solar radiation measurements. Total of sq, based on four 15-minuite values, to
or microscale influences) and compatible solar radiation or net radiation should be minimize ‘‘double counting’’ of plume spread
with the input requirements of the model to measured with a reliable pyranometer or net associated with meander. The calculation of
be used. It should be noted that, while site- radiometer, sited and operated in accordance hourly sq discussed above is automatically
specific measurements are frequently made with established site-specific meteorological applied within AERMET when sub-hourly
‘‘on-property’’ (i.e., on the source’s premises), guidance.108 111 data are processed. There are other
acquisition of adequately representative site- e. Temperature measurements. dispersion models that employ P–G stability
specific data does not preclude collection of Temperature measurements should be made categories for the characterization of the
data from a location off property. Conversely, at standard shelter height (2m) in accordance vertical and lateral dispersion. Methods for
collection of meteorological data on a with established site-specific meteorological using site-specific turbulence data for the
source’s property does not of itself guarantee guidance.108 characterization of P–G stability categories
adequate representativeness. For help in f. Temperature difference measurements. are discussed in reference 108. When
determining representativeness of site- Temperature difference (DT) measurements turbulence data are used in this manner to
specific measurements, technical should be obtained using matched determine the P–G stability category, the
guidance 108 is available. Site-specific data thermometers or a reliable thermocouple averaging time for the turbulence
should always be reviewed for system to achieve adequate accuracy. Siting, measurements should be 15 minutes, with
representativeness and adequacy by an probe placement, and operation of DT hourly averaged values based on
experienced meteorologist, atmospheric systems should be based on guidance found methodology in paragraph 8.4.4.2(b).
scientist, or other qualified scientist. in Chapter 3 of reference 108 and such i. Stability categories. For dispersion
guidance should be followed when obtaining models that employ P–G stability categories
8.4.4.2 Recommendations vertical temperature gradient data. AERMET for the characterization of the vertical and
a. The EPA guidance 108 provides may employ the Bulk Richardson scheme, lateral dispersion, the P–G stability
recommendations on the collection and use which requires measurements of temperature categories, as originally defined, couple near-
of site-specific meteorological data. difference, in lieu of cloud cover or surface measurements of wind speed with
Recommendations on characteristics, siting, insolation data. To ensure correct application subjectively determined insolation
and exposure of meteorological instruments and acceptance, AERMOD users should assessments based on hourly cloud cover and
and on data recording, processing, consult with the appropriate reviewing ceiling height observations. The wind speed
completeness requirements, reporting, and authority (paragraph 3.0(b)) before using the measurements are made at or near 10m. The
archiving are also included. This publication Bulk Richardson scheme for their analysis. insolation rate is typically assessed using
should be used as a supplement to other g. Wind measurements. For simulation of observations of cloud cover and ceiling
limited guidance on these subjects.5 91 109 110 plume rise and dispersion of a plume emitted height based on criteria outlined by Turner.72
Detailed information on quality assurance is from a stack, characterization of the wind It is recommended that the P–G stability
also available.111 As a minimum, site-specific profile up through the layer in which the category be estimated using the Turner
measurements of ambient air temperature, plume disperses is desirable. This is method with site-specific wind speed
transport wind speed and direction, and the especially important in complex terrain and/ measured at or near 10m and representative
variables necessary to estimate atmospheric or complex wind situations where wind cloud cover and ceiling height.
dispersion should be available in measurements at heights up to hundreds of Implementation of the Turner method, as
meteorological datasets to be used in meters above stack base may be required in well as considerations in determining
modeling. Care should be taken to ensure some circumstances. For tall stacks when representativeness of cloud cover and ceiling
that meteorological instruments are located site-specific data are needed, these winds height in cases for which site-specific cloud
observations are unavailable, may be found analysis domain. In most cases, this may be direction data where the speed is greater than
in section 6 of reference 108. In the absence the grid cell containing the emission source the wind speed threshold, and the height of
of requisite data to implement the Turner of interest. Since the dispersion modeling the measurement is between seven times the
method, the solar radiation/delta-T (SRDT) may involve multiple sources and the local surface roughness and 100 meters. If the
method or wind fluctuation statistics (i.e., the domain may cover several grid cells, only valid observation of the reference wind
sE and sA methods) may be used. depending on grid resolution of the speed between these heights is less than the
j. The SRDT method, described in section prognostic model, professional judgement threshold, the hour is considered calm, and
6.4.4.2 of reference 108, is modified slightly may be needed to select the appropriate grid no concentration is calculated. None of the
from that published from earlier work 112 and cell to use. In such cases, the selected grid observed wind speeds in a measured wind
has been evaluated with three site-specific cell should be adequately representative of profile that are less than the threshold speed
databases.113 The two methods of stability the entire domain. are used in construction of the modeled wind
classification which use wind fluctuation c. Grid resolution. The grid resolution of speed profile in AERMOD.
statistics, the sE and sA methods, are also the prognostic meteorological data should be
described in detail in section 6.4.4 of considered and evaluated appropriately, 8.4.6.2 Recommendations
reference 108 (note applicable tables in particularly for projects involving complex a. Hourly concentrations calculated with
section 6). For additional information on the terrain. The operational evaluation of the steady-state Gaussian plume models using
wind fluctuation methods, several references modeling data should consider whether a calms should not be considered valid; the
are available.114 115 116 117 finer grid resolution is needed to ensure that wind and concentration estimates for these
8.4.5 Prognostic Meteorological Data the data are representative. The use of output hours should be disregarded and considered
from prognostic mesoscale meteorological to be missing. Critical concentrations for
8.4.5.1 Discussion models is contingent upon the concurrence 3-, 8-, and 24-hour averages should be
a. For some modeling applications, there with the appropriate reviewing authority calculated by dividing the sum of the hourly
may not be a representative NWS or (paragraph 3.0(b)) that the data are of concentrations for the period by the number
comparable meteorological station available acceptable quality. of valid or non-missing hours. If the total
(e.g., complex terrain), and it may be cost 8.4.6 Treatment of Near-Calms and Calms number of valid hours is less than 18 for 24-
prohibitive or infeasible to collect adequately hour averages, less than 6 for 8-hour averages
8.4.6.1 Discussion
representative site-specific data. For these or less than 3 for 3-hour averages, the total
cases, it may be necessary to use prognostic a. Treatment of calm or light and variable concentration should be divided by 18 for the
meteorological data in a regulatory modeling wind poses a special problem in modeling 24-hour average, 6 for the 8-hour average and
application. applications since steady-state Gaussian 3 for the 3-hour average. For annual averages,
b. The EPA has developed a processor, the plume models assume that concentration is the sum of all valid hourly concentrations is
MMIF (Mesoscale Model Interface Program) inversely proportional to wind speed, divided by the number of non-calm hours
to process MM5 (Mesoscale Model 5) or WRF depending on model formulations.
during the year. AERMOD has been coded to
(Weather Research and Forecasting) model Procedures have been developed to prevent
implement these instructions. For hours that
data for input into various models including the occurrence of overly conservative
are calm or missing, the AERMOD hourly
AERMOD. MMIF can process data for input concentration estimates during periods of
concentrations will be zero. For other models
into AERMET or AERMOD for a single grid calms. These procedures acknowledge that a
listed in appendix A, a post-processor
cell or multiple grid cells. MMIF output has steady-state Gaussian plume model does not
apply during calm conditions, and that our computer program, CALMPRO 119 has been
been found to compare favorably against prepared, is available on the EPA’s SCRAM
observed data (site-specific or NWS).118 knowledge of wind patterns and plume
behavior during these conditions does not, at Web site (section 2.3), and should be used.
Specific guidance on processing MMIF for b. Stagnant conditions that include
AERMOD can be found in reference 104. present, permit the development of a better
technique. Therefore, the procedures extended periods of calms often produce
When using MMIF to process prognostic data high concentrations over wide areas for
for regulatory applications, the data should disregard hours which are identified as calm.
The hour is treated as missing and a relatively long averaging periods. The
be processed to generate AERMET inputs and standard steady-state Gaussian plume models
the data subsequently processed through convention for handling missing hours is
recommended. With the advent of the are often not applicable to such situations.
AERMET for input into AERMOD. If an When stagnation conditions are of concern,
alternative method of processing data for AERMINUTE processor, when processing
NWS ASOS data, the inclusion of hourly other modeling techniques should be
input into AERMET is used, it must be considered on a case-by-case basis (see also
approved by the appropriate reviewing averaged winds from AERMINUTE will, in
some instances, dramatically reduce the section 7.2.1.2).
authority (paragraph 3.0(b)). c. When used in steady-state Gaussian
number of calm and missing hours,
8.4.5.2 Recommendations especially when the ASOS wind are derived plume models, measured site-specific wind
a. Prognostic model evaluation. from a sonic anemometer. To alleviate speeds of less than 1 m/s but higher than the
Appropriate effort should be devoted to the concerns about low winds, especially those response threshold of the instrument should
process of evaluating the prognostic introduced with AERMINUTE, the EPA be input as 1 m/s; the corresponding wind
meteorological data. The modeling data implemented a wind speed threshold in direction should also be input. Wind
should be compared to NWS observational AERMET for use with ASOS derived observations below the response threshold of
data in an effort to show that the data are winds.96 Winds below the threshold will be the instrument should be set to zero, with the
accurately replicating the observed treated as calms. input file in ASCII format. For input to
