Responses to Significant Comments on the

2009 Proposed Rule on the

Primary National Ambient Air Quality Standards

for Nitrogen Dioxide

(July 15, 2009; 74 FR 34404)

Docket Number OAR-2006-0922

U.S. Environmental Protection Agency

January 2010

Table of Contents

  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc252046863"  I.	Introduction
  PAGEREF _Toc252046863 \h  1  

  HYPERLINK \l "_Toc252046864"  II.	Responses to Significant Comments on
the Scientific Evidence and Exposure/Risk Information	  PAGEREF
_Toc252046864 \h  2  

  HYPERLINK \l "_Toc252046865"  A.	General comments on the scientific
evidence	  PAGEREF _Toc252046865 \h  2  

  HYPERLINK \l "_Toc252046866"  B.	Comments on the epidemiologic
evidence	  PAGEREF _Toc252046866 \h  2  

  HYPERLINK \l "_Toc252046867"  C.	Comments on the controlled human
exposure evidence	  PAGEREF _Toc252046867 \h  10  

  HYPERLINK \l "_Toc252046868"  D.	Comments on air quality, exposure,
and risk analyses	  PAGEREF _Toc252046868 \h  13  

  HYPERLINK \l "_Toc252046869"  III.	Responses to Significant Comments
on the Adequacy of the Current Standard	  PAGEREF _Toc252046869 \h  35  

  HYPERLINK \l "_Toc252046870"  IV.	Comments on A New Short-Term NO2
Primary Standard	  PAGEREF _Toc252046870 \h  37  

  HYPERLINK \l "_Toc252046871"  A.	Indicator	  PAGEREF _Toc252046871 \h 
37  

  HYPERLINK \l "_Toc252046872"  B.	Averaging time	  PAGEREF
_Toc252046872 \h  38  

  HYPERLINK \l "_Toc252046873"  C.	Form	  PAGEREF _Toc252046873 \h  39  

  HYPERLINK \l "_Toc252046874"  D.	Approach and level	  PAGEREF
_Toc252046874 \h  40  

  HYPERLINK \l "_Toc252046875"  1.	Comments on the approach to setting
the standard	  PAGEREF _Toc252046875 \h  40  

  HYPERLINK \l "_Toc252046876"  2.	Comments on standard level	  PAGEREF
_Toc252046876 \h  45  

  HYPERLINK \l "_Toc252046877"  3.	Comments on the annual standard	 
PAGEREF _Toc252046877 \h  51  

  HYPERLINK \l "_Toc252046878"  V.	Technical Issues with Monitoring
Requirements	  PAGEREF _Toc252046878 \h  53  

  HYPERLINK \l "_Toc252046879"  A.	Comments on near-road monitor siting
requirements	  PAGEREF _Toc252046879 \h  53  

  HYPERLINK \l "_Toc252046880"  B.	Comments on area-wide monitor
requirements	  PAGEREF _Toc252046880 \h  59  

  HYPERLINK \l "_Toc252046881"  C.	Comments on the need for a research
monitoring network	  PAGEREF _Toc252046881 \h  61  

  HYPERLINK \l "_Toc252046882"  D.	Justification of meteorological
measurements	  PAGEREF _Toc252046882 \h  65  

  HYPERLINK \l "_Toc252046883"  E.	Monitoring technology	  PAGEREF
_Toc252046883 \h  65  

  HYPERLINK \l "_Toc252046884"  F.	Timing of monitor deployment	 
PAGEREF _Toc252046884 \h  68  

  HYPERLINK \l "_Toc252046885"  G.	Other monitoring issues	  PAGEREF
_Toc252046885 \h  69  

  HYPERLINK \l "_Toc252046886"  VI.	Air Quality Index	  PAGEREF
_Toc252046886 \h  72  

  HYPERLINK \l "_Toc252046887"  VII.	Comments on the Process for
Reviewing the NO2 Primary NAAQS	  PAGEREF _Toc252046887 \h  74  

  HYPERLINK \l "_Toc252046888"  VIII.	Interpretation of the Clean Air
Act	  PAGEREF _Toc252046888 \h  76  

  HYPERLINK \l "_Toc252046889"  IX.	Comments on Implementation	  PAGEREF
_Toc252046889 \h  78  

  HYPERLINK \l "_Toc252046890"  X.	References	  PAGEREF _Toc252046890 \h
 86  

 

Frequently Cited Documents tc \l1 "Frequently Cited Documents 

The following documents are frequently cited throughout EPA(s response
to comments, often by means of the short names listed below: 

Integrated Science Assessment (ISA):

	EPA.  (2008a).  ISA for Oxides of Nitrogen - Health Criteria.  National
Center for Environmental Assessment, Research Triangle Park, NC. 
Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=194645.

 

Preamble to the final rule (final notice):

	Preamble to the Final Rule on the Review of the Primary National
Ambient Air Quality Standards for Nitrogen Dioxide; to be published in
the Federal Register in January or February of 2010.

Proposal notice:	

Primary National Ambient Air Quality Standard for Nitrogen Dioxide: 
Proposed Rule.  74 FR 34404, July 15, 2009.

Risk and Exposure Assessment (REA):

EPA.  (2008b). Risk and Exposure Assessment to Support the Review of the
NO2 Primary National Ambient Air Quality Standard. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. Available at: 

  HYPERLINK
"http://www.epa.gov/fedrgstr/EPA-AIR/2009/July/Day-15/a15944.pdf" 
http://www.epa.gov/fedrgstr/EPA-AIR/2009/July/Day-15/a15944.pdf 

Acronyms and Abbreviations

AADT	Annual average daily traffic

AAM	Alliance of Auto Manufacturers

AAQM	Ambient Air Quality Monitoring

AASHTO	American Association of State Highway and Transportation
Officials

ACC	American Chemistry Council

ACSBPP	Annapolis Center for Science-Based Public Policy

AGCA	Association General Contractors of America 

AHR	Airways hyperresponsiveness

AQRL	AirQuality Research and Logistics

ALA	American Lung Association

ANPR	Advance Notice Of Proposed Rulemaking

API	American Petroleum Institute

AQI	Air quality index

ATS	American Thoracic Society

CAA	Clean Air Act

CAAPCOA	California Air Pollution Control Officer's Association

CAC	Clean Air Council

CASAC	Clean Air Scientific Advisory Committee

CBSA	Core-based statistical area

CE	Consumers Energy

CO	Carbon monoxide

COPA	Colorado Petroleum Association

CSE	Coalition for a Safe Environment

Dow	Dow Chemical Company

EDF	Environmental Defense Fund

EEI	Edison Electric Institute

EJ	EarthJustice

EMA	Engine Manufacturers Association

EPA	U.S. Environmental Protection Agency

Exxon	ExxonMobil Refining and Supply Company

FEM	Federal Equivalent Method

FEV	Forced Expiratory Volume

FRM	Federal Reference Method

GASP	Group Against Smog and Pollution

HCPHES	Harris County Public Health and Environmental Services

HEI	Health Effects Institute

HLI	Healthy Lungs Initiative

IADNR	Iowa Department of Natural Resources

INDEM	Indiana Department of Environmental Management

INGAA	Interstate Natural Gas Association of America

IPAMS	Independent Petroleum Association of Mountain States

ISA	Integrated Science Assessment

MET	Meteorological

MODNR	Missouri Department of Natural Resources Air Pollution Control
Program

NAAQS	National Ambient Air Quality Standards

NACAA	National Association of Clean Air Agencies

NAM	National Association of Manufacturers

NCDENR	North Carolina Department of Environment and Natural Resources

NESCAUM	Northeast States for Coordinated Air Use Management

NIOSH	National Institute of Occupational Safety and Health

NMA	National Mining Association

NMED	New Mexico Environment Department Air Quality Bureau

NO2	Nitrogen Dioxide

NOX	Nitrogen Oxides

NPRA	National Petrochemical and Refiners Association

NRDC	National Resource Defense Council 

NSR	New Source Review

NTAA	National Tribal Air Association

NYDOH	New York Department of Health

NYSDEC	New York State Department of Environmental Conservation

OIPA	Oklahoma Independent Petroleum Association

OLM	Ozone Limiting Method

OSHA	Occupational Safety and Health Association

PAW	Petroleum Association of Wyoming

PEL	Permissible Exposure Level

PM	Particulate Matter

ppb	Parts Per Billion

ppbv	Parts Per Billion by Volume

ppmv	Parts Per Million by Volume

PSD	Prevention of Significant Deterioration

PVMRM	Plume Volume Molar Ratio Method

QC	Quality control

REA	Risk and Exposure Assessment

RHAMC	Respiratory Health Association of Metropolitan Chicago

RIA	Regulatory Impact Analysis

SCAQMD	South Coast Air Quality Management District

SDDENR	South Dakota Department of Environment and Natural Resources

SIL	Significant Impact Level

SIP	State Implementation Plan

STEL	Short-Term Exposure Level

TFI	The Fertilizer Institute

TXCEQ	Texas Commission on Environmental Quality

UPA	Utah Petroleum Association

UARG	Utility Air Regulatory Group

VADOT	Virginia Department of Transportation

Introduction

This document, together with the preamble to the final rule on the
review of the primary national ambient air quality standards (NAAQS) for
nitrogen dioxide (NO2), presents the responses of the Environmental
Protection Agency (EPA) to the public comments received on the 2009 NO2
NAAQS proposal notice (74 FR 34404).  All significant issues raised in
timely public comments have been addressed.  Where comments were
submitted after the close of the public comment period, EPA has
responded to the extent feasible. 