meteorological conditions of the time periods b. AERMOD, while fundamentally a AERMOD, no adjustment should be made to
modeled. An operational evaluation of the steady-state Gaussian plume model, contains the site-specific wind data. For NWS ASOS
modeling data for all model years (i.e., algorithms for dealing with low wind speed data, especially data using the 1-minute
statistical, graphical) should be completed.93 (near calm) conditions. As a result, AERMOD ASOS winds, a wind speed threshold option
The use of output from prognostic mesoscale can produce model estimates for conditions is allowed with a recommended speed of 0.5
meteorological models is contingent upon the when the wind speed may be less than m/s. 94 When using prognostic data processed
concurrence with the appropriate reviewing 1 m/s, but still greater than the instrument by MMIF, a 0.5 m/s threshold is also invoked
authority (paragraph 3.0(b)) that the data are threshold. Required input to AERMET for by MMIF for input into AERMET.
of acceptable quality, which can be site-specific data, the meteorological Observations with wind speeds less than the
demonstrated through statistical comparisons processor for AERMOD, includes a threshold threshold are considered calm, and no
with meteorological observations aloft and at wind speed and a reference wind speed. The concentration is calculated. In all cases
the surface at several appropriate locations.93 threshold wind speed is typically the involving steady-state Gaussian plume
b. Representativeness. When processing threshold of the instrument used to collect models, calm hours should be treated as
MMIF data for use with AERMOD, the grid the wind speed data. The reference wind missing, and concentrations should be
cell used for the dispersion modeling should speed is selected by the model as the lowest calculated as in paragraph (a) of this
be adequately spatially representative of the level of non-missing wind speed and subsection.
9.0 Regulatory Application of Models the CAA, the NAAQS are subjected to required to demonstrate that emissions from
extensive review every 5 years and the the construction or operation of a proposed
9.1 Discussion standards, including the level and the form, new source or modification will not cause or
a. Standardized procedures are valuable in may be revised as part of that review. The contribute to a violation of the NAAQS or
the review of air quality modeling and data criteria pollutants have either long-term PSD increments.
analyses conducted to support SIP submittals (annual or quarterly) and/or short-term (24- i. For a NAAQS assessment, the design
and revisions, NSR, including PSD, or other hour or less) forms that are not to be concentration is the combination of the
EPA requirements to ensure consistency in exceeded more than a certain frequency over appropriate background concentration
their regulatory application. This section a period of time (e.g., no exceedance on a (section 8.3) with the estimated modeled
recommends procedures specific to NSR, rolling 3-month average, no more than once impact of the source. The NAAQS design
including PSD, that facilitate some degree of per year, or no more than once per year concentration is then compared to the
standardization while at the same time averaged over 3 years), are averaged over a applicable NAAQS.
allowing the flexibility needed to assure the period of time (e.g., an annual mean or an ii. For a PSD increment assessment, the
technically best analysis for each regulatory annual mean averaged over 3 years), or are design concentration includes impacts after
application. For SIP attainment some percentile that is averaged over a the appropriate baseline date from all
demonstrations, refer to the appropriate EPA period of time (e.g., annual 99th or 98th increment consuming and increment
guidance 51 60 for the recommended percentile averaged over 3 years). The 3-year expanding sources. The PSD increment
procedures. period for ambient monitoring design values design concentration is then compared to the
b. Air quality model estimates, especially does not dictate the length of the data periods applicable PSD increment.
with the support of measured air quality recommended for modeling (i.e., 5 years of b. The specific form of the NAAQS for the
data, are the preferred basis for air quality NWS meteorological data, at least 1 year of pollutant(s) of concern will also influence
demonstrations. A number of actions have site-specific, or at least 3 years of prognostic how the background and modeled data
been taken to ensure that the best air quality meteorological data). should be combined for appropriate
model is used correctly for each regulatory e. This section discusses general comparison with the respective NAAQS in
application and that it is not arbitrarily recommendations on the regulatory such a modeling demonstration. Given the
imposed. application of models for the purposes of potential for revision of the form of the
• First, the Guideline clearly recommends NSR, including PSD permitting, and NAAQS and the complexities of combining
that the most appropriate model be used in particularly for estimating design background and modeled data, specific
each case. Preferred models are identified, concentration(s), appropriately comparing details on this process can be found in
based on a number of factors, for many uses. these estimates to NAAQS and PSD applicable modeling guidance available on
• Second, the preferred models have been increment, and developing emissions limits. the EPA’s SCRAM Web site (section 2.3).
subjected to a systematic performance Lastly, this section provides the criteria Modeled concentrations should not be
evaluation and a peer scientific review. necessary for considering use of analysis rounded before comparing the resulting
Statistical performance measures, including based on measured ambient data in lieu of design concentration to the NAAQS or PSD
measures of difference (or residuals) such as modeling as the sole basis for demonstrating increments. Ambient monitoring and
bias, variance of difference and gross compliance with NAAQS and PSD dispersion modeling address different issues
variability of the difference, and measures of increments. and needs relative to each aspect of the
correlation such as time, space, and time and overall air quality assessment.
space combined as described in section 2.1.1, 9.2 Recommendations c. The PSD increments for criteria
were generally followed. 9.2.1 Modeling Protocol pollutants are listed in 40 CFR 52.21(c) and
• Third, more specific information has 40 CFR 51.166(c). For short-term increments,
a. Every effort should be made by the
been provided for considering the these maximum allowable increases in
appropriate reviewing authority (paragraph
incorporation of new models into the pollutant concentrations may be exceeded
3.0(b)) to meet with all parties involved in
Guideline (section 3.1) and the Guideline once per year at each site, while the annual
either a SIP submission or revision or a PSD
contains procedures for justifying the case- increment may not be exceeded. The highest,
permit application prior to the start of any
by-case use of alternative models and second-highest increase in estimated
work on such a project. During this meeting,
obtaining EPA approval (section 3.2). concentrations for the short-term averages as
a protocol should be established between the
The Guideline, therefore, provides preparing and reviewing parties to define the determined by a model should be less than
objective methods that allow a determination procedures to be followed, the data to be or equal to the permitted increment. The
to be made as to what air quality model or collected, the model to be used, and the modeled annual averages should not exceed
technique is most appropriate for a particular analysis of the source and concentration data the increment.
application. to be performed. An example of the content d. Receptor sites for refined dispersion
c. Air quality modeling is the preferred for such an effort is contained in the Air modeling should be located within the
basis for air quality demonstrations. Quality Analysis Checklist posted on the modeling domain (section 8.1). In designing
Nevertheless, there are rare circumstances EPA’s SCRAM Web site (section 2.3). This a receptor network, the emphasis should be
where the performance of the preferred air checklist suggests the appropriate level of placed on receptor density and location, not
quality model may be shown to be less than detail to assess the air quality resulting from total number of receptors. Typically, the
reasonably acceptable or where no preferred the proposed action. Special cases may density of receptor sites should be
air quality model, screening model or require additional data collection or analysis progressively more resolved near the new or
technique, or alternative model are suitable and this should be determined and agreed modifying source, areas of interest, and areas
for the situation. In these unique instances, upon at this pre-application meeting. The with the highest concentrations with
there is the possibility of assuring protocol should be written and agreed upon sufficient detail to determine where possible
compliance and establishing emissions limits by the parties concerned, although it is not violations of a NAAQS or PSD increment are
for an existing source solely on the basis of intended that this protocol be a binding, most likely to occur. The placement of
observed air quality data in lieu of an air formal legal document. Changes in such a receptor sites should be determined on a
quality modeling analysis. Comprehensive protocol or deviations from the protocol are case-by-case basis, taking into consideration
air quality monitoring in the vicinity of the often necessary as the data collection and the source characteristics, topography,
existing source with proposed modifications analysis progresses. However, the protocol climatology, and monitor sites. Locations of
will be necessary in these cases. The same establishes a common understanding of how particular importance include: (1) The area of
attention should be given to the detailed the demonstration required to meet maximum impact of the point source; (2) the
analyses of the air quality data as would be regulatory requirements will be made. area of maximum impact of nearby sources;
applied to a model performance evaluation. and (3) the area where all sources combine
d. The current levels and forms of the 9.2.2 Design Concentration and Receptor to cause maximum impact. Depending on the
NAAQS for the six criteria pollutants can be Sites complexities of the source and the
found on the EPA’s NAAQS Web site at a. Under the PSD permitting program, an environment to which the source is located,
http://www.epa.gov/air/criteria.html. Under air quality analysis for criteria pollutants is a dense array of receptors may be required in
some cases. In order to avoid unreasonably a significant impact, then the analysis is alternative model is available, model results,
large computer runs due to an excessively generally sufficient to demonstrate that the including the appropriate background, are
large array of receptors, it is often desirable source will not cause or contribute to an sufficient for air quality demonstrations and
to model the area twice. The first model run exceedance. However, if the concentration establishing emissions limits, if necessary. In
would use a moderate number of receptors estimates from the refined modeling analysis instances when the modeling technique
more resolved nearby the new or modifying indicate that significant impacts may occur, available is only a screening technique, the
source and over areas of interest. The second then a cumulative impact analysis should be addition of air quality monitoring data to the
model run would modify the receptor undertaken. The receptors that indicate the analysis may lend credence to the model
network from the first model run with a location of significant impacts should be results. However, air quality monitoring data
denser array of receptors in areas showing used to define the modeling domain for use alone will normally not be acceptable as the
potential for high concentrations and in the cumulative impact analysis (section sole basis for demonstrating compliance with
possible violations, as indicated by the 8.2.2). the NAAQS and PSD increments or for
results of the first model run. Accordingly, d. The cumulative impact analysis, or the determining emissions limits.
the EPA neither anticipates nor encourages second stage of an air quality analysis, b. There may be rare circumstances where
that numerous iterations of modeling runs be should be conducted with the same refined the performance of the preferred air quality
made to continually refine the receptor model or technique to characterize the model will be shown to be less than
network. project source and then include the reasonably acceptable when compared with
appropriate background concentrations air quality monitoring data measured in the
9.2.3 NAAQS and PSD Increments
(section 8.3). The resulting design vicinity of an existing source. Additionally,
Compliance Demonstrations for New or
concentrations are used to determine there may not be an applicable preferred air
Modified Sources quality model, screening technique, or
whether the source will cause or contribute
a. As described in this subsection, the to a NAAQS or PSD increment violation. justifiable alternative model suitable for the
recommended procedure for conducting This determination should be based on: (1) situation. In these unique instances, there
either a NAAQS or PSD increment The appropriate design concentration for may be the possibility of establishing
assessment under PSD permitting is a multi- each applicable NAAQS (and averaging emissions limits and demonstrating
stage approach that includes the following period); and (2) the significance of the compliance with the NAAQS and PSD
two stages: source’s contribution, in a temporal and increments solely on the basis of analysis of
i. The first stage is referred to as a single- spatial sense, to any modeled violation, i.e., observed air quality data in lieu of an air
source impact analysis, since only the new or where and when the predicted design quality modeling analysis. However, only in
modifying source is considered in the concentration is greater than the NAAQS. For the case of a modification to an existing
analysis. There are two possible levels of PSD increment, the cumulative impact source should air quality monitoring data
detail in conducting a single-source impact analysis should also consider the amount of alone be a basis for determining adequate
analysis with the model user beginning with the air quality increment that has already emissions limits or for demonstration that the
use of a screening model and proceeding to been consumed by other sources, or, modification will not cause or contribute to
use of a refined model as necessary. conversely, whether increment has expanded a violation of any NAAQS or PSD increment.
ii. The second stage is referred to as a relative to the baseline concentration. c. The following items should be
cumulative impact analysis, since it takes Therefore, the applicant should model the considered prior to the acceptance of an
into account all sources affecting the air existing or permitted nearby increment- analysis of measured air quality data as the
quality in an area. In addition to the project consuming and increment-expanding sole basis for an air quality demonstration or
source impact, it includes consideration of sources, rather than using past modeling determining an emissions limit:
background, which includes contributions analyses of those sources as part of i. Does a monitoring network exist for the
from natural, nearby, and unknown sources. background concentration. This would pollutants and averaging times of concern in
b. Each stage involves increasing permit the use of newly acquired data or the vicinity of the existing source?
complexity and details, as required to fully improved modeling techniques if such data ii. Has the monitoring network been
demonstrate a new or modifying source will and/or techniques have become available designed to locate points of maximum
not cause of contribution to a violation of any since the last source was permitted. concentration?
NAAQS or PSD increment. As such, starting iii. Do the monitoring network and the data
with a single-source impact analysis may 9.2.3.1 Considerations in Developing reduction and storage procedures meet EPA
alleviate the need for a more time consuming Emissions Limits monitoring and quality assurance
and comprehensive cumulative modeling a. Emissions limits and resulting control requirements?
analysis. requirements should be established to iv. Do the dataset and the analysis allow
c. The single-source impact analysis, or provide for compliance with each applicable impact of the most important individual
first stage of an air quality analysis, begins by NAAQS (and averaging period) and PSD sources to be identified if more than one
determining the potential of a proposed new increment. It is possible that multiple source or emission point is involved?
or modifying source to cause or contribute to emissions limits will be required for a source v. Is at least one full year of valid ambient
a NAAQS or PSD increment violation. In to demonstrate compliance with several data available?
certain circumstances, a screening model or criteria pollutants (and averaging periods) vi. Can it be demonstrated through the
technique may be used instead of the and PSD increments. Case-by-case comparison of monitored data with model
preferred model because it will provide determinations must be made as to the results that available air quality models and
estimated worst-case ambient impacts from appropriate form of the limits, i.e., whether techniques are not applicable?
the proposed new or modifying source. If the emissions limits restrict the emission c. Comprehensive air quality monitoring in
these worst case ambient concentration factor (e.g., limiting lb/MMBTU), the the area affected by the existing source with
estimates indicate that there will not be a emission rate (e.g., lb/hr), or both. The proposed modifications will be necessary in
significant impact, then the analysis is appropriate reviewing authority (paragraph these cases. Additional meteorological
sufficient for the required demonstration 3.0(b)) and appropriate EPA guidance should monitoring may also be necessary. The
under PSD. If the ambient concentration be consulted to determine the appropriate appropriate number of air quality and
estimates indicate that significant impacts emissions limits on a case-by-case basis. meteorological monitors from a scientific and
may occur, then the use of a refined model technical standpoint is a function of the
9.2.4 Use of Measured Data in Lieu of situation being considered. The source
to estimate the source’s impact should be
pursued. The refined modeling analysis Model Estimates configuration, terrain configuration, and
should use a model or technique consistent a. As described throughout the Guideline, meteorological variations all have an impact
with the Guideline (either a preferred model modeling is the preferred method for on number and optimal placement of
or technique or an alternative model or demonstrating compliance with the NAAQS monitors. Decisions on the monitoring
technique) and follow the requirements and and PSD increments and for determining the network appropriate for this type of analysis
recommendations for model inputs outlined most appropriate emissions limits for new can only be made on a case-by-case basis.
in section 8. If the estimated ambient and existing sources. When a preferred d. Sources should obtain approval from the
concentrations indicate that there will not be model or adequately justified and approved appropriate reviewing authority (paragraph
3.0(b)) and the EPA Regional Office for the Corp., Novato, CA. Prepared under 21. ‘‘Ten years of Harmonisation activities:
monitoring network prior to the start of contract No. EP–D–07–102 for U.S. Past, present and future’’ at http://
monitoring. A monitoring protocol agreed to Environmental Protection Agency, www.dmu.dk/AtmosphericEnvironment/
by all parties involved is necessary to assure Research Triangle Park, NC. http:// Harmoni/Conferences/Belgirate/
that ambient data are collected in a www.epa.gov/ttn/scram/reports/Plume_ BelgiratePapers.asp.
consistent and appropriate manner. The Eval_Final_Sep_2012v5.pdf. 22. Weil, Sykes, and Venkatram, 1992.
design of the network, the number, type, and 9. McMurry, P.H., Shepherd, M.F., Vickery, Evaluating Air-Quality Models: Review
location of the monitors, the sampling J.S., 2004. Particulate matter science for and Outlook. Journal of Applied
period, averaging time as well as the need for policy makers: A NARSTO assessment. Meteorology, 31: 1121–1145.
meteorological monitoring or the use of Cambridge University Press. 23. Environmental Protection Agency, 1988.
mobile sampling or plume tracking 10. Baker, K.R., Foley, K.M., 2011. A Model Clearinghouse: Operational Plan
techniques, should all be specified in the nonlinear regression model estimating (Revised). Staff Report. Office of Air
protocol and agreed upon prior to start-up of single source concentrations of primary Quality Planning & Standards, Research
the network. and secondarily formed PM2.5. Triangle Park, NC. (Docket No. A–88–04,
e. Given the uniqueness and complexities Atmospheric Environment, 45: 3758– II–J–1)
of these rare circumstances, the procedures 3767. 24. American Meteorological Society, 1983.
can only be established on a case-by-case 11. Bergin, M.S., Russell, A.G., Odman, M.T., Synthesis of the Rural Model Reviews.
basis for analyzing the source’s emissions Cohan, D.S., Chameldes, W.L., 2008. Publication No. EPA–600/3–83–108.
data and the measured air quality monitoring Single-Source Impact Analysis Using Office of Research & Development,
data and for projecting with a reasoned basis Three-Dimensional Air Quality Models. Research Triangle Park, NC. (NTIS No.
the air quality impact of a proposed Journal of the Air & Waste Management PB 84–121037)
modification to an existing source in order to Association, 58: 1351–1359. 25. Hanna, S., M. Garrison and B. Turner,
demonstrate that emissions from the 12. Zhou, W., Cohan, D.S., Pinder, R.W., 1998. AERMOD Peer Review report.
construction or operation of the modification Neuman, J.A., Holloway, J.S., Peischl, J., Prepared by SAI, Inc. under EPA
will not cause or contribute to a violation of Ryerson, T.B., Nowak, J.B., Flocke, F., Contract No. 68–D6–0064/1–14 for
the applicable NAAQS and PSD increment, Zheng, W.G., 2012. Observation and Environmental Protection Agency,
modeling of the evolution of Texas Research Triangle Park, NC. 12pp. &
and to determine adequate emissions limits.
power plant plumes. Atmospheric appendices. (Docket No. A–99–05, II–A–
The same attention should be given to the
Chemistry and Physics, 12: 455–468. 6)
detailed analyses of the air quality data as
13. Chen, J., Lu, J., Avise, J.C., DaMassa, J.A., 26. Scire, J.S. and L.L. Schulman, 1981.
would be applied to a comprehensive model
Kleeman, M.J., Kaduwela, A.P., 2014. Evaluation of the BLP and ISC Models
performance evaluation. In some cases, the with SF6 Tracer Data and SO2
monitoring data collected for use in the Seasonal modeling of PM 2.5 in
Measurements at Aluminum Reduction
performance evaluation of preferred air California’s San Joaquin Valley.
Plants. APCA Specialty Conference on
quality models, screening technique, or Atmospheric Environment, 92: 182–190.