Due to the number of comments that addressed similar issues, this
response-to-comments document does not generally cross-reference each
response to the commenter(s) who raised the particular issue involved,
although commenters are identified in some cases where they provided
particularly detailed comments that were used to frame the overall
response on an issue.  

	The responses presented in this document are intended to augment the
responses to comments that appear in the preamble to the final rule or
to address comments not discussed in the preamble to the final rule. 
Although portions of the preamble to the final rule are paraphrased in
this document where useful to add clarity to responses, to the extent
any ambiguity is introduced by this paraphrasing, the preamble itself
remains the definitive statement of the rationale for the revisions to
the standards adopted in the final rule.

	In many instances, particular responses presented in this document
include cross references to responses on related issues that are located
either in the preamble to the NO2 primary NAAQS final rule, or in this
Response to Comments document.  In other instances the comment is
appropriately addressed by the Agency’s discussion in other parts of
the record.  All issues on which the Administrator is taking final
action in the NO2 primary NAAQS final rule are addressed in the NO2
NAAQS rulemaking record.    

         Accordingly, this Response to Comments document, together with
the preamble to the NO2 primary NAAQS final rule and the information
contained in the Integrated Science Assessment (ISA) (EPA, 2008a), the
Risk and Exposure Assessment (REA) (EPA, 2008b), and the Notice of
Proposed Rulemaking should be considered collectively as EPA’s
response to all of the significant comments submitted on EPA’s 2009
NO2 primary NAAQS proposed rule.  This document incorporates directly or
by reference the significant public comments addressed in the preamble
to the NO2 NAAQS final rule as well as other significant public comments
that were submitted on the proposed rule.

Consistent with the final decisions presented in the notice of final
rulemaking, comments on the following topics are addressed in this
document: the scientific evidence and exposure/risk information (section
II); the adequacy of the current NO2 standard to protect public health
(section III); revisions to the current standard in terms of indicator
(section IV.A), averaging time (section IV.B), form (section IV.C), and
level (section IV.D); revisions to the NO2 monitoring network (section
V); the air quality index (VI); the process for reviewing the standard
(section VII); interpretation of the Clean Air Act (section VIII); and
implementation of the standard (section IX);.  

Responses to Significant Comments on the Scientific Evidence and
Exposure/Risk Information

General comments on the scientific evidence

(1)	Comment:  Some commenters (e.g., AAM, ACC) stated that it is not
clear what was meant in the ISA by a “likely causal” relationship,
stating that this classification is “somewhat ambiguous” and “at
odds with” frameworks developed by IARC and NAS/IOM (ACC, p. 2) These
groups concluded that it would have been preferable not to include this
characterization of the evidence in the ISA.  

	Response:  EPA’s causality framework is not at odds with other
frameworks. EPA’s use of a five-level hierarchy in the ISA is
consistent with EPA’s Guidelines for Carcinogen Risk Assessment (EPA,
2005).  Excerpts from those guidelines are included in Annex A (AX1 pp.
1-6 to 1-9); the second of five descriptors is “Likely to be
Carcinogenic to Humans.” IARC’s classification of carcinogens is
also excerpted in Annex A (AX1 pp. 1-14 to 1-20), and it includes Group
1 (carcinogenic), Group 2A (probably carcinogenic), Group 2B (possibly
carcinogenic), Group 3 (not classifiable) and Group 4 (probably not
carcinogenic). In contrast with commenters’ assertion that IARC’s
system differs from EPA’s, these five groups and subgroups are clearly
analogous to the five tiers in EPA’s framework. Commenters also refer
to the NAS/IOM framework (IOM, 2007) that was used as a resource for
EPA’s causal framework and excerpted in Annex A (AX1 pp. 1-9 to 1-12).
The four categories used in that system – “sufficient”,
“equipoise and above”, “below equipoise” and “against” –
are different from those used in most other classifications. However,
the IOM report included detailed discussion of the approach to
evaluating evidence from across scientific disciplines that served as an
important resource for EPA’s weight of evidence determination. In
developing its framework drawing causal inferences, EPA carefully
considered and drew upon the work of previous assessments to build a
framework for determination of causality that is consistent with EPA’s
cancer risk assessment guidelines as well as those of other expert
organizations. The development and implementation of this framework has
been lauded by the CASAC in recent ISA reviews such as that in the PM
NAAQS review (Samet, 2009a).   

Comments on the epidemiologic evidence

(1)	Comment:  Several industry groups (e.g., API, AAM, ACC) commented
that the epidemiologic studies did not support a causal relationship
between NO2 and health effects due to uncertainties and issues related
to model specification. These commenters state that EPA has ignored or
understated issues such as selection of degrees of smoothing appropriate
to adjust for weather and time trends and selection of lag period.

	Response: EPA has not ignored issues related to model specification for
epidemiologic studies, such as selection of models and approaches to
adjust for meteorological and temporal variables. EPA agrees with
commenters that these are important issues. In this and previous NAAQS
reviews, EPA has carefully evaluated such issues, including sponsorship
of workshops and special analyses of epidemiologic data, such as the
Health Effects Institute (HEI) report (2003) cited by several commenters
(e.g., ACC, AAM). As observed in the NOx ISA (p. 3-1), extensive
discussions of the issues surrounding model selection and model
specification have been presented in the PM AQCD (EPA, 2004) and the
Ozone AQCD (EPA, 2006) and are thus not reiterated at length in this
ISA. The NOx ISA makes clear, however, that these issues were were
carefully considered in selecting studies for inclusion in the ISA and
interpreting the results of the body of epidemiologic evidence. 

	Commenters specifically refer to the Health Effects Institute report on
the reanalysis of a series of time-series epidemiologic studies to
evaluate alternative modeling strategies, and include a brief quote from
the Health Review Committee Panel (HEI, 2003): “Neither the
appropriate degree of control for time, nor the appropriate
specification of the effects of weather, has been determined for
time-series analysis.” (ACC, p. 7, quoting HEI, 2003, p. 269). Taken
in full context, it can be seen not as condemnation of all epidemiologic
time-series studies but rather as a call for further investigation. The
primary conclusion drawn by the HEI panel was, rather: “The overall
impact of revising these studies include: While the number of studies
showing an association of PM with mortality was slightly smaller, the PM
association persisted in the majority of studies. In some of the large
number of studies in which the PM association persisted, the estimates
of PM effect were substantially smaller. In the few studies in which
investigators performed further sensitivity analyses, some showed marked
sensitivity of the PM effect estimate to the degree of smoothing and/or
the specification of weather” (HEI, 2003, p. 269). The results of this
extensive set of reanalyses was thoroughly discussed in the PM AQCD
(EPA, 2004, Section 8.4.2). EPA has not ignored or downplayed these
issues, in fact, EPA has scrutinized epidemiologic studies and has
funded independent reanalyses of epidemiologic studies.

EPA observes that the reanalyses discussed above and cited by commenters
are primarily on PM health effects studies. Few reanalyses have
specifically focused on NOx; however EPA has recognized that model
specification issues are generally similar for both PM and the gaseous
criteria air pollutants. EPA agrees that model specification and control
of time-varying factors such as weather, remains important and has
considered the available evidence in its evaluation of epidemiologic
studies.  The EPA also agrees with commenters that season-specific
analyses have not been widely used in epidemiologic studies for NOx,
however, EPA observes that a number of studies included in the ISA did
report results from seasonal models.  The EPA disagrees that the
available studies can support a conclusion that air pollution effects
are consistently and strongly modified by season. In summary, EPA has
not ignored the potential influence of model selection and specification
in reviewing epidemiologic studies; rather, EPA has scrutinized such
issues in previous assessments as well as the NOx ISA. These issues were
fully considered in drawing the conclusion that there was a likely
causal relationship between short-term exposure to NOx and respiratory
morbidity.

(2)	Comment:  AAM challenges ISA conclusions on several health outcome
categories on pp. 31-44. The commenter contends that “EPA overstates
the strength and consistency of epidemiological evidence regarding
various potential health effects”.