Dispersion Modeling for Complex
existing alternative models may help inform 14. Russell, A.G., 2008. EPA Supersites
Sources, St. Louis, MO.
the development of a suitable new alternative program-related emissions-based
27. Environmental Protection Agency, 2003.
model. Early coordination with the particulate matter modeling: initial AERMOD: Latest Features and
appropriate reviewing authority (paragraph applications and advances. Journal of the Evaluation Results. Publication No.
3.0(b)) and the EPA Regional Office is Air & Waste Management Association, EPA–454/R–03–003. Office of Air
fundamental with respect to any potential 58: 289–302. Quality Planning & Standards, Research
use of measured data in lieu of model 15. Tesche, T., Morris, R., Tonnesen, G., Triangle Park, NC. http://www.epa.gov/
estimates. McNally, D., Boylan, J., Brewer, P., 2006. ttn/scram/7thconf/aermod/aermod_
CMAQ/CAMx annual 2002 performance mep.pdf.
10.0 References evaluation over the eastern US. 28. ASTM D6589: Standard Guide for
1. Code of Federal Regulations; Title 40 Atmospheric Environment, 40: 4906– Statistical Evaluation of Atmospheric
(Protection of Environment); Part 51; 4919. Dispersion Model Performance. (2010).
Sections 51.112, 51.117, 51.150, 51.160. 16. Fox, D.G., 1984. Uncertainty in air quality 29. Environmental Protection Agency, 1992.
2. Environmental Protection Agency, 1990. modeling. Bulletin of the American Protocol for Determining the Best
New Source Review Workshop Manual: Meteorological Society, 65(1): 27–36. Performing Model. Publication No. EPA–
Prevention of Significant Deterioration 17. Bowne, NE., 1981. Validation and 454/R–92–025. Office of Air Quality
and Nonattainment Area Permitting Performance Criteria for Air Quality Planning & Standards, Research Triangle
(Draft). Office of Air Quality Planning & Models. Appendix F in Air Quality Park, NC. http://www.epa.gov/ttn/scram/
Standards, Research Triangle Park, NC. Modeling and the Clean Air Act: guidance/guide/modleval.zip. (NTIS No.
http://www.epa.gov/nsr. Recommendations to EPA on Dispersion PB 93–226082)
3. Code of Federal Regulations; Title 40 Modeling for Regulatory Applications. 30. Hanna, S.R., 1982. Natural Variability of
(Protection of Environment); Part 51; American Meteorological Society, Observed Hourly SO2 and CO
Sections 51.166 and 52.21. Boston, MA; pp. 159–171. (Docket No. Concentrations in St. Louis. Atmospheric
4. Code of Federal Regulations; Title 40 A–80–46, II–A–106). Environment, 16(6): 1435–1440.
(Protection of Environment); Part 93; 18. Fox, D.G., 1981. Judging Air Quality 31. Pasquill, F., 1974. Atmospheric Diffusion,
Sections 93.116, 93.123, and 93.150. Model Performance. Bulletin of the 2nd Edition. John Wiley and Sons, New
5. Code of Federal Regulations; Title 40 American Meteorological Society, 62(5): York, NY; 479pp.
(Protection of Environment); Part 58 599–609. 32. Rhoads, R.G., 1981. Accuracy of Air
(Ambient Air Quality Surveillance). 19. Simon, H., Baker, K.R., Phillips, S., 2012. Quality Models. Staff Report. Office of
6. Code of Federal Regulations; Title 40 Compilation and interpretation of Air Quality Planning & Standards,
(Protection of Environment); Part 50 photochemical model performance Research Triangle Park, NC. (Docket No.
(National Primary and Secondary statistics published between 2006 and A–80–46, II–G–6)
Ambient Air Quality Standards). 2012. Atmospheric Environment, 61: 33. Hanna, S.R., 1993. Uncertainties in air
7. Baker, K.R., Kelly, J.T., 2014. Single source 124–139. quality model predictions. Boundary-
impacts estimated with photochemical 20. Burton, C.S., 1981. The Role of Layer Meteorology, 62: 3–20.
model source sensitivity and Atmospheric Models in Regulatory 34. Hanna, S.R., 1989. Confidence limits for
apportionment approaches. Atmospheric Decision-Making: Summary Report. air quality model evaluations, as
Environment, 96: 266–274. Systems Applications, Inc., San Rafael, estimated by bootstrap and jackknife
8. ENVIRON, 2012. Evaluation of Chemical CA. Prepared under contract No. 68–01– resampling methods. Atmospheric
Dispersion Models using Atmospheric 5845 for U.S. Environmental Protection Environment, 23(6): 1385–1398.
Plume Measurements from Field Agency, Research Triangle Park, NC. 35. Cox, W.M. and J.A. Tikvart, 1990. A
Experiments. ENVIRON International, (Docket No. A–80–46, II–M–6). statistical procedure for determining the
best performing air quality simulation Applications. Part I: General Model Publication No. EPA–454–R–04–002.
model. Atmospheric Environment, Formulation and Boundary Layer Office of Air Quality Planning &
24A(9): 2387–2395. Characterization. Journal of Applied Standards, Research Triangle Park, NC.
36. Environmental Protection Agency, 2015. Meteorology, 44(5): 682–693. http://www.epa.gov/airtrends/aqtrnd04/
Technical Support Document (TSD) for 46. L.L. Schulman, D.G. Strimaitis and J.S. pmreport03/report_2405.pdf#page=1.
AERMOD-Based Assessments of Long- Scire, 2002. Development and evaluation 59. Environmental Protection Agency, 2014.
Range Transport Impacts for Primary of the PRIME plume rise and building Guidance for PM2.5 Modeling. May 20,
Pollutants. Publication No. EPA–454/B– downwash model. Journal of the Air & 2014, Publication No. EPA–454/B–14–
15–003. Office of Air Quality Planning & Waste Management Association, 50: 001. Office of Air Quality Planning &
Standards, Research Triangle Park, NC. 378–390. Standards, Research Triangle Park, NC.
37. Environmental Protection Agency, 2015: 47. Environmental Protection Agency, 2015, http://www.epa.gov/ttn/scram/guidance/
AERSCREEN User’s Guide. Publication Technical Support Document (TSD) for guide/Guidance_for_PM25_Permit_
No. EPA–454-/B–15–005. Office of Air Replacement of CALINE3 with AERMOD Modeling.pdf.
Quality Planning & Standards, Research for Transportation Related Air Quality 60. Environmental Protection Agency, 2014.
Triangle Park, NC. http://www.epa.gov/ Analyses. Publication No. EPA–454/B– Modeling Guidance for Demonstrating
ttn/scram/dispersion_ 15–002. Office of Air Quality Planning & Attainment of Air Quality Goals for
screening.htm#aerscreen. Standards, Research Triangle Park, NC. Ozone, PM2.5, and Regional Haze. Office
38. Environmental Protection Agency, 2011: 48. Environmental Protection Agency, 1997. of Air Quality Planning & Standards,
AERSCREEN Released as the EPA Guidance for Siting Ambient Air Research Triangle Park, NC. http://
Recommended Screening Model. Tyler Monitors around Stationary Lead www.epa.gov/ttn/scram/guidance/guide/
Fox Memorandum dated April 11, 2011, Sources. Publication No. EPA–454/R– Draft_O3–PM–RH_Modeling_Guidance-
Office of Air Quality Planning & 92–009R. Office of Air Quality Planning 2014.pdf.
Standards, Research Triangle Park, NC. & Standards, Research Triangle Park, NC. 61. Environmental Protection Agency, 2013.
http://www.epa.gov/ttn/scram/ (NTIS No. PB 97–208094) Transportation Conformity Guidance for
20110411_AERSCREEN_Release_ 49. LEADPOST processor: http:// Quantitative Hot-Spot Analyses in PM2.5
Memo.pdf. www.epa.gov/ttn/scram/models/aermod/ and PM10 Nonattainment and
39. Perry, S.G., D.J. Burns and A.J. Cimorelli, leadpost.zip. Maintenance Areas. Publication No.
1990. User’s Guide to CTDMPLUS: 50. Environmental Protection Agency, 1993. EPA–420–B–053. Office of
Volume 2. The Screening Mode Lead Guideline Document. Publication Transportation and Air Quality, Ann
(CTSCREEN). Publication No. EPA–600/ No. EPA–452/R–93–009. Office of Air Arbor, MI. http://www.epa.gov/otaq/
8–90–087. U.S. Environmental Quality Planning & Standards, Research stateresources/transconf/policy/
Protection Agency, Research Triangle Triangle Park, NC. (NTIS No. PB 94– 420b13053-sec.pdf.
Park, NC. http://www.epa.gov/ttn/scram/ 111846) 62. Environmental Protection Agency, 1987.
dispersion_prefrec.htm#ctdmplus. (NTIS 51. U.S. EPA, 2014: Guidance for 1-Hour SO2 PM10 SIP Development Guideline.
No. PB 91–136564) Nonattaiment Area SIP Submissions. Publication No. EPA–450/2–86–001.
40. Environmental Protection Agency, 1992. Stephen D. Page Memorandum dated Office of Air Quality Planning &
Screening Procedures for Estimating the April 23, 2011, Office of Air Quality Standards, Research Triangle Park, NC.
Air Quality Impact of Stationary Sources, Planning & Standards, Research Triangle (NTIS No. PB 87–206488)
Revised. Publication No. EPA–454/R– Park, NC. http://www.epa.gov/airquality/ 63. Environmental Protection Agency, 2012.
92–019. Office of Air Quality Planning & sulfurdioxide/pdfs/ Memorandum: Haul Road Workgroup
Standards, Research Triangle Park, NC. 20140423guidance.pdf. Final Report Submission to EPA–
(NTIS No. PB 93–219095) 52. U.S. EPA, 2013: SO2 NAAQS OAQPS. March 2, 2012. Office of Air
41. Burns, D.J., S.G. Perry and A.J. Cimorelli, Designations Modeling Technical Quality Planning & Standards, Research
1991. An Advanced Screening Model for Assistance Document. Office of Air Triangle Park, NC. http://www.epa.gov/
Complex Terrain Applications. Paper Quality Planning & Standards, Research ttn/scram/reports/Haul_Road_
presented at the 7th Joint Conference on Triangle Park, NC. http://www.epa.gov/ Workgroup-Final_Report_Package-
Applications of Air Pollution airquality/sulfurdioxide/pdfs/ 20120302.pdf.
Meteorology (cosponsored by the SO2ModelingTAD.pdf. 64. Seinfeld, J.H., Pandis, S.N., 2012.
American Meteorological Society and the 53. Turner, D.B., 1964. A Diffusion Model for Atmospheric chemistry and physics:
Air & Waste Management Association), an Urban Area. Journal of Applied from air pollution to climate change.
January 13–18, 1991, New Orleans, LA. Meteorology, 3(1):83–91. John Wiley & Sons.
42. Mills, M.T., R.J. Paine, E.A. Insley and 54. Environmental Protection Agency, 2015. 65. Simon, H., Baker, K.R., Phillips, S., 2012.
B.A. Egan, 1987. The Complex Terrain Technical Support Document (TSD) for Compilation and interpretation of
Dispersion Model Terrain Preprocessor NO2-Related AERMOD Options and photochemical model performance
System—User’s Guide and Program Modifications, Publication No. EPA–454/ statistics published between 2006 and
Description. Publication No. EPA–600/ B–15–004. Office of Air Quality Planning 2012. Atmospheric Environment, 61,
8–88–003. U.S. Environmental & Standards, Research Triangle Park, NC. 124–139.
Protection Agency, Research Triangle 55. Podrez, M. 2015. An Update to the 66. Environmental Protection Agency, 2015.
Park, NC. (NTIS No. PB 88–162094) Ambient Ratio Method for 1-h NO2 Air Guidance on the use of models for
43. Environmental Research and Technology, Quality Standards Dispersion Modeling. assessing the impacts from single sources
1987. User’s Guide to the Rough Terrain Atmospheric Environment, 103: 163– on secondarily formed pollutants ozone
Diffusion Model (RTDM), Rev. 3.20. ERT 170. and PM2.5. Publication No. EPA 454/P–
Document No. P–D535–585. 56. Cole, H.S. and J.E. Summerhays, 1979. A 15–001. Office of Air Quality Planning &
Environmental Research and Review of Techniques Available for Standards, Research Triangle Park, NC.
Technology, Inc., Concord, MA. (NTIS Estimation of Short-Term NO2 67. U.S. Department of the Interior, 2010.
No. PB 88–171467) Concentrations. Journal of the Air Federal Land Managers’ Air Quality
44. Environmental Protection Agency, 2004. Pollution Control Association, 29(8): Related Values Work Group (FLAG)
AERMOD: Description of Model 812–817. Phase I Report—Revised 2010. http://
Formulation. Publication No. EPA–454/ 57. Hanrahan, P.L., 1999. The Polar Volume www.nature.nps.gov/air/pubs/pdf/flag/
R–03–004. Office of Air Quality Planning Polar Ratio Method for Determining FLAG_2010.pdf. Natural Resource Report
& Standards, Research Triangle Park, NC; NO2/NOX Ratios in Modeling—Part I: NPS/NPRC/NRR–2010/232.
September 2004. http://www.epa.gov/ Methodology. Journal of the Air & Waste 68. National Acid Precipitation Assessment