	Response:  The EPA disagrees with these commenters approach to
assessing health effects evidence as well as their conclusion regarding
the lack of a scientific basis to support the continuation of NAAQS to
protect against the health effects of NO2.  The EPA contrasts these
commenters’ narrow focus on counting the numbers of epidemiologic
studies that report results with statistical significance, without
regard to other considerations that are important to consider in a
comprehensive appraisal of the evidence, with the approach taken by EPA
in the ISA.  Specifically, EPA recognizes the distinction between
evaluation of the relative scientific quality of individual study
results, and evaluation of the pattern of results in a body of evidence.
 The EPA has done both.  The more detailed characterizations of
individual studies include assessment of the quality of the study, based
on criteria for assessment of the epidemiologic studies that are
described in Section AX 1.1.2 (pg. 1-3) and Figure AX1.1-1 (pg. 1-2) of
the Annexes to the ISA for Oxides of Nitrogen.  Statistical significance
is an indicator of the precision of a study’s results, which is
influenced by the size of the study, as well as exposure and measurement
error, model specifications, and other such factors.

	In developing an integrated assessment of the health effects evidence
for NO2, EPA has emphasized the importance of examining the pattern of
results across various studies, and not focusing solely on statistical
significance as a criterion.  It is important not to focus on results of
statistical tests to the exclusion of other information.  As observed by
Rothman (1998): 

	

Many data analysts appear to remain oblivious to the qualitative nature
of significance testing.  Although calculations based on mountains of
valuable quantitative information may go in to it, statistical
significance is itself only a dichotomous indicator.  As it has only two
values, significant or not significant, it cannot convey much useful
information….Nevertheless, P-values still confound effect size with
study size, the two components of estimation that we believe need to be
reported separately.   Therefore, we prefer that P-values be omitted
altogether, provided that point and interval estimates, or some
equivalent, are available.  (Rothman, 1998, p. 334)

The concepts underlying EPA’s approach to integrated assessment of
statistical associations reported for the health effects on NO2 have
been discussed in numerous publications, including a recent report by
the U.S. Surgeon General on the health consequences of smoking (Centers
for Disease Control and Prevention, 2004).  This report also cautions
against over-reliance on statistical significance in evaluating the
overall evidence for an exposure-response relationship.

Hill made a point of commenting on the value, or lack thereof, of
statistical testing in the determination of cause: “No formal tests of
significance can answer those [causal] questions.  Such tests can, and
should, remind us of the effects the play of chance can create, and they
will instruct us in the likely magnitude of those effects.  Beyond that,
they contribute nothing to the ‘proof’ of our hypothesis” (Hill
1965, p. 299).

Hill’s warning was in some ways prescient, as the reliance on
statistically significant testing as a substitute for judgment in a
causal inference remains today (Savitz et al., 1994; Holman et al.,
2001; Poole 2001).  To understand the basis for this warning, it is
critical to recognize the difference between inductive inferences about
the truth of underlying hypotheses, and deductive statistical
calculations that are relevant to those inferences, but that are not
inductive statements themselves.  The latter include p values,
confidence intervals, and hypothesis tests (Greenland 1998; Goodman
1999).  The dominant approach to statistical inference today, which
employs those statistical measures, obscures this important distinction
between deductive and inductive inferences (Royall 1997), and has
produced the mistaken view that inferences flow directly and inevitably
from data.  There is no mathematic formula that can transform data into
a probabilistic statement about the truth of an association without
introducing some formal quantification of external knowledge, such as in
Bayesian approaches to inference (Goodman 1993; Howson and Urbach 1993).
 Significance testing and the complementary estimation of confidence
intervals remain useful for characterizing the role of chance in
producing the association in hand (CDC, 2003, pp. 23-24).

Accordingly, the statistical significance of individual study findings
has played an important role in EPA’s evaluation of the study’s
results and EPA has placed greater emphasis on studies reporting
statistically significant results.  However, in the broader evaluation
of the evidence from many epidemiologic studies, EPA has also emphasized
the pattern of results for drawing conclusions on the relationship
between NO2 and health outcomes, as well as consideration of the
integration of epidemiologic evidence with findings from laboratory
studies.

The EPA considered the results of studies conducted in many different
countries to draw conclusions about the likelihood of a causal
relationship between NO2 and health outcomes.  Because EPA places
greater weight on US and Canadian studies in determining standard
levels, the ISA, REA and proposal notice present graphical results from
epidemiologic studies in these two countries, standardized to a common
increment of NO2, and based on similar analytic strategies (i.e.
single-pollutant results).  EPA believes that the examination of
multipollutant model results and the inherent instability that often
occurs in effect estimates for correlated pollutants in such studies
justifies the use of single pollutant model results, in addition to
multi-pollutant model results, for comparing effect estimates for NO2
and health outcomes.  

The comparisons across studies in the ISA begins with an evaluation of
the overall pattern of excess risk results – whether generally
positive or centered around zero - the consistency in size of effect
estimates, the precision of the studies evidenced in the width of the
confidence intervals, with special attention to comparisons of similar
effect categories.  

Finally, it is important to reiterate that the EPA’s evaluation of the
scientific evidence used in the current NOX NAAQS review was the subject
of exhaustive and detailed review by the CASAC and the public.  Two
drafts of the ISA were released for CASAC and public review at public
meetings.  Evidence related to the substantive issues raised by
submitted comments were evaluated in the ISA drafts and discussed at
length in public CASAC meetings.  This process ensured that overemphasis
on any study or group of studies was addressed.  Following the final
meeting with CASAC on the ISA, the CASAC panel found the coverage of the
literature in the ISA to be appropriate. 

(3)	Comment:  Several industry groups (e.g., API, AAM, ACC, ACSBPP, CE,
COPA, Dow, EEI, EMA, Exxon, NMA, TFI, UARG) and an individual (Roger
McClellan) commented that, given the presence of numerous copollutants
in the air, epidemiologic studies do not support the contention that NO2
itself is causing health effects.  These groups concluded that EPA has
overstated the strength and consistency of the epidemiologic evidence. 
For example, API commented that “These [epidemiologic] studies do not
provide support for a causal association between short-term NO2 exposure
and respiratory effects and, thus, do not provide appropriate support
for a 1-hour daily maximum NO2 standard.”  API also stated that
“Most [epidemiologic studies] only report statistically significant
findings in single-pollutant, but not multi-pollutant, models, and no
statistically significant finding is large, robust, or consistent.” 
Several industry commenters (e.g., API, ACC, ACSBPP) further noted that
the ISA relies primarily on single-pollutant models when interpreting
epidemiologic studies, rather than evaluating the results within the
context of the full suite of air pollutants.  These commenters noted
that this can lead to overestimating NO2 effects and to double or triple
counting by attributing effects to NO2 in the current review that are
attributed to different pollutants in other NAAQS reviews. Similar
conclusions were also expressed by other commenters. 

Response: These comments and EPA’s responses are discussed in section
II.E.2.a of the final rule.  The Administrator’s consideration of the
epidemiologic evidence specifically as it relates to decisions on the
adequacy of the current standard, the averaging time of a new short-term
standard, and the level of the short-term standard are discussed in
sections II.E and II.F of the final rule.  

EPA has not focused solely on the results of single pollutant models,
but has also carefully examined the implications on multiple pollutant
results.  The greatest weight of evidence for multi-pollutant results
has been placed on two-pollutant models (NO2 plus one additional
pollutant), as the inclusion of each additional pollutant in the model
can decrease model stability.  This decrease in model stability is often
reflected in wider confidence intervals, making it less likely for a
statistically significant result to be observed.  Thus, when a
statistically significant effect estimate observed in a single pollutant
model is no longer statistically significant in a copollutant model
(even though the magnitude and direction of the effect estimate has not
changed substantially), this may be an artifact of model instability. 
Results in Figures 3.1-5, 3.1-7, 3.1-10 and 3.1-11, indicate that the
association of NO2 with respiratory morbidity is robust to the addition
of copollutants.  While some individual studies may report specific
findings that are more influenced by copollutants, Figures 3.1-5, 3.1-7,
3.1-10 and 3.1-11 clearly demonstrate EPA’s conclusion of robustness. 
  

The EPA strongly disagrees that the agency is “double or triple
counting by attributing effects to NO2 in the current review that are
attributed to different pollutants in other NAAQS reviews”.  The EPA
consistently recognizes that other pollutants are also associated with
health outcomes, as is reflected in the fact that EPA has established
regulations to limit emissions of the particulate criteria pollutants as
well as other gaseous criteria pollutants.  In its assessment of the
health evidence regarding NO2, EPA has carefully evaluated the potential
for confounding, effect measure modification or other interactions
between NO2 and other criteria pollutants, and concluded that the
results attributable to NO2 are robust.

(4)	Comment:  Some industry groups (e.g., API, Exxon Mobil (Exxon))
commented that reliance on central monitors in epidemiologic studies
leads to a high degree of exposure misclassification.  Specifically,
Exxon commented that “there is a lack of correlation between
measurements from ambient monitors for NO2 and those from personal
monitors for NO2. This makes the observational epidemiology data invalid
for assessing the connection between exposure and health effects for
NO2.”  API commented that “All [epidemiologic] studies used
measurements from central monitors, which likely led to a high degree of
exposure misclassification.”  

	Response:  As discussed in the ISA (section 5.2.2), EPA agrees that
there is variability in the extent to which NO2 concentrations measured
by ambient monitors correlate with personal exposures.  However, EPA
disagrees with the conclusion of the commenter that “This makes the
observational epidemiology data invalid for assessing the connection
between exposure and health effects for NO2.”  Rather, as noted in the
conclusions of the ISA (section 5.2.2, p. 5-3), “The errors and
uncertainties associated with the use of ambient NO2 concentrations as a
surrogate for personal exposure to ambient NO2 generally tend to reduce
rather than increase effect estimates, and therefore are not expected to
change the principal conclusions from NO2 epidemiologic studies.” 
Therefore, EPA continues to judge that it is appropriate to consider the
NO2 epidemiologic evidence in reviewing the NAAQS.

(5)	Comment:  UARG also stated the following with regard to the
epidemiologic evidence: 

Another compelling source of uncertainty regarding the epidemiologic
studies on which EPA relies is the discrepancy between the levels at
which health effects were observed in these studies versus human
clinical studies. The epidemiologic studies on which EPA relies in the
Proposed Rule report effects at NO2 concentrations much lower than the
doses used in human clinical studies.

Response:  As discussed in detail in the final notice (sections II.A and
II.F.4), controlled human exposure studies reported effects down to the
lowest concentration evaluated; no studies have been conducted at lower
concentrations.  While these concentrations are higher than the mean
concentrations reported in epidemiologic studies, it is important to
note that epidemiologic studies conducted in the United States have
reported associations between ambient NO2 concentrations measured at
area-wide monitors in the current network and increased respiratory
symptoms, emergency department visits, and hospital admissions. 
Area-wide monitors in the urban areas in which these epidemiologic
studies were conducted are not sited in locations where localized peak
concentrations are likely to occur.  Thus, they do not measure the
highest ambient NO2 concentrations across the area.  Rather, the
area-wide NO2 concentrations measured by these monitors are used as
surrogates for the distribution of ambient NO2 concentrations across the
area, a distribution that includes NO2 concentrations that are higher
than the area-wide concentrations measured in study locations.  

As noted in the ISA (section 5.3.2.1 and Figure 2.4-13), 1-hour NO2
concentrations measured at area-wide monitors in the United States have
been observed in the range of 100 to 200 ppb.  In addition, even higher
1-hour NO2 concentrations could occur on and/or around major roads,
where NO2 concentrations could be 30-100% higher than measured by
existing area-wide monitors (section II.A.2 of the final notice). 
Therefore, EPA does not agree that the NO2 concentrations that occurred
in the locations of epidemiologic studies did not include concentrations
that overlapped with those reported to cause respiratory effects in
controlled human exposure studies.  

(6)	Comment: AAM contends that “the ISA uses the results of
respiratory symptom studies to claim coherence with the hospital and ED
admission results.  However the authors of the Mortimer et al. and
Schildcrout et al. multi-city studies that the Agency relies on do not
implicate NO2, per se, but summer time air pollution and fine PM,
respectively”, and claims that the characterization of the results in
studies by Mortimer et al. (2002), and Schildcrout et al. (2006) is not
consistent with the authors’ conclusions.  

	Response: The commenter is wrong in the characterization of these
studies.  They have inaccurately characterized the authors’
conclusions from these studies.  In fact, Mortimer et al. (2002)
concluded that “Nitrogen dioxide, sulphur dioxide, and particles with
a 50% cut-off aerodynamic diameter of 10 µm were associated with
increases in symptoms, with nitrogen dioxide exhibiting the strongest
influence”.  Similarly in the Schildcrout et al. (2006) study, the
authors conclude that “Among the pollutants studies, carbon monoxide
and nitrogen dioxide appeared to capture the most relevant health
information. Sulfur dioxide showed less evidence for a relation with
asthma exacerbations unless it was considered in combination with carbon
monoxide or nitrogen dioxide.  There was no evidence of a warm-season
effect of ozone or of a year-round effect of PM10.” In fact, fine PM
(PM2.5) was not even included in the study by Schildcrout et al. (2006),
though the commenter contends that the authors implicate "fine PM" as
the agent causing the respiratory morbidity.

(7)	Comment:  Several industry groups (e.g., ACC, API, NAM) commented on
the reliance on the epidemiologic study by Delfino et al. (2002). 
Specifically, as part of a Request for Correction (RFC) submitted under
EPA’s Information Quality Guidelines, NAM stated the following: 

In the Final REA, EPA relied on a purported association between
short-term NO2 exposure and asthma from a study that was not properly
reviewed in the Final ISA to support selection of a lower bound for
potential short-term NO2 standards. This use of a study that has not
been fully reviewed by EPA scientists violates EPA Guidelines requiring
use of the “best available science.” The study in question did not
find any association between asthma symptoms and NO2 exposure after
controlling for the effect of particulates. EPA must include a proper
review of this study in the Final ISA, and must explain why it believes
the study would provide any support for selection of a standard for NO2.


Response:  These comments and EPA’s responses are discussed in detail
in section II.E.2.a of the final rule.    

(8)	Comment:  As part of that RFC, NAM also stated the following: 

EPA assessments of several studies in the Final ISA differ materially
from analyses of these same studies in EPA documents for prior NAAQS.
Differing scientific evaluations by EPA which appear to depend on the
regulatory purpose for which data are being evaluated violate EPA
Guidelines requiring “objectivity.” EPA must either correct its
current analyses to be consistent with its prior conclusions or explain
why it believes those prior conclusions were incorrect.

Response:  These comments and EPA’s responses are discussed in detail
in section II.E.2.a of the final rule.  

Comments on the controlled human exposure evidence

(1)	Comment:  A number of industry commenters disagreed with EPA’s
reliance on a meta-analysis of controlled human exposure studies, which
was included in the final ISA.  API stated that “Evidence from
controlled human exposure studies of airway hyperresponsiveness in
sensitive individuals does not support the need for a short term
standard at the levels on which US EPA is taking comment.”  ACC
described a number of shortcomings of the ISA meta-analysis.  These
commenters generally referenced a recently published meta-analysis
(Goodman et al., 2009) and recommended that EPA rely on that
meta-analysis rather than the analysis included in the final ISA.  Roger
McClellan made similar comments and also referenced a recently published
article in support of his views (Hesterberg et al., 2009).  

Response: EPA generally disagrees with the commenters’
characterization of the “meta-analysis” of NO2-induced airway
responsiveness. Furthermore, EPA has considered the Goodman et al.
(2009) study and has provided a detailed response to these comments in
section II.E.2.b of the final rule.  

(2)	Comment:  NAM provided the following comment as part of their RFC.  

If EPA elects to retain the meta-analysis in the Final ISA, EPA must
correct the Final ISA to resolve inconsistencies between the conclusions
in the section on Airway Hyperresponsiveness and in the summary chapter.

Response:  EPA disagrees with this comment and finds there is no
inconsistency to be resolved between the discourse presented in the
Final ISA Section 3.1.3 (Airway Hyperresponsiveness) and associated text
presented in the Final ISA Section 5.3.2.1. (Respiratory Effects Related
to Short-Term Exposure). The section in Chapter 3 of the Final ISA to
which the commenter refers was clearly labeled as a summary of the
previous section. Specifically, section 3.1.3.3. was entitled “Summary
of Short-Term Exposure on Airway Responsiveness.” Summaries are
intended to provide brief statement of main points or substance of a
matter. However, just as an abstract of a paper, summaries do not
present all substantive or pertinent matter contained in the larger body
of a paper or, in this case, the previous section on airways
responsiveness. Furthermore, in Section 3.1.7 (Summary and
Integration–Respiratory Health Effects with Short-term Exposure) it is
specifically stated on pg 3-42 that, “Nonspecific responsiveness was
also increased following 30-min exposures of resting asthmatic subjects
to 0.2- to 0.3-ppm NO2 and following 1-h exposures to 0.1-ppm NO2.” In
Section 5.3.2.1 it is stated on pg 5-10 and 5-11 that, “…increases
in nonspecific hyperresponsiveness were observed … between 0.2 and 0.3
ppm NO2 for 30-min exposures and at 0.1 ppm NO2 for 60-min exposures in
asthmatics.”  Thus, there is clearly no merit to the commenters’
assertion of inconsistency in presentation of levels to which asthmatics
experience NO2-induced airway hyperresponsiveness between Chapters 3 and
5. 

(3)	Comment:  Some commenters (e.g., API, COPA) concluded that AHR is
not an appropriate endpoint to inform decisions on standard level. 
Because exposure to other agents in combination with exposure to NO2 is
required to trigger AHR, the commenters said EPA would need to
demonstrate the frequency and extent of exposure to these other agents
and whether these exposures would occur at sufficient concentrations to
actually trigger AHR.  And if AHR is occurring subsequent to ambient NO2
exposure, the commenters argued that the frequency of NO2-induced AHR
versus background rates of AHR has not been documented; therefore, the
significance to public health of a NAAQS based on this endpoint cannot
be determined (COPA, API).  Specifically, API stated the following:

EPA has not documented the frequency of NO2-induced AHR, if in fact it
is occurring, subsequent to exposure to NO2 in ambient air. Nor has the
Agency compared the estimated frequency of AHR induced through combined
exposure to ambient NO2 and provocative agents in the environment to
background rates of AHR produced by other causes. Thus, it is not
possible to judge the signficance to public health of a NAAQS based on
reducing the estimated reduction in the incidence or magnitude of AHR,
or whether reducing ambient NO2 will have any measurable impact on the
overall incidence of AHR. 