ttn/scram/7thconf/aermod/aermod_ Management Association, 49: 1324– Program (NAPAP), 1991. Acid
mfd.pdf. 1331. Deposition: State of Science and
45. Cimorelli, A. et al., 2005. AERMOD: A 58. Environmental Protection Agency, 2004. Technology. Volume III Terrestrial,
Dispersion Model for Industrial Source The Particle Pollution Report, Materials, Health and Visibility Effects.
Report 24, Visibility: Existing and 81. Environmental Protection Agency, 1981. eidocs/eiguid/
Historical Conditions—Causes and Guideline for Use of Fluid Modeling to 2014revisedeiguidance.pdf.
Effects. Edited by Patricia M. Irving. Determine Good Engineering Practice 91. Environmental Protection Agency, 1987.
Washington, DC 129pp. (GEP) Stack Height. Publication No. Ambient Air Monitoring Guidelines for
69. National Research Council, 1993. EPA–450/4–81–003. Office of Air Prevention of Significant Deterioration
Protecting Visibility in National Parks Quality Planning & Standards, Research (PSD). Publication No. EPA–450/4–87–
and Wilderness Areas. National Triangle Park, NC. (NTIS No. PB 82– 007. Office of Air Quality Planning &
Academy Press, Washington, DC 446pp. 145327) Standards, Research Triangle Park, NC.
70. U.S. Environmental Protection Agency, 82. Lawson, Jr., R.E. and W.H. Snyder, 1983. (NTIS No. PB 90–168030)
1992. Workbook for plume visual impact Determination of Good Engineering 92. Environmental Protection Agency, 2011.
screening and analysis (revised). Practice Stack Height: A Demonstration Additional Clarification Regarding
Publication No. EPA–454/R–92–023. Study for a Power Plant. Publication No. Application of Appendix W Modeling
Office of Air Quality Planning & EPA–600/3–83–024. Office of Research & Guidance for the 1-hour NO2 National
Standards, Research Triangle Park, NC Development, Research Triangle Park, Ambient Air Quality Standard. Office of
(NTIS No. PB 93–223592). NC. (NTIS No. PB 83–207407) Air Quality Planning & Standards,
71. Nilsson, J., Grennfelt, P., Ministerråd, N., 83. Environmental Protection Agency, 1985. Research Triangle Park, NC. http://
1988. Critical Loads for Sulphur and Guideline for Determination of Good www.epa.gov/ttn/scram/guidance/
Nitrogen: Report from a Workshop Held Engineering Practice Stack Height clarification/Additional_Clarifications_
at Skokloster, Sweden, 19–24 March, (Technical Support Document for the AppendixW_Hourly-NO2-NAAQS_
1988. Nordic Council of Ministers. Stack Height Regulations), Revised. FINAL_03-01-2011.pdf.
72. Turner, D.B., 1969. Workbook of Publication No. EPA–450/4–80–023R. 93. Environmental Protection Agency, 2014.
Atmospheric Dispersion Estimates. PHS Office of Air Quality Planning & Modeling Guidance For Demonstrating
Publication No. 999–AP–26. U.S. Standards, Research Triangle Park, NC. Attainment of Air Quality Goals for
Department of Health, Education and http://www.epa.gov/ttn/scram/guidance/ Ozone, PM2.5, and Regional Haze (Draft).
Welfare, Public Health Service, guide/gep.pdf. (NTIS No. PB 85–225241) U.S. Office of Air Quality Planning &
Cincinnati, OH. (NTIS No. PB–191482) 84. Snyder, W.H. and R.E. Lawson, Jr., 1985. Standards, Research Triangle Park, NC.
73. McElroy, J.L. and F. Pooler, Jr., 1968. St. Fluid Modeling Demonstration of Good http://www.epa.gov/ttn/scram/guidance/
Louis Dispersion Study, Volume II— Engineering-Practice Stack Height in guide/Draft_O3-PM-RH_Modeling_
Analysis. National Air Pollution Control Complex Terrain. Publication No. EPA– Guidance-2014.pdf.
Administration Publication No. AP–53, 600/3–85–022. Office of Research & 94. Environmental Protection Agency, 2013.
U.S. Department of Health, Education Development, Research Triangle Park, Memorandum: Use of ASOS
and Welfare, Public Health Service, NC. (NTIS No. PB 85–203107) meteorological data in AERMOD
Arlington, VA. (NTIS No. PB–190255) 85. Briggs, G.A., 1975. Plume Rise dispersion modeling. March 8, 2013.
74. Irwin, J.S., 1978. Proposed Criteria for Predictions. Chapter 3 in Lectures on Air Office of Air Quality Planning &
Selection of Urban Versus Rural Pollution and Environmental Impact Standards, Research Triangle Park, NC.
Dispersion Coefficients. (Draft Staff Analyses. American Meteorological http://www.epa.gov/ttn/scram/guidance/
Report). Meteorology and Assessment Society, Boston, MA; pp. 59–111. clarification/20130308_Met_Data_
Division, U.S. Environmental Protection 86. Hanna, S.R., G.A. Briggs and R.P. Hosker, Clarification.pdf.
Agency, Research Triangle Park, NC. Jr., 1982. Plume Rise. Chapter 2 in 95. Environmental Protection Agency, 2004.
(Docket No. A–80–46, II–B–8) Handbook on Atmospheric Diffusion. User’s Guide for the AERMOD
75. Auer, Jr., A.H., 1978. Correlation of Land Technical Information Center, U.S. Meteorological Preprocessor (AERMET).
Use and Cover with Meteorological Department of Energy, Washington, DC; Publication No. EPA–454/B–03–002.
Anomalies. Journal of Applied pp. 11–24. DOE/TIC–11223 (DE Office of Air Quality Planning &
Meteorology, 17(5): 636–643. 82002045) Standards, Research Triangle Park, NC.
76. Environmental Protection Agency, 2015. 87. Weil, J.C., L.A. Corio and R.P. Brower, http://www.epa.gov/ttn/scram/
AERMOD Implementation Guide. U.S. 1997. A PDF dispersion model for metobsdata_procaccprogs.htm#aermet.
Office of Air Quality Planning & buoyant plumes in the convective 96. U.S Environmental Protection Agency.
Standards, Research Triangle Park, NC. boundary layer. Journal of Applied AERMINUTE User’s Guide. Office of Air
http://www.epa.gov/ttn/scram/ Meteorology, 36: 982–1003. Quality Planning & Standards, Research
dispersion_prefrec.htm#aermod. 88. L.L. Schulman, D.G. Strimaitis and J.S. Triangle Park, NC. http://www.epa.gov/
77. Pasquill, F., 1976. Atmospheric Scire, 2002. Development and evaluation ttn/scram/metobsdata_
Dispersion Parameters in Gaussian of the PRIME plume rise and building procaccprogs.htm#aermet.
Plume Modeling, Part II. Possible downwash model. Journal of the Air & 97. Environmental Protection Agency, 1993.
Requirements for Change in the Turner Waste Management Association, 50: PCRAMMET User’s Guide. Publication
Workbook Values. Publication No. EPA– 378–390. No. EPA–454/R–96–001. Office of Air
600/4–76–030b. Office of Research & 89. Environmental Protection Agency, 1995. Quality Planning & Standards, Research
Development, Research Triangle Park, Compilation of Air Pollutant Emission Triangle Park, NC. (NTIS No. PB 97–
NC. (NTIS No. PB–258036/3BA) Factors, Volume I: Stationary Point and 147912)
78. Stull, R.B., 1988. An Introduction to Area Sources (Fifth Edition, AP–42: GPO 98. Environmental Protection Agency, 1996.
Boundary Layer Meteorology. Kluwer Stock No. 055–000–00500–1), and Meteorological Processor for Regulatory
Academic Publishers, Boston, MA. Supplements A- D; Volume II: Mobile Models (MPRM). Publication No. EPA–
666pp. Sources (Fifth Edition). Office of Air 454/R–96–002. Office of Air Quality
79. Environmental Protection Agency, 1987. Quality Planning & Standards, Research Planning & Standards, Research Triangle
Analysis and Evaluation of Statistical Triangle Park, NC. Volume I can be Park, NC. (NTIS No. PB 96–180518)
Coastal Fumigation Models. Publication downloaded from EPA’s Web site at 99. Paine, R.J., 1987. User’s Guide to the
No. EPA–450/4–87–002. Office of Air http://www.epa.gov/ttn/chief/ap42.html; CTDM Meteorological Preprocessor
Quality Planning & Standards, Research Volume II can be downloaded from Program. Publication No. EPA–600/8–
Triangle Park, NC. (NTIS No. PB 87– http://www.epa.gov/omswww/ap42.htm. 88–004. Office of Research &
175519) 90. Environmental Protection Agency, 2014. Development, Research Triangle Park,
80. Wesely, M.L, P.V. Doskey, and J.D. Draft Emissions Inventory Guidance for NC. (NTIS No. PB–88–162102)
Shannon, 2002: Deposition Implementation of Ozone and Particulate 100. Perry, S.G., D.J. Burns, L.H. Adams, R.J.
Parameterizations for the Industrial Matter National Ambient Air Quality Paine, M.G. Dennis, M.T. Mills, D.G.
Source Complex (ISC3) Model. Draft Standards (NAAQS) and Regional Haze Strimaitis, R.J. Yamartino and E.M.
ANL report ANL/ER/TRB01/003, DOE/ Regulations. Office of Air Quality Insley, 1989. User’s Guide to the
xxnnnn, Argonne National Laboratory, Planning & Standards, Research Triangle Complex Terrain Dispersion Model Plus
Argonne, Illinois 60439. Park, NC. http://www.epa.gov/ttn/chief/ Algorithms for Unstable Situations
(CTDMPLUS). Volume 1: Model 26 West Martin Luther King Dr., demonstration of applicability provided they
Descriptions and User Instructions. EPA Cincinnati, OH 45268. satisfy the recommendations for regulatory
Publication No. EPA–600/8–89–041. 112. Bowen, B.M., J.M. Dewart and A.I. Chen, use; not all options in the models are
Environmental Protection Agency, 1983. Stability Class Determination: A necessarily recommended for regulatory use.
Research Triangle Park, NC. http:// Comparison for One Site. Proceedings, (2) Many of these models have been
www.epa.gov/ttn/scram/dispersion_ Sixth Symposium on Turbulence and subjected to a performance evaluation using
prefrec.htm#ctdmplus. (NTIS No. PB 89– Diffusion. American Meteorological comparisons with observed air quality data.
181424) Society, Boston, MA; pp. 211–214. Where possible, several of the models
101. Environmental Protection Agency, 2008. (Docket No. A–92–65, II–A–7) contained herein have been subjected to
AERSURFACE User’s Guide. Publication 113. Environmental Protection Agency, 1993. evaluation exercises, including (1) statistical
No. EPA–454/B–08–001. Office of Air An Evaluation of a Solar Radiation/ performance tests recommended by the
Quality Planning & Standards, Research Delta-T (SRDT) Method for Estimating American Meteorological Society and (2)
Triangle Park, NC. http://www.epa.gov/ Pasquill-Gifford (P–G) Stability peer scientific reviews. The models in this
ttn/scram/dispersion_ Categories. Publication No. EPA–454/R– appendix have been selected on the basis of
related.htm#aersurface. 93–055. Office of Air Quality Planning & the results of the model evaluations,
102. Wesson, K., R. Brode, W. Peters, and C. Standards, Research Triangle Park, NC. experience with previous use, familiarity of
Tillerson, 2006. AERMOD Sensitivity to (NTIS No. PB 94–113958)
the model to various air quality programs,
the Choice of Surface Characteristics. 114. Irwin, J.S., 1980. Dispersion Estimate
and the costs and resource requirements for
A&WMA presentation. Suggestion #8: Estimation of Pasquill
use.
103. Environ, 2014. The Mesocale Model Stability Categories. Office of Air Quality
(3) Codes and documentation for all
Interface Program (MMIF) Version 3.1 Planning & Standards, Research Triangle
models listed in this appendix are available
User’s Manual. Park, NC. (Docket No. A–80–46, II–B–10)
115. Mitchell, Jr., A.E. and K.O. Timbre, from the EPA’s Support Center for Regulatory
104. Environmental Protection Agency, 2015
1979. Atmospheric Stability Class from Air Models (SCRAM) Web site at http://
Guidance on the Use of the Mesoscale
Horizontal Wind Fluctuation. Presented www.epa.gov/ttn/scram. Codes and
Model Interface Program (MMIF) for
at 72nd Annual Meeting of Air Pollution documentation may also available from the
AERMOD Applications. Publication No.
Control Association, Cincinnati, OH; National Technical Information Service
EPA–454/B–15–001. Office of Air
June 24–29, 1979. (Docket No. A–80–46, (NTIS), http://www.ntis.gov, and, when
Quality Planning & Standards, Research
II–P–9) available, is referenced with the appropriate
Triangle Park, NC.
116. Smedman—Hogstrom, A. and V. NTIS accession number.
105. Solar and Meteorological Surface
Observation Network, 1961–1990; 3- Hogstrom, 1978. A Practical Method for A.1 AERMOD (AMS/EPA Regulatory
volume CD–ROM. Version 1.0, Determining Wind Frequency Model)
September 1993. Produced jointly by Distributions for the Lowest 200m from
National Climatic Data Center and Routine Meteorological Data. Journal of References
National Renewable Energy Laboratory. Applied Meteorology, 17(7): 942–954. Environmental Protection Agency, 2004.
Can be ordered from NOAA National 117. Smith, T.B. and S.M. Howard, 1972. AERMOD: Description of Model
Data Center’s Web site at http:// Methodology for Treating Diffusivity. Formulation. Publication No. EPA–454/
www.nndc.noaa.gov. MRI 72 FR–1030. Meteorology Research, R–03–004. Office of Air Quality Planning
106. Hourly United States Weather Inc., Altadena, CA. (Docket No. A–80– & Standards, Research Triangle Park, NC;
Observations, 1990–1995 (CD–ROM). 46, II–P–8) September 2004. http://www.epa.gov/
October 1997. Produced jointly by 118. Environmental Protection Agency, 2015. ttn/scram/7thconf/aermod/aermod_
National Climatic Data Center and Evaluation of Prognostic Meteorological mfd.pdf.
Environmental Protection Agency. Can Data in AERMOD Applications. Cimorelli, A., et al., 2005. AERMOD: A
be ordered from NOAA National Data Publication No. EPA–454/R–15–004. Dispersion Model for Industrial Source
Center’s Web site at http:// Office of Air Quality Planning & Applications. Part I: General Model
www.ncdc.noaa.gov. Standards, Research Triangle Park, NC. Formulation and Boundary Layer
107. Radiosonde Data of North America, 119. Environmental Protection Agency, 1984. Characterization. Journal of Applied
1946–1996; 4-volume CD–ROM. Calms Processor (CALMPRO) User’s Meteorology, 44(5): 682–693.
August1996. Produced jointly by Guide. Publication No. EPA–901/9–84– Perry, S. et al., 2005. AERMOD: A Dispersion
Forecast Systems laboratory and 001. Office Of Air Quality Planning & Model for Industrial Source
Standards, Region I, Boston, MA. (NTIS Applications. Part II: Model Performance
National Climatic Data Center. Can be
No. PB 84–229467)
ordered from NOAA National Data against 17 Field Study Databases. Journal
Center’s Web site at http:// Appendix A to Appendix W of Part of Applied Meteorology, 44(5): 694–708.
lwf.ncdc.noaa.gov/oa/ncdc.html. 51—Summaries of Preferred Air Environmental Protection Agency, 2004.
108. Environmental Protection Agency, 2000. User’s Guide for the AMS/EPA
Quality Models
Meteorological Monitoring Guidance for Regulatory Model—AERMOD.
Regulatory Modeling Applications. Table of Contents Publication No. EPA–454/B–03–001.
Publication No. EPA–454/R–99–005. A.0 Introduction and Availability Office of Air Quality Planning &
Office of Air Quality Planning & A.1 AERMOD (AMS/EPA Regulatory Standards, Research Triangle Park, NC;
Standards, Research Triangle Park, NC. Model) September 2004. http://www.epa.gov/
(NTIS No. PB 2001–103606) A.2 CTDMPLUS (Complex Terrain ttn/scram/dispersion_
109. ASTM D5527: Standard Practice for Dispersion Model Plus Algorithms for prefrec.htm#aermod.
Measuring Surface Winds and Unstable Situations) Environmental Protection Agency, 2004.
Temperature by Acoustic Means. (2011) A.3 OCD (Offshore and Coastal Dispersion User’s Guide for the AERMOD
110. ASTM D5741: Standard Practice for Model) Meteorological Preprocessor (AERMET).
Characterizing Surface Wind Using Wind Publication No. EPA–454/B–03–002.
A.0 Introduction and Availability
Vane and Rotating Anemometer. (2011) Office of Air Quality Planning &
111. Environmental Protection Agency, 1995. (1) This appendix summarizes key features Standards, Research Triangle Park, NC;
Quality Assurance for Air Pollution of refined air quality models preferred for November 2004. http://www.epa.gov/ttn/
Measurement Systems, Volume IV— specific regulatory applications. For each scram/metobsdata_
Meteorological Measurements. model, information is provided on procaccprogs.htm#aermet.
Publication No. EPA600/R–94/038d. availability, approximate cost (where User’s Guide for the AERMOD Terrain
Office of Air Quality Planning & applicable), regulatory use, data input, Preprocessor (AERMAP). Publication No.
Standards, Research Triangle Park, NC. output format and options, simulation of EPA–454/B–03–003. Office of Air
Note: For copies of this handbook, you atmospheric physics, and accuracy. These Quality Planning & Standards, Research
may make inquiry to ORD Publications, models may be used without a formal Triangle Park, NC; October 2004.
http://www.epa.gov/ttn/scram/ regulatory default mode, pollutant half-life or applications with multi-level site-specific
dispersion_related.htm#aermap. decay options are not employed, except in meteorological data, the profile contains the
Schulman, L. L., D.G. Strimaitis and J.S. the case of an urban source of sulfur dioxide observations made at each level of the
Scire, 2000. Development and evaluation where a four-hour half-life is applied. Terrain meteorological tower (or remote sensor).
of the PRIME plume rise and building elevation data from the U.S. Geological When using prognostic data, the surface file
downwash model. Journal of the Air and Survey 7.5-Minute Digital Elevation Model contains surface variables calculated by the
Waste Management Association, 50: (DEM) or equivalent (approx. 30-meter prognostic model and AERMET. The profile
378–390. resolution) should be used in all file contains the observations made at each
Schulman, L. L., and Joseph S. Scire, 1980. applications. Starting in 2011, data from the level of a meteorological tower (or remote
Buoyant Line and Point Source (BLP) National Elevation Dataset (NED, http:// sensor), the one-level observations taken
Dispersion Model User’s Guide. ned.usgs.gov) can also be used in AERMOD, from other representative data (e.g., National
Document P–7304B. Environmental which includes a range of resolutions, Weather Service surface observations), one
Research and Technology, Inc., Concord, ranging from 1-m to 2 arc seconds and such record per level per hour, or in the case of
MA. (NTIS No. PB 81–164642). high resolution would always be preferred. In prognostic data, the prognostic modeled
some cases, exceptions of the terrain data values of temperature and winds at user-
Availability requirement may be made in consultation specified levels.
The model codes and associated with the appropriate reviewing authority (i) Data used as input to AERMET should
documentation are available on EPA’s (paragraph 3.0(b)). possess an adequate degree of
SCRAM Web site (paragraph A.0(3)). b. Input Requirements representativeness to insure that the wind,
Abstract (1) Source data: Required input includes temperature and turbulence profiles derived
source type, location, emission rate, stack by AERMOD are both laterally and vertically
AERMOD is a steady-state plume
height, stack inside diameter, stack gas exit representative of the source area. The
dispersion model for assessment of pollutant
velocity, stack gas temperature, area and adequacy of input data should be judged
concentrations from a variety of sources.
volume source dimensions, and source independently for each variable. The values
AERMOD simulates transport and dispersion
elevation. Building dimensions and variable for surface roughness, Bowen ratio, and
from multiple point, area, or volume sources
emission rates are optional. Buoyant line albedo should reflect the surface
based on an up-to-date characterization of the
atmospheric boundary layer. Sources may be sources require coordinates of the end points characteristics in the vicinity of the
located in rural or urban areas, and receptors of the line, release height, emission rate, meteorological tower or representative grid
may be located in simple or complex terrain. average line source width, average building cell when using prognostic data, and should
AERMOD accounts for building wake effects width, average spacing between buildings, be adequately representative of the modeling
(i.e., plume downwash) based on the PRIME and average line source buoyancy parameter. domain. Finally, the primary atmospheric
building downwash algorithms. The model For mobile sources, traffic volume; emission input variables including wind speed and
employs hourly sequential preprocessed factor, source height, and mixing zone width direction, ambient temperature, cloud cover,
meteorological data to estimate are needed. and a morning upper air sounding should
concentrations for averaging times from 1- (2) Meteorological data: The AERMET also be adequately representative of the
hour to 1-year (also multiple years). meteorological preprocessor requires input of source area, when using observed data.
AERMOD can be used to estimate the surface characteristics, including surface (ii) For recommendations regarding the
concentrations of nonreactive pollutants from roughness (zo), Bowen ratio, and albedo, as length of meteorological record needed to
highway traffic. AERMOD also handles well as, hourly observations of wind speed perform a regulatory analysis with AERMOD,
unique modeling problems associated with between 7zo and 100m (reference wind speed see section 8.4.2.
aluminum reduction plants, and other measurement from which a vertical profile (3) Receptor data: Receptor coordinates,
industrial sources where plume rise and can be developed), wind direction, cloud elevations, height above ground, and hill
downwash effects from stationary buoyant cover, and temperature between zo and 100m height scales are produced by the AERMAP
line sources are important. AERMOD is (reference temperature measurement from terrain preprocessor for input to AERMOD.
designed to operate in concert with two pre- which a vertical profile can be developed). Discrete receptors and/or multiple receptor
processor codes: AERMET processes Meteorological data can be in the form of grids, Cartesian and/or polar, may be
meteorological data for input to AERMOD, observed data or prognostic modeled data as employed in AERMOD. AERMAP requires
and AERMAP processes terrain elevation discussed in paragraph 8.4.1(d). Surface input of DEM terrain data produced by the
data and generates receptor and hill height characteristics may be varied by wind sector U.S. Geological Survey (USGS), or other
information for input to AERMOD. and by season or month. When using equivalent data. AERMAP can be used
observed meteorological data, a morning optionally to estimate source elevations.
a. Recommendations for Regulatory Use
sounding (in National Weather Service c. Output
(1) AERMOD is appropriate for the format) from a representative upper air
following applications: station is required. Latitude, longitude, and Printed output options include input
• Point, volume, and area sources; time zone of the surface, site-specific (if information, high concentration summary
• Buoyant, elevated line sources (e.g., applicable) and upper air meteorological tables by receptor for user-specified
aluminum reduction plants); stations are required. The wind speed averaging periods, maximum concentration
• Mobile (line) sources; starting threshold is also required in summary tables, and concurrent values
• Surface, near-surface, and elevated AERMET for applications involving site- summarized by receptor for each day
releases; specific data). When using prognostic data, processed. Optional output files can be
• Rural or urban areas; modeled profiles of temperature and winds generated for: A listing of occurrences of
• Simple and complex terrain; are input into AERMET. These can be hourly exceedances of user-specified threshold
• Transport distances over which steady- or a time that represents a morning sounding. value; a listing of concurrent (raw) results at
state assumptions are appropriate, up to Additionally, measured profiles of wind, each receptor for each hour modeled, suitable
50km; temperature, vertical and lateral turbulence for post-processing; a listing of design values
• 1-hour to annual averaging times; and may be required in certain applications (e.g., that can be imported into graphics software
• Continuous toxic air emissions.
in complex terrain) to adequately represent for plotting contours; a listing of results