Response:  EPA disagrees with these commenters’ views on NO2-induced
airway responsiveness. EPA believes that its interpretation of airway
hyperresponsiveness following short-term exposure to NO2 is appropriate
and supported by the available scientific information.  Human clinical
studies provided evidence for airway hyperresponsiveness, i.e., a
heightened bronchoconstrictive response to a challenge agent, following
short-term exposure to NO2 (see sections 3.1.3 and 3.1.7 of Final ISA).
Since these experimental human studies evaluated the airway
responsiveness to challenge agents following exposure to both NO2 and
clean air as a control, they allowed determination of the independent
contribution of NO2 on airway responsiveness. An evaluation of all other
agents or factors that might affect airway responsiveness, as suggested
by the commenters, would not better define the independent effect of NO2
on airway responsiveness. Table 3.1-3 in the Final ISA specifically
provided the fraction (or frequency) of NO2-exposed asthmatics
experiencing non-specific airway hyperresponsiveness.  The ISA concluded
that “[t]ransient increases in airway responsiveness following NO2
exposure have the potential to increase symptoms and worsen asthma
control” (ISA, section 5.4).  Given this, combined with the large size
of the asthmatic population in the U.S., the REA concluded that it is
appropriate to consider NO2-induced airway hyperresponsiveness in
characterizing NO2-associated health risks (REA, section 10.3.2).  CASAC
endorsed this conclusion in their letters to the Administrator on the
final REA and on the proposal (Samet, 2008b; Samet, 2009b). 

(4)	Comment: Comment: API expressed the concern that EPA is not using a
standard approach to establishing NAAQS.  As an example they note that,
while EPA has characterized Forced Expiratory Volume (FEV) decrements in
the past as adverse health effects, the NO2 Gradient meta-analysis
(Goodman et al, (2009) showed decrements in FEV of only 1.6 percent.  In
revising the ozone NAAQS, a decrement of 3 percent was characterized by
EPA as “relatively small” and seemingly not adverse. 

Response:  EPA disagrees with the commenters’ interpretation of
NO2-induced changes in airway responsiveness and their approach taken to
discern the magnitude of this response. Due to differences in study
protocols in the NO2-airway response literature (ISA, section 3.1.3),
EPA judged it appropriate in the ISA meta-analysis to assess only the
fraction of asthmatics experiencing increased or decreased airway
responsiveness following NO2 exposure. Examples of differences in the
study protocols include the NO2 exposure method (i.e., mouthpiece versus
chamber), subject activity level (viz., rest versus exercise) during NO2
exposure, choice of airway challenge agent, and physiological endpoint
used to quantify airway responses. Therefore, EPA judged it
inappropriate in the ISA meta-analysis to try to assess the magnitude of
the NO2-induced change in airway responsiveness. EPA further notes that
Goodman et al. (2009) also recognized heterogeneity among studies as a
limitation in their analyses.  

EPA further disagrees with the commenter’s comparison and
interpretation of the effect of ozone on FEV1 versus the effect of NO2
on airway responsiveness to challenge which was subsequently assessed by
FEV1 in some studies and by other physiological endpoints such as
specific airway resistance in other studies.  Ozone itself causes
transient decrements in FEV1, whereas, NO2 does not. Rather, in
individuals with asthma, NO2-induces an increase in the responsiveness
of the airways to subsequent challenge. FEV1 is only one of several
physiological endpoints used to discern that a change in airway
responsiveness has occurred. Based on studies that evaluated these
different endpoints, the ISA concluded that “[t]ransient increases in
airway responsiveness following NO2 exposure have the potential to
increase symptoms and worsen asthma control” (ISA, section 5.4). 
Given this, combined with the large size of the asthmatic population in
the U.S., the REA concluded that it is appropriate to consider
NO2-induced airway hyperresponsiveness in characterizing NO2-associated
health risks (REA, section 10.3.2).  CASAC endorsed this conclusion in
their letters to the Administrator on the final REA and on the proposal
(Samet, 2008b; Samet, 2009b). In summary, ozone causes transient changes
in FEV1, whereas, NO2 causes transient changes in airway responsiveness.
 These are different health endpoints and EPA disagrees with the
commenter’s comparison and interpretation.

(5) 	Comment:  AAM concludes that “[r]eliance on the Orehek et al.
1976 study in an unpublished meta-analysis to claim an effect at 0.10
ppm is scientifically unsound.”  

	Response:  Orehek et al. (1976) was peer-reviewed and published in a
scientific journal. EPA specifically recognized criticism of the
statistical approach used by Orehek in section 15.3.1 of the 1993 NOx
CD. However, the statistical significance of findings presented by
Orehek et al. (1976) were not material to inclusion of data into the
meta-analysis provided in the 1993 NOx CD or the Final ISA. EPA rejects
the assertion that inclusion of the Orehek et al. (1976) study in its
meta-analysis was scientifically unsound. EPA further notes that the
Orehek et al. (1976) study data were also included in the analysis by
Goodman et al. (2009).

(6)	Comment:  AAM also concludes that “[t]he effects of NO2 from
controlled studies have not changed materially since the last review.”
 Given this, they question EPA’s conclusion that exposure to NO2
concentrations at or above 100 ppb could increase airway responsiveness
in asthmatics.  

, 66% of resting asthmatics (p ≤ 0.05) experienced an increase in
airway responsiveness. Therefore, as discussed in section II.E.2.b of
the final rule, EPA concludes that scientific evidence reviewed in both
the 1993 NOx CD and Final ISA support that exposure of asthmatic
individuals to levels of 0.1 ppm NO2 and greater results in an increase
in their airway responsiveness.  In addition, since the 1993 NOx CD,
there is a substantial new body of epidemiological evidence showing that
short-term NO2 exposure is associated with a broad range of respiratory
morbidity effects (Final ISA, Section 3.1). This new epidemiological
evidence is consistent with and supported by controlled human exposure
studies showing effects of NO2 on airway responsiveness.

(7)	Comment:  AAM concluded that, because clinical studies of
NO2-induced airway hyperresponsiveness have not reported increased
respiratory symptoms or medicine use in exposed subjects, the NO2 effect
on airway responsiveness is not adverse.  

	Response:  The intent of the clinical studies to which the commenter
refers was to evaluate the effect of NO2 exposure on airway
responsiveness, not to evaluate the effect of airway hyperresponsiveness
on respiratory symptoms or medicine use. EPA disagrees with the
commenters’ characterization of these studies.

Comments on air quality, exposure, and risk analyses 

(1)	Comment:  Several commenters discussed the analyses of
NO2-associated exposures and health risks presented in the REA.  As in
past reviews (EPA 2005a, 2007a, 2007b), EPA has estimated allowable
risks associated with the current standard and potential alternative
standards to inform judgments on the public health risks that could
exist under different standard options.  A few industry commenters
(e.g., API, NMA, UARG) concluded that the Administrator should consider
modeled exposures and risks associated with actual NO2 air quality
rather than with NO2 concentrations adjusted to simulate just meeting
the current annual standard or potential alternative 1-hour standards. 
These commenters noted that such simulations require large adjustments
to air quality and are highly uncertain and that NAAQS are intended to
address actual, rather than highly improbable, risks to health.    

Response:  These comments and EPA responses are discussed in detail in
section II.E.2.c of the final rule.  

(2)	Comment: API/AECOM (December 1, 2008 memo) specifically commented on
the EPA approach used to assess air quality that would just meet the
current and alternative standards.  The three major comments were
focused results in the EPA memo by Rizzo (2008) that investigated trends
in NO2 air quality concentrations over time.  The commenters charged
that (1) the high concentration to low concentration air quality
comparisons are not linear and not proportional, (2) even with “a good
fit to the data, the peak-to-mean ratio depends on the extremes of the
distribution”, and added that the good statistical fits Rizzo (2008)
provided are the result of the “middle percentile values, e.g., from
10% to 90%”, and (3) the ambient concentration adjustment “does not
consider the atmospheric chemistry that is involved in the formation of
NO2 from NOX emissions”.

Response:  The current and alternative standard air quality scenarios
are hypothetical scenarios, with the adjustments to the ambient
concentrations determined by an analysis of the historical trends in air
quality data at each monitor.  EPA noted in the REA (section 7.4.5) that
“there is uncertainty in adjusting concentrations” considering
uncertainty in “future source emission scenarios and how these would
relate to observed trends in current and historical air quality.” 
However, given the demonstrated strong relationship of these
concentrations over time by Rizzo (2008), EPA believed a proportional
approach would best represent the modeled air quality scenarios.  EPA
agrees that demonstration of a linear relationship between the high and
low concentration years is not the same as a proportional relationship. 
EPA acknowledged the presence of features that deviate from
proportionality in the air quality comparisons (i.e., presence of
positive and negative regression intercepts, curvilinear relationships)
as potentially contributing to uncertainty in the estimated
concentrations used in the alternative air quality standard scenarios
(REA page 131).  

     

≥95th percentile), an even greater percentage (74%) occurred at the
absolute maximum value.  This indicates that the linear relationship
extends well beyond the 90th percentile value for the majority of the
comparisons; that is on average, the upper bound of the linear range
extends through the 99th percentile of the distribution.  

EPA acknowledged that, given the scenario modeled and the deviations
from proportionality, the selected approach may contribute to both over
and under-prediction in ambient concentrations (Table 7-31, REA).  As
far as the impact of deviations from proportionality, Rizzo (2008) added
that “[i]n the majority of the monitor cases, the y-intercept of the
best fit line is positive” and that this “indicates a larger decline
between high and low year concentrations in the upper end of the
distribution than in the lower percentiles”.  Further Rizzo (2008)
stated “[m]ost although not all of the 98th and higher percentile
points that deviate from their best fit lines do fall below the line”,
meaning “that percentile point has declined even more strongly than
the middle of the distribution”.  For these cases (where a positive
intercept exists and upper percentile points are below the linear
regression line), when adjusting concentrations upwards based on the
annual average concentration, the adjusted upper percentile
concentrations would be lower than that observed during a comparable
high concentration year.  Most of the alternative standard scenarios
required an upward adjustment (see comment below).  Therefore, it is
possible that most of the extreme upper percentile concentrations used
in the alternative standard scenarios were under-estimated using the
proportional adjustment approach rather than over-estimated.  At the
limited number of monitor-years noted by Rizzo (2008) where a negative
regression intercept was present and upper percentile points are above
the linear regression line, it is possible that the extreme upper
percentile concentrations used in the alternative standard scenarios
were over-estimated. 

EPA agrees with the commenter that the primary transformation reaction
for NO2 is through reaction with atmospheric oxidants, primarily ozone. 
However, this transformation reaction and the potential amount of ozone
needed for the reaction to occur is not needed to address the
hypothetical alternative air quality scenarios . One focus of the NAAQS
review is to determine the level of the pollutant that is protective of
public health.  In this instance, the integrated review plan poses a
series of questions relevant to the NOX review including “Do exposure
estimates suggest that exposures of concern for NOX-induced health
effects will occur with current ambient levels of NO2 or with levels
that just meet current, or potential alternative, standards?” (US EPA,
2007a).  As stated, the objective is to determine what level of NO2 is
considered protective against health effects associated with NO2
exposures, not what level of NO2 is plausible under certain atmospheric
conditions.  It can also be stated that future strategies that control
ambient ozone concentrations such that the ozone NAAQS would be met are
to protect against the health effects associated with ambient ozone (and
other atmospheric oxidants ozone is serving as a surrogate for), not
ambient NO2.  

(3)	Comment: Comments from the American Chemistry Council (ACC; May 30,
2008 memo) were also directed towards analyses of the current and
alternative air quality scenarios, however these commenters primarily
questioned the relationship between analyses conducted by McCurdy (1994)
used in the prior 1995 NAAQS review and the current analyses.  These
commenters suggest that conclusions drawn during this current NAAQS
review regarding the relationship between concentrations and benchmark
exceedances differ from conclusions drawn in the previous NAAQS review. 
The commenters also charge that “for each monitor in the city, EPA
determined the annual average concentration and scaled all hourly
concentrations by the ratio of the current standard (53 ppb) and the
annual average concentration.”  A series of comments were received in
three technical memos dated by the ACC on May 30, 2008 (regarding the
1st draft REA), September 26, 2008 (regarding the 2nd draft REA without
chapter 8), October 22, 2008 (regarding the 2nd draft REA chapter 8),
and summarized in their September 14, 2009 memo. 

Response: The conclusions drawn in the current NAAQS review regarding
the relationship between annual average NO2 concentrations and frequency
of 1-hour benchmark exceedances are not necessarily contrary to
conclusions made in the prior NO2 primary NAAQS review, as the ACC
suggests.  In fact, section A-6 of Appendix A (US EPA, 2008c) documents
the investigation of both linear and non-linear relationships between
number of benchmark level exceedances and the annual average NO2
concentrations using recent ambient monitoring data (i.e., years
1995-2006).  As described in the Scope and Methods Plan (US EPA, 2007),
EPA planned to perform a similar non-linear regression modeling approach
first described by McCurdy (1994).  The non-linear regression approach
used by McCurdy (1994) to support the previous 1995 NAAQS review was
developed to estimate the number of exceedances associated with varying
annual average concentrations, including the current standard of 0.053
ppm.

As part of the current review, EPA expanded the regression approach used
by McCurdy (1994) and evaluated four different regression models (either
using normal or Poisson distributions by linear and exponential links)
and using either all data combined or data stratified by location (see
Appendix A-6, US EPA, 2008c).  The current regression analysis included
construction of a model similar to that used by McCurdy (1994) (i.e., a
normal distribution, an exponential link, and stratified by location). 
The performance of all regression models constructed was evaluated using
R-square and log-likelihood statistics.  While a model employing a
Poisson distribution and using an exponential link function demonstrated
the best performance of the models currently investigated (i.e., it was
still a non-linear relationship though different from that derived by
McCurdy (1994)), the overall predictive capabilities of such a
regression approach was generally poor, particularly for evaluating
annual average NO2 concentrations just meeting the current standard. 
For many locations, there were extremely wide 95% prediction intervals
estimated using the regression models, a direct result of data
limitations (i.e., there were very few exceedances of benchmark
concentrations > 150 ppb using recent ambient concentrations and most
locations had annual average NO2 concentrations well below the current
standard).  These results presented in the REA are entirely consistent
with the statement made in the 1995 staff paper that was cited by the
ACC as a different conclusion:

However, because all areas of the country reporting NO2 air quality data
are attaining the existing standard and because of the nonlinear
relationship of 1-hour peaks and annual averages, it is not possible to
estimate with any degree of confidence what the frequency and magnitude
of 1-hour peaks would be if the standard was selected from the upper end
of the suggested range.

Based on the performance results of the regression analyses (and
documented in Appendix A-6), an alternative approach was selected to
evaluate the number of benchmark exceedances associated with annual
average and other time-averaged NO2 concentrations.  Rather than using a
regression model constructed from the as is air quality to extrapolate
to ambient levels approaching the current or alternative standards, EPA
adjusted the ambient NO2 concentrations such that they simulate levels
of NO2 that just meet a  standard.  These adjusted ambient data “allow
comparisons of the level of public health protection that could be
associated with just meeting the current and potential alternative
standards” (REA, page 59).

A proportional approach was used to adjust the as is air quality to
simulate air quality just meeting the current and alternative standards
(see REA section 6.3.1 and Appendix A-7).  Because all locations
currently have ambient NO2 concentrations below the current standard, an
approach was needed to, at a minimum, adjust concentrations upwards to
reflect concentrations that just meet the current annual average
standard in the location.  The approach also needed to be appropriate
for adjusting air quality that just meets several potential alternative
standards, standards of varying form (98th and 99th percentile),
concentration level (50, 100, 150, 200 ppb), and averaging time
(1-hour).  A proportional adjustment approach was selected by EPA and
justified by the results of two analyses: the evaluation of the
stability in hourly and annual average concentration variability over
the span of the data used in the analysis (Appendix A-7) and the
comparisons of concentration distributions of historical and recent air
quality (Rizzo, 2008).  While annual and hourly NO2 concentrations have
steadily declined over time, the concentration variability in each of
these measures has generally remained constant over the same time period
(Figure A-100).  Concentration trends within monitoring sites in six
cities were evaluated to determine how the distribution of air quality
changed with time at each monitor.  In comparing high concentration
(historical air quality) to low concentration years (recent air quality)
at the same ambient monitor, the majority of monitors demonstrated
features of proportionality, that is the proportional change in
concentration at each percentile of the air quality distribution was
similar.  The results of both of these analyses indicated that a
proportional approach can be used to adjust ambient concentrations
(upwards or downwards) in representing alternative air quality
conditions.  