(2) For regulatory applications of the meteorology affecting plume transport suitable for NAAQS analyses including
AERMOD, the regulatory default option and dispersion. Optionally, measurements of NAAQS exceedances and culpability
should be set, i.e., the parameter DFAULT solar, or net radiation may be input to analyses; an unformatted listing of raw
should be employed in the MODELOPT AERMET. Two files are produced by the results above a threshold value with a special
record in the Control Pathway. The DFAULT AERMET meteorological preprocessor for structure for use with the TOXX model
option requires the use of terrain elevation input to the AERMOD dispersion model. component of TOXST; a listing of
data, stack-tip downwash, sequential date When using observed data, the surface file concentrations by rank (e.g., for use in
checking, and does not permit the use of the contains observed and calculated surface quantile-quantile plots); and, a listing of
model in the SCREEN mode. In the variables, one record per hour. For concentrations, including arc-maximum
normalized concentrations, suitable for modeling domain, straight-line plume n. Evaluation Studies
model evaluation studies. trajectories are assumed, with adjustment in American Petroleum Institute, 1998.
d. Type of Model the plume/receptor geometry used to account Evaluation of State of the Science of Air
for the terrain effects. Quality Dispersion Model, Scientific
AERMOD is a steady-state plume model,
using Gaussian distributions in the vertical h. Horizontal Winds Evaluation, prepared by Woodward-
and horizontal for stable conditions, and in Vertical profiles of wind are calculated for Clyde Consultants, Lexington,
the horizontal for convective conditions. The each hour based on measurements and Massachusetts, for American Petroleum
vertical concentration distribution for surface-layer similarity (scaling) Institute, Washington, DC, 20005–4070.
convective conditions results from an relationships. At a given height above Brode, R.W., 2002. Implementation and
assumed bi-Gaussian probability density ground, for a given hour, winds are assumed Evaluation of PRIME in AERMOD.
function of the vertical velocity. constant over the modeling domain. The Preprints of the 12th Joint Conference on
effect of the vertical variation in horizontal Applications of Air Pollution
e. Pollutant Types Meteorology, May 20–24, 2002;
wind speed on dispersion is accounted for
AERMOD is applicable to primary through simple averaging over the plume American Meteorological Society,
pollutants and continuous releases of toxic depth. Boston, MA.
and hazardous waste pollutants. Chemical Brode, R.W., 2004. Implementation and
i. Vertical Wind Speed
transformation is treated by simple Evaluation of Bulk Richardson Number
exponential decay. In convective conditions, the effects of Scheme in AERMOD. 13th Joint
random vertical updraft and downdraft Conference on Applications of Air
f. Source-Receptor Relationships
velocities are simulated with a bi-Gaussian Pollution Meteorology, August 23–26,
AERMOD applies user-specified locations probability density function. In both 2004; American Meteorological Society,
for sources and receptors. Actual separation convective and stable conditions, the mean Boston, MA.
between each source-receptor pair is used. vertical wind speed is assumed equal to zero. Environmental Protection Agency, 2003.
Source and receptor elevations are user input AERMOD: Latest Features and
j. Horizontal Dispersion
or are determined by AERMAP using USGS Evaluation Results. Publication No.
DEM terrain data. Receptors may be located Gaussian horizontal dispersion coefficients
are estimated as continuous functions of the EPA–454/R–03–003. Office of Air
at user-specified heights above ground level. Quality Planning & Standards, Research
parameterized (or measured) ambient lateral
g. Plume Behavior turbulence and also account for buoyancy- Triangle Park, NC. http://www.epa.gov/
(1) In the convective boundary layer (CBL), induced and building wake-induced ttn/scram/7thconf/aermod/aermod_
the transport and dispersion of a plume is turbulence. Vertical profiles of lateral mep.pdf.
characterized as the superposition of three turbulence are developed from measurements Heist, D., et al, 2013. Estimating near-road
modeled plumes: The direct plume (from the and similarity (scaling) relationships. pollutant dispersion: A model inter-
stack), the indirect plume, and the penetrated Effective turbulence values are determined comparison. Transportation Research
plume, where the indirect plume accounts from the portion of the vertical profile of Part D: Transport and Environment, 25:
for the lofting of a buoyant plume near the lateral turbulence between the plume height pp 93–105.
top of the boundary layer, and the penetrated and the receptor height. The effective lateral A.2 CTDMPLUS (Complex Terrain
plume accounts for the portion of a plume turbulence is then used to estimate Dispersion Model Plus Algorithms for
that, due to its buoyancy, penetrates above horizontal dispersion. Unstable Situations)
the mixed layer, but can disperse downward k. Vertical Dispersion
and re-enter the mixed layer. In the CBL, References
plume rise is superposed on the In the stable boundary layer, Gaussian
vertical dispersion coefficients are estimated Perry, S.G., D.J. Burns, L.H. Adams, R.J.
displacements by random convective Paine, M.G. Dennis, M.T. Mills, D.G.
velocities (Weil et al., 1997). as continuous functions of parameterized
vertical turbulence. In the convective Strimaitis, R.J. Yamartino and E.M.
(2) In the stable boundary layer, plume rise Insley, 1989. User’s Guide to the
is estimated using an iterative approach to boundary layer, vertical dispersion is
characterized by a bi-Gaussian probability Complex Terrain Dispersion Model Plus
account for height-dependent lapse rates, Algorithms for Unstable Situations
similar to that in the CTDMPLUS model (see density function, and is also estimated as a
continuous function of parameterized (CTDMPLUS). Volume 1: Model
A.2 in this appendix). Descriptions and User Instructions. EPA
(3) Stack-tip downwash and buoyancy vertical turbulence. Vertical turbulence
profiles are developed from measurements Publication No. EPA–600/8–89–041.
induced dispersion effects are modeled. Environmental Protection Agency,
Building wake effects are simulated for stacks and similarity (scaling) relationships. These
turbulence profiles account for both Research Triangle Park, NC. http://
subject to building downwash using the www.epa.gov/ttn/scram/dispersion_
methods contained in the PRIME downwash convective and mechanical turbulence.
Effective turbulence values are determined prefrec.htm#ctdmplus. (NTIS No. PB 89–
algorithms (Schulman, et al., 2000). For 181424)
plume rise affected by the presence of a from the portion of the vertical profile of
vertical turbulence between the plume height Perry, S.G., 1992. CTDMPLUS: A Dispersion
building, the PRIME downwash algorithm Model for Sources near Complex
uses a numerical solution of the mass, energy and the receptor height. The effective vertical
turbulence is then used to estimate vertical Topography. Part I: Technical
and momentum conservation laws (Zhang Formulations. Journal of Applied
and Ghoniem, 1993). Streamline deflection dispersion.
Meteorology, 31(7): 633–645.
and the position of the stack relative to the l. Chemical Transformation
building affect plume trajectory and Availability
Chemical transformations are generally not
dispersion. Enhanced dispersion is based on treated by AERMOD. However, AERMOD The model codes and associated
the approach of Weil (1996). Plume mass does contain an option to treat chemical documentation are available on the EPA’s
captured by the cavity is well-mixed within transformation using simple exponential SCRAM Web site (paragraph A.0(3)).
the cavity. The captured plume mass is re- decay, although this option is typically not
emitted to the far wake as a volume source. Abstract
used in regulatory applications, except for
(4) For elevated terrain, AERMOD sources of sulfur dioxide in urban areas. CTDMPLUS is a refined point source
incorporates the concept of the critical Either a decay coefficient or a half-life is Gaussian air quality model for use in all
dividing streamline height, in which flow input by the user. Note also that the Plume stability conditions for complex terrain
below this height remains horizontal, and Volume Molar Ratio Method and the Ozone applications. The model contains, in its
flow above this height tends to rise up and Limiting Method (section 4.2.3.4) and for entirety, the technology of CTDM for stable
over terrain (Snyder et al., 1985). Plume point-source NO2 analyses are available. and neutral conditions. However,
concentration estimates are the weighted sum CTDMPLUS can also simulate daytime,
of these two limiting plume states. However, m. Physical Removal unstable conditions, and has a number of
consistent with the steady-state assumption AERMOD can be used to treat dry and wet additional capabilities for improved user
of uniform horizontal wind direction over the deposition for both gases and particles. friendliness. Its use of meteorological data
and terrain information is different from • Receptor information, and CTDMPLUS compute ground-level
other EPA models; considerable detail for • Source-receptor location (line printer concentrations resulting from plume material
both types of input data is required and is map). in each of these flows.
supplied by preprocessors specifically (2) In addition, if the case-study option is (3) The model calculates on an hourly (or
designed for CTDMPLUS. CTDMPLUS selected, the listing includes: appropriate steady averaging period) basis
requires the parameterization of individual • Meteorological variables at plume height, how the plume trajectory (and, in stable/
hill shapes using the terrain preprocessor and • Geometrical relationships between the neutral conditions, the shape) is deformed by
the association of each model receptor with source and the hill, and each hill. Hourly profiles of wind and
a particular hill. • Plume characteristics at each receptor, temperature measurements are used by
a. Recommendation for Regulatory Use i.e., CTDMPLUS to compute plume rise, plume
—Distance in along-flow and cross flow penetration (a formulation is included to
CTDMPLUS is appropriate for the handle penetration into elevated stable
direction
following applications: layers, based on Briggs (1984)), convective
—Effective plume-receptor height difference
• Elevated point sources; scaling parameters, the value of Hc, and the
—Effective sy & sz values, both flat terrain
• Terrain elevations above stack top; Froude number above Hc.
and hill induced (the difference shows the
• Rural or urban areas;
effect of the hill) h. Horizontal Winds
• Transport distances less than 50
—Concentration components due to WRAP, CTDMPLUS does not simulate calm
kilometers; and
LIFT and FLAT
• 1-hour to annual averaging times when meteorological conditions. Both scalar and
used with a post-processor program such as (3) If the user selects the TOPN option, a vector wind speed observations can be read
CHAVG. summary table of the top four concentrations by the model. If vector wind speed is
at each receptor is given. If the ISOR option unavailable, it is calculated from the scalar
b. Input Requirements is selected, a source contribution table for wind speed. The assignment of wind speed
(1) Source data: For each source, user every hour will be printed. (either vector or scalar) at plume height is
supplies source location, height, stack (4) A separate output file of predicted (1- done by either:
diameter, stack exit velocity, stack exit hour only) concentrations (‘‘CONC’’) is • Interpolating between observations
temperature, and emission rate; if variable written if the user chooses this option. Three above and below the plume height, or
emissions are appropriate, the user supplies forms of output are possible: • Extrapolating (within the surface layer)
hourly values for emission rate, stack exit (i) A binary file of concentrations, one from the nearest measurement height to the
velocity, and stack exit temperature. value for each receptor in the hourly plume height.
(2) Meteorological data: For applications of sequence as run;
i. Vertical Wind Speed
CTDMPLUS, multiple level (typically three (ii) A text file of concentrations, one value
or more) measurements of wind speed and for each receptor in the hourly sequence as Vertical flow is treated for the plume
direction, temperature and turbulence (wind run; or component above the critical dividing
fluctuation statistics) are required to create (iii) A text file as described above, but with streamline height (Hc); see ‘‘Plume
the basic meteorological data file a listing of receptor information (names, Behavior.’’
(‘‘PROFILE’’). Such measurements should be positions, hill number) at the beginning of j. Horizontal Dispersion
obtained up to the representative plume the file. Horizontal dispersion for stable/neutral
height(s) of interest (i.e., the plume height(s) (5) Hourly information provided to these conditions is related to the turbulence
under those conditions important to the files besides the concentrations themselves velocity scale for lateral fluctuations, sv, for
determination of the design concentration). includes the year, month, day, and hour which a minimum value of 0.2 m/s is used.
The representative plume height(s) of interest information as well as the receptor number Convective scaling formulations are used to
should be determined using an appropriate with the highest concentration. estimate horizontal dispersion for unstable
complex terrain screening procedure (e.g., d. Type of Model conditions.
CTSCREEN) and should be documented in
the monitoring/modeling protocol. The CTDMPLUS is a refined steady-state, point k. Vertical Dispersion
necessary meteorological measurements source plume model for use in all stability Direct estimates of vertical dispersion for
should be obtained from an appropriately conditions for complex terrain applications. stable/neutral conditions are based on
sited meteorological tower augmented by e. Pollutant Types observed vertical turbulence intensity, e.g.,
SODAR and/or RASS if the representative CTDMPLUS may be used to model non- sw (standard deviation of the vertical
plume height(s) of interest is above the levels reactive, primary pollutants. velocity fluctuation). In simulating unstable
represented by the tower measurements. (convective) conditions, CTDMPLUS relies
Meteorological preprocessors then create a f. Source-Receptor Relationship on a skewed, bi-Gaussian probability density
SURFACE data file (hourly values of mixed Up to 40 point sources, 400 receptors and function (pdf) description of the vertical
layer heights, surface friction velocity, 25 hills may be used. Receptors and sources velocities to estimate the vertical distribution
Monin-Obukhov length and surface are allowed at any location. Hill slopes are of pollutant concentration.
roughness length) and a RAWINsonde data assumed not to exceed 15°, so that the l. Chemical Transformation
file (upper air measurements of pressure, linearized equation of motion for Boussinesq
Chemical transformation is not treated by
temperature, wind direction, and wind flow are applicable. Receptors upwind of the
CTDMPLUS.
speed). impingement point, or those associated with
(3) Receptor data: Receptor names (up to any of the hills in the modeling domain, m. Physical Removal
400) and coordinates, and hill number (each require separate treatment. Physical removal is not treated by
receptor must have a hill number assigned). g. Plume Behavior CTDMPLUS (complete reflection at the
(4) Terrain data: User inputs digitized ground/hill surface is assumed).
contour information to the terrain (1) As in CTDM, the basic plume rise
algorithms are based on Briggs’ (1975) n. Evaluation Studies
preprocessor which creates the TERRAIN
data file (for up to 25 hills). recommendations. Burns, D.J., L.H. Adams and S.G. Perry, 1990.
(2) A central feature of CTDMPLUS for Testing and Evaluation of the