Ambient measurements at a single monitor were used within each of the 17
named locations and with the two aggregate locations to estimate the
specific adjustment factors needed to simulate air quality just meeting
the current and alternative standards.  This single design monitor in
each location was selected by having the highest annual average NO2
concentration or the highest 98th or 99th 1-hour daily maximum NO2
concentration to estimate the respective adjustment factors.  Because
the concentration adjustment was considered proportional, all 1-hour
concentrations were multiplied by the same specific factor developed for
each air quality scenario and in each location.  The adjustment factors
developed from the single monitor in each location was applied to
concentrations at all monitors within each location, such that only one
monitor had concentrations that just met the current or alternative
standard (i.e., the design monitor), while all other monitors in that
location had ambient concentrations below that of the current or
alternative standards (page 61, REA).  Following the proportional
adjustment of ambient NO2 concentrations (as well as when using the as
is air quality), the number of exceedances of each benchmark were
counted.  This approach does not assume there is a linear relationship
between the annual average NO2 concentration and the number of benchmark
exceedances.  The commenter is confusing the proportional approach used
to adjust the ambient concentrations (where comparisons of high
concentration with low concentrations demonstrate features of
proportionality – see Rizzo, 2008) with the regression models that
related the number of exceedances with the annual average concentrations
(which is not linear, as demonstrated by the conclusions in the prior
1995 review, is discussed above, and is documented in REA Appendix A-6).
 The two approaches are not comparable in this regard, though the
objective of each analysis is the same; to estimate of the number of
1-hour benchmark exceedances when considering alternative air quality
scenarios.

There is uncertainty associated with such a concentration adjustment and
has been qualitatively described in REA section 7.4.5.  However, EPA
notes that there is likely less uncertainty in the approach and results
generated when using adjusted concentrations that are closest to
existing as is air quality.  For most locations, this was the either a
98th or 99th percentile alternative standard at a 1-hour daily maximum
concentration of 50 ppb (see Table A-107).  On average, to have air
quality just meet the alternative 1-hour daily maximum standard level of
50 ppb, the minimum concentration adjustment was less than 10% and
ranged from a low of zero percent (i.e., the as is air quality for a
given location had either the 98th or 99th 1-daily maximum equal to 50
ppb) to only as high as 28 percent.  To simulate just meeting the 50 ppb
level of either the 98th or 99th percentile 1-hour daily maximum
standard, a downward adjustment of air quality was applied in most
locations and year-group of data considered.  To just meet the 100 ppb
level (either form of the 1-hour daily maximum standard), all locations
required an upward adjustment.  At this 100 ppb level, the 99th
percentile daily maximum standard required the lowest concentration
adjustment at all locations; on average, the upward concentration
adjustment was 70% (min 1%, max 131%).  Taken together, this means that
most locations have as is concentrations within the range of the two
alternative standard levels of 50 to 100 ppb and within the range of the
newly proposed standard.  As the data for simulating just meeting these
standard levels required the least concentration adjustment, there is
less of an impact of the uncertainty associated with the air quality
adjustment procedure on the estimated number of daily benchmark
exceedances associated with these air quality scenarios. 

(4)	Comment: AAM (September 14, 2009 memo) claimed that the approaches
used in the REA to characterize health risks from ambient NO2 lead to
significant overestimation of actual risk.  Specifically, they argued
that the first approach (comparing monitoring data with potential health
benchmarks) is known to overestimate the distribution of actual human
exposures because monitored concentrations do not necessarily equal
human exposures. 

Response: EPA believes that the ambient air quality has been
appropriately characterized in chapter 7 of the REA.  The REA also
provides ample context for how the ambient air quality data can serve as
a useful indicator of human exposure (sections 2.3.2, 6.1, 9.6, and
10.3.2).  EPA recognizes that population personal exposure is not
directly equivalent to ambient air quality measurements however there is
no justification for discounting the ambient monitoring data simply
because they are perceived by AAM to “overestimate personal
exposures.”  The data represent actual measurements of NO2 levels that
a population may be exposed to, provided the population encounters the
given concentration in both space and time.

EPA notes that the available evidence indicates that at-risk populations
can be exposed to ambient NO2 concentrations greater than that
represented by an ambient monitor.  EPA has clearly stated in the REA
and NPRM that the current ambient monitoring network does not have a
sufficient number or appropriate siting of monitors to reflect NO2
concentrations occurring on or near major roads in most, if not all,
urban areas of the U.S.  The existing monitoring network is not designed
to capture the spatial gradient in NO2 concentrations surrounding
roadways in any location, which is why modeling approaches were selected
by EPA staff to estimate on- and near-road NO2 concentrations.

(5)	Comment: AAM (September 14, 2009 memo) commented that “there are
no reliable data indicating either on-road or near-road exposure in
recent years that exceed the 0.20 or 0.30 benchmarks” and referenced
several studies to support their claim.

	Response: EPA notes first that the benchmark values have an averaging
time of 1-hour.  None of the studies cited by the AAM reported 1-hour
NO2 concentrations.  For example, AAM cites Beckerman et al. (2008) and
Singer et al. (2004), which were both studies  where weekly average
concentrations were measured.  A paper by Roorda-Knape et al. (1998) is
also mentioned, where two-week average concentrations were measured. 
Given the expected diurnal and day-of-week variation of NO2
concentrations (see ISA, section 2.4.4) it is not a surprise that the
average concentrations reported in these studies do not approach a peak
1-hour NO2 concentration at or above 200 ppb.   

There were only two studies cited by AAM that measured NO2
concentrations inside vehicles: a study by Westerdahl et al. (2005)
conducted in Los Angeles, CA and one by Riediker et al. (2003) conducted
in Raleigh, NC.  EPA notes that the researchers in the Los Angeles study
did measure NO2 concentrations across a two-hour averaging period, which
is the closest any measurement study came to a 1-hour averaging period. 
However, based on the study design, the primary purpose of the study is
to collect data more relevant for PM exposure than for NO2 exposure. 
The time-of-year the data were collected may not correspond to when NO2
concentrations are highest.  The NOx ISA reported that “in Los Angeles
and Riverside, CA, monthly maxima tend to occur from autumn through
early winter, with minima occurring from spring through early summer”.
  EPA notes that the Westerdahl et al. (2005) measured NO2 on February
14th and 20th and April 7th and 16th (see Table 2 of Westerdahl et al.,
2005).  In addition, the study drive times when measurements were
collected did not necessarily correspond to when NO2 peak concentrations
are expected to occur.  According to the NOx ISA, “NO2 typically
exhibits daily maxima during the morning rush hours”.  Drive times
selected for evaluation in the Westerdahl et al. (2005) study were as
follows: 2-6PM, 11AM-1PM, 9AM-12PM, and 11AM-2PM, times-of-day that do
not correspond to the morning rush hours.  

In the second study cited by AAM that reported NO2 concentrations
on-roads and inside vehicles, Riediker et al (2003) reported that sample
averaging times were about 9-hours in duration and all occurred
“during late-shift patrols (3 p.m. to midnight)”.  As discussed
above by EPA, these collection times are not consistent with the
averaging time of interest (i.e., 1-hour) or the time-of-day when
maximum concentrations are expected to occur (i.e., morning rush hours).
 Second, the monitoring events in the Riediker et al. (2003) study did
not always occur on- or near major roadways.  Often times the automobile
was parked in a lot, potentially further limiting the measurement of any
maximum NO2 concentrations.  Riediker et al. (2003) states “[o]n
average, troopers spent 35% of their shift away from the car, either in
the office, in jail, in hospitals, for dinner”, etc.  The authors of
the study also note that “pollutant concentrations in the occupied
cars were significantly higher compared to the parked cars (Mann-Whitney
U-test)” (Riediker et al., 2003).  In their comments, AAM stated that
the maximum concentration of 0.548 ppm reported by Riedeker et al.
(2003) was “flawed” and “invalid”.  Nowhere in the Riediker et
al. (2003) paper was it stated that the 9-hour average NO2 measurement
was flawed or invalid, only mentioned as an “outlier”.  Thus, this
maximum reported NO2 concentration remains a valid measurement.  In
further review of this paper, EPA notes also that the reported maximum
8-hour average roadside NO2 concentration did in fact exceed 200 ppb
(212.1 ppb; see Table 1 of Riedeker et al. (2003)), contradicting the
above AAM comment.  Given the diurnal profile of NO2, this suggests that
on this day there was likely more than one hour where roadside
concentrations exceeded 200 ppb. 

(6)	Comment:  AAM (September 14, 2009 memo) commented that the
distribution of ratios used to represent the relationship between
on-road and ambient concentrations in the air quality characterization
(chapter 7 of the REA) resulted in overestimates of on-road NO2
concentrations.  AAM refer to a statement made by Westerdahl et al.
(2005) regarding the relationship between on-road and ambient
concentrations that “roadway concentrations of [CO and] NO2 were
usually no more than about twice the ambient concentrations”.  They
further comment that higher ratios developed using lower background
concentrations will not be appropriate in urban situations having a high
background and will “substantially over-estimate the on-road
increment”.