c. Output neutral/stable conditions is its use of a CTDMPLUS Dispersion Model: Daytime
(1) When CTDMPLUS is run, it produces critical dividing-streamline height (Hc) to Convective Conditions. Environmental
a concentration file, in either binary or text separate the flow in the vicinity of a hill into Protection Agency, Research Triangle
format (user’s choice), and a list file two separate layers. The plume component in Park, NC.
containing a verification of model inputs, i.e., the upper layer has sufficient kinetic energy Paumier, J.O., S.G. Perry and D.J. Burns,
• Input meteorological data from to pass over the top of the hill while 1990. An Analysis of CTDMPLUS Model
‘‘SURFACE’’ and ‘‘PROFILE’’, streamlines in the lower portion are Predictions with the Lovett Power Plant
• Stack data for each source, constrained to flow in a horizontal plane Data Base. Environmental Protection
• Terrain information, around the hill. Two separate components of Agency, Research Triangle Park, NC.
Paumier, J.O., S.G. Perry and D.J. Burns, temperature, vertical wind direction shear where complete vertical mixing through the
1992. CTDMPLUS: A Dispersion Model (optional), vertical temperature gradient thermal internal boundary layer (TIBL)
for Sources near Complex Topography. (optional), turbulence intensities (optional). occurs as soon as the plume intercepts the
Part II: Performance Characteristics. (3) Meteorological data: TIBL.
Journal of Applied Meteorology, 31(7): Over land: Surface weather data from a h. Horizontal Winds
646–660. preprocessor such as PCRAMMET which
provides hourly stability class, wind (1) Constant, uniform wind is assumed for
A.3 OCD (Offshore and Coastal Dispersion direction, wind speed, ambient temperature, each hour.
Model) and mixing height are required. (2) Overwater wind speed can be estimated
Over water: Hourly values for mixing from overland wind speed using relationship
Reference
height, relative humidity, air temperature, of Hsu (1981).
DiCristofaro, D.C. and S.R. Hanna, 1989. (3) Wind speed profiles are estimated using
and water surface temperature are required;
OCD: The Offshore and Coastal similarity theory (Businger, 1973). Surface
if wind speed/direction are missing, values
Dispersion Model, Version 4. Volume I: layer fluxes for these formulas are calculated
over land will be used (if available); vertical
User’s Guide, and Volume II: from bulk aerodynamic methods.
wind direction shear, vertical temperature
Appendices. Sigma Research
gradient, and turbulence intensities are i. Vertical Wind Speed
Corporation, Westford, MA. http://
optional. Vertical wind speed is assumed equal to
www.epa.gov/ttn/scram/dispersion_
(4) Receptor data: Location, height above zero.
prefrec.htm#ocd. (NTIS Nos. PB 93–
local ground-level, ground-level elevation
144384 and PB 93–144392) j. Horizontal Dispersion
above the water surface.
Availability c. Output (1) Lateral turbulence intensity is
The model codes and associated recommended as a direct estimate of
(1) All input options, specification of horizontal dispersion. If lateral turbulence
documentation are available on EPA’s sources, receptors and land/water map
SCRAM Web site (paragraph A.0(3)). Official intensity is not available, it is estimated from
including locations of sources and receptors. boundary layer theory. For wind speeds less
contact at Minerals Management Service: Mr. (2) Summary tables of five highest
Dirk Herkhof, Parkway Atrium Building, 381 than 8 m/s, lateral turbulence intensity is
concentrations at each receptor for each assumed inversely proportional to wind
Elden Street, Herndon, VA 20170, Phone: averaging period, and average concentration
(703) 787–1735. speed.
for entire run period at each receptor. (2) Horizontal dispersion may be enhanced
Abstract (3) Optional case study printout with because of obstructions near the source. A
hourly plume and receptor characteristics. virtual source technique is used to simulate
(1) OCD is a straight-line Gaussian model Optional table of annual impact assessment
developed to determine the impact of the initial plume dilution due to downwash.
from non-permanent activities. (3) Formulas recommended by Pasquill
offshore emissions from point, area or line (4) Concentration output files can be used
sources on the air quality of coastal regions. (1976) are used to calculate buoyant plume
by ANALYSIS postprocessor to produce the
OCD incorporates overwater plume transport enhancement and wind direction shear
highest concentrations for each receptor, the
and dispersion as well as changes that occur enhancement.
cumulative frequency distributions for each
as the plume crosses the shoreline. Hourly (4) At the water/land interface, the change
receptor, the tabulation of all concentrations
meteorological data are needed from both to overland dispersion rates is modeled using
exceeding a given threshold, and the
offshore and onshore locations. These a virtual source. The overland dispersion
manipulation of hourly concentration files.
include water surface temperature, overwater rates can be calculated from either lateral
air temperature, mixing height, and relative d. Type of Model turbulence intensity or Pasquill-Gifford
humidity. OCD is a Gaussian plume model curves. The change is implemented where
(2) Some of the key features include constructed on the framework of the MPTER the plume intercepts the rising internal
platform building downwash, partial plume model. boundary layer.
penetration into elevated inversions, direct e. Pollutant Types k. Vertical Dispersion
use of turbulence intensities for plume
dispersion, interaction with the overland OCD may be used to model primary (1) Observed vertical turbulence intensity
internal boundary layer, and continuous pollutants. Settling and deposition are not is not recommended as a direct estimate of
shoreline fumigation. treated. vertical dispersion. Turbulence intensity
f. Source-Receptor Relationship should be estimated from boundary layer
a. Recommendations for Regulatory Use theory as default in the model. For very
OCD has been recommended for use by the (1) Up to 250 point sources, 5 area sources, stable conditions, vertical dispersion is also
Minerals Management Service for emissions or 1 line source and 180 receptors may be a function of lapse rate.
located on the Outer Continental Shelf (50 FR used. (2) Vertical dispersion may be enhanced
12248; 28 March 1985). OCD is applicable for (2) Receptors and sources are allowed at because of obstructions near the source. A
overwater sources where onshore receptors any location. virtual source technique is used to simulate
are below the lowest source height. Where (3) The coastal configuration is determined the initial plume dilution due to downwash.
onshore receptors are above the lowest by a grid of up to 3600 rectangles. Each (3) Formulas recommended by Pasquill
source height, offshore plume transport and element of the grid is designated as either (1976) are used to calculate buoyant plume
dispersion may be modeled on a case-by-case land or water to identify the coastline. enhancement.
basis in consultation with the appropriate g. Plume Behavior (4) At the water/land interface, the change
reviewing authority (paragraph 3.0(b)). to overland dispersion rates is modeled using
(1) As in ISC, the basic plume rise
b. Input Requirements algorithms are based on Briggs’ a virtual source. The overland dispersion
recommendations. rates can be calculated from either vertical
(1) Source data: Point, area or line source
(2) Momentum rise includes consideration turbulence intensity or the Pasquill-Gifford
location, pollutant emission rate, building
of the stack angle from the vertical. coefficients. The change is implemented
height, stack height, stack gas temperature,
where the plume intercepts the rising
stack inside diameter, stack gas exit velocity, (3) The effect of drilling platforms, ships,
or any overwater obstructions near the source internal boundary layer.
stack angle from vertical, elevation of stack
base above water surface and gridded are used to decrease plume rise using a l. Chemical Transformation
specification of the land/water surfaces. As revised platform downwash algorithm based Chemical transformations are treated using
an option, emission rate, stack gas exit on laboratory experiments. exponential decay. Different rates can be
velocity and temperature can be varied (4) Partial plume penetration of elevated specified by month and by day or night.
hourly. inversions is included using the suggestions
(2) Meteorological data (over water): Wind of Briggs (1975) and Weil and Brower (1984). m. Physical Removal
direction, wind speed, mixing height, relative (5) Continuous shoreline fumigation is Physical removal is also treated using
humidity, air temperature, water surface parameterized using the Turner method exponential decay.
n. Evaluation Studies Guide, Revised. OCS Study, MMS 84– of the Air Pollution Control Association,
DiCristofaro, D.C. and S.R. Hanna, 1989. 0069. Environmental Research & 35: 1039–1047.
OCD: The Offshore and Coastal Technology, Inc., Concord, MA. (NTIS Hanna, S.R. and D.C. DiCristofaro, 1988.
Dispersion Model. Volume I: User’s No. PB 86–159803). Development and Evaluation of the
Guide. Sigma Research Corporation, Hanna, S.R., L.L. Schulman, R.J. Paine, J.E. OCD/API Model. Final Report, API Pub.
Pleim and M. Baer, 1985. Development 4461, American Petroleum Institute,
Westford, MA.
Washington, DC.
Hanna, S.R., L.L. Schulman, R.J. Paine and and Evaluation of the Offshore and
J.E. Pleim, 1984. The Offshore and Coastal Dispersion (OCD) Model. Journal [FR Doc. 2015–18075 Filed 7–28–15; 8:45 am]
Coastal Dispersion (OCD) Model User’s BILLING CODE 6560–50–P

Environmental Protection Agency: Vol. 80 Wednesday, No. 145 July 29, 2015

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Environmental Protection Agency: Vol. 80 Wednesday, No. 145 July 29, 2015

Uploaded by

Copyright:

Available Formats

Vol.

Environmental Protection Agency

you for clarification, the EPA may not 1. Clarifications To Distinguish

6 U.S. EPA, 2012. Gina McCarthy Letter to Robert

be longer than 10 minutes, the presenter appropriateness of these revisions in

OAQPS, Research Triangle Park, NC. July 9, 1993.

photochemical grid models, if done in model in appendix A for such long-

clarification in the application of models. Quality Assessments

information concerning air quality models of transformation and removal (e.g.,

standard deviation of horizontal and vertical 4.2.2.1 AERMOD 4.2.2.3 OCD

Averaging time (lb/MMBtu) 2 (lb/MMBtu) 2 (e.g., hr/yr. hr/day)

Proposed Major New or Modified Source

permit limit.5 most recent 2 years.6 7 under consideration (for all

in complex terrain) to adequately represent for plotting contours; a listing of results

(2) A central feature of CTDMPLUS for Testing and Evaluation of the

You might also like