Response: EPA notes that this statement made by Westerdahl et al. (2005)
supports an upper bound for the ratio between on-road and away from road
concentrations of at least two.  In other words, on-road NO2
concentrations may be at least 100% higher than away from road
concentrations.  This percent difference is consistent with statements
made in the REA about the on-road concentrations that “simulated
on-road annual average NO2 concentrations are, on average, 80% higher
than the respective ambient levels at distances ≥100 m from a road”
(section 7.3.2, page 97) and statements made in the ISA regarding the
relationship between the two concentrations (section 2.5.4, page 2-36). 
In addition, the use of “usually” by Westerdahl et al. (2005) in
their description indicates that most of the time the ratio was a factor
of two or less, but also suggests that there were some instances that
the factor difference was greater than two.  This statement also
supports the ratios EPA used to adjust ambient NO2 concentrations in
estimating on-road concentrations.  Table 7-10 (REA) indicates that most
of the ratios (64% and 79%, for the summer and not summer ratios,
respectively) used by EPA in estimating on-road NO2 concentrations were
less than a factor of two.  EPA also note that in addition to the
several papers EPA used to develop the on-road ratios used in Chapter 7
of the REA, an average factor of two enhancement on- or near roads
concentrations is also reported by Beckerman et al. (2008) using the
normalized NO2 concentrations provided in that study (see Figure 4 of
Beckerman et al., 2008).

EPA believes that the ratio approach developed from measurement data
available in near-road studies and used to simulate the on-road NO2
concentrations in the REA chapter 7 was reasonable.  EPA acknowledged in
the REA that there is uncertainty in the estimated on-road NO2
concentrations, particularly when using ambient monitoring
concentrations that may have been influenced by non-road sources (e.g.,
section 7.4.6 REA, page 135).  We agree with AAM that this particular
uncertainty in the approach used to simulate on-road concentrations
would tend to lead to overestimation of on-road NO2 concentrations. 
However, EPA notes that there are other sources of uncertainty
indentified in the approach used that could also contribute to
underestimation of on-road concentrations.  Localized areas such as
those within urban street canyons, tunnels, and major intersections may
have instances where on-road NO2 concentrations are greater than a
factor of two times the ambient monitor concentration.  For example,
Vardoulakis et al. (2004) reported mean NO2 concentrations at a major
intersection were about 2.1 times greater than on-road NO2
concentrations measured a few hundred meters distance away.  The ISA
(section 2.5.4) reported that NOX concentrations can be 7 times greater
at a tunnel exit when compared with NOX concentrations measured at the
tunnel entrance.  Note that the maximum ratio used to simulate on-road
concentrations in the air quality characterization was 3.7 (REA, Table
7-10).

In addition, EPA also acknowledged that there could be uncertainty in
the application of ratios developed from rural locations to urban
locations (see REA page 137).  However, given the limited information on
the characterization of the study areas used to develop the data, the
limited number of studies containing relevant data, and that there was
no apparent difference observed in the distribution of developed ratios
when considering the study locations as having a rural or urban
designation, EPA elected to stratify and apply the ratios based on
observed seasonal differences in the distributions. 

(7)	Comment:  AAM commented (September 14, 2009 memo) that AERMOD was
developed and tested primarily for stationary source applications.  AAM
referenced a recent EPA review of 21 air-quality modeling tools that
simulate line-type sources to use in near roadways applications
(EPA/600/R-09/001) including an evaluation of AERMOD and several other
roadway dispersion models.  AAM states that this “EPA review indicates
that AERMOD has not been compared rigorously for line source
applications and that it contains a very simplistic algorithm for line
sources.”

Response: First, the highways in Atlanta were simulated not as line
sources, but as a set of area sources, a common use of this regulatory
model.

Second, the referenced EPA review (US EPA, 2008) explicitly cites
numerous comparisons of AERMOD in various configurations to the
performance of some of the other available models, with AERMOD
consistently providing reasonable concentration predictions.  For
example, “AERMOD was evaluated with respect to other models such as
ADMS-Roads, ISCST3, and CTDMPLUS.  When considering only the highest
predicted and observed concentrations, it was found that ISCST3
overpredicts by a factor of seven, on average, whereas ADMS-Roads and
AERMOD underpredicted, on average, by about 20%.  It also was determined
that ADMS-Roads performance is slightly better than AERMOD (Hanna et
al., 1999).  In complex terrain, AERMOD consistently produced lower
regulatory design concentrations than ISCST3, not an unexpected result
because ISCST3 uses algorithms from a screening model (COMPLEX1) in its
calculations.  In comparisons with CTDMPLUS and observed data, AERMOD
consistently performed better than CTDMPLUS, a model approved by EPA for
regulatory applications in complex terrain.  The model has not been
compared rigorously for line-source applications. Because AERMOD is used
most commonly for dispersion analyses of stationary point sources, area
sources, and volume sources, there is no accommodation for different
roadway geometries (e.g., bridges and deep roadway cuts).”  Although
AERMOD does not currently accommodate complex roadway geometries,
enhanced turbulence, plume rise, and other characteristics of near road
dispersion can be accounted for by selection of appropriate input
parameters, such as initial sigma-y, initial sigma-z, and emission
release heights (each of which are documented in the REA).

 

Furthermore, according to Appendix W (70 FR 68218), “AERMOD is a best
state-of-the-practice Gaussian plume dispersion model whose formulation
is based on planetary boundary layer principles” and is the current
preferred model for near-field and urban-scale air dispersion modeling
applications.  While Appendix W lists CALINE-3 as the preferred model
for roadways, this model is not generally applicable for modeling long
periods in very large, complex modeling domains such as that considered
here.  In fact, CALINE- 3 and CAL3QHCR rely on older dispersion
algorithms from ISCST2 (Bailey, 2009), further justifying the use of
AERMOD rather than CALINE-3 or CAL3QHCR. 

We note also that currently available versions of the MOVES emissions
model will predict NOX emissions 30-50 percent higher than the MOBILE
model used in the present analysis (Beardsley, 2009), suggesting that
the newer emissions model would have predicted even higher NOX
concentrations and therefore higher NO2 concentrations. 

Moreover, following analysis of this dispersion model review report (US
EPA, 2008) by Isakov (2009), AERMOD was chosen as the best platform for
further model development to address roadway scenarios.  Given the
diversity of source emissions, complexity of the scenario to be modeled
in an urban environment, and requisite spatial and temporal resolution
for detailed analyses used in estimating exposures, AERMOD was also
judged here to be the most appropriate model for this application.

(8)	Comment: AAM (September 14, 2009 memo) believes that the second
approach used to characterize health risk (estimating risk based on
detailed exposure modeling), for a variety of reasons, overestimates the
risk from NO2.  One reason identified by AAM was that since AERMOD is
not a photochemical model, it does not include the key reactions
converting NO to NO2 in urban areas.  In particular, concern was
expressed by the commenter regarding EPA assumptions made for the
photochemical conversion rate as an air pacel moves downwind, and
questioning whether peak NO2 concentrations are likely to occur on or
next to the roadway.

API/AECOM (September 14, 2009 memo) had similar comments regarding the
OLM and PVMRM methods used by AERMOD, that is “[b]oth methods are
simple and conservative because they assume that only nitric oxide
titration by ozone (Equation 2) governs NO2 formation and that all of
the ambient ozone can be consumed”.  Later they commented
“consideration should be given to using a model platform that includes
plume transport time so that if time dependant chemical reactions are
needed, they can be included as a function of transport time”. 

Response:  As discussed in the REA, the AERMOD model was selected to
provide the high spatial resolution required for this exposure
assessment.  Although the chemical parameterizations of the AERMOD model
are far less detailed than the chemistry algorithms in photochemical
models, in the near field, no photochemical model would have provided
the necessary spatial resolution for this study.   Furthermore, most
photochemical models are optimized for predictions of O3, and may not
provide adequate performance in predicting NO2, even at the coarser
spatial scales at which they are applied.  Given these considerations,
AERMOD was judged to be the most appropriate model for this application.


EPA recognizes that numerous photochemical reactions take place in the
troposphere that influence the relative concentrations of NO and NO2
from NOX emissions.  The complex reactions can be commonly summarized
as:

 (O3 + PAN + HNO3… + Particles, etc. 	(1)

(Finlayson-Pitts and Pitts, 1999).  However, to a first order, the
characterization of the ratio of concentration of NO to NO2 can be
represented by the photostationary state equation:

 	(2)

where j3 results from the following 

	NO2 + h (NO + O 	(3)

and k4 from the following

	O3 + NO (NO2 + O2. 	(4)

(e.g., Atkinson et al., 1988). 

This describes a photostationary state where NO and O3 cannot coexist in
significant concentrations and the concentration of NO2 is proportional
to the concentration of O3. The commenter(s) stated that the mechanisms
in AERMOD do not appropriately parameterize the role of organics in the
processing of NO to NO2 via equation (2). However, although NO is
converted to NO2 in a reaction sequence initiated by the hydroxyl
radical in an attack on organic compounds and involving numerous free
radicals, much of the OH and HO2 radicals involved are themselves
created by photo-reactions involving O3.  Thus, to the first
approximation, equation (2) is a reasonable approximation of the balance
of NO2 and NOX, particularly during the daytime when NO2 concentrations
are highest. 

While the OLM/ARM workgroup (1998) noted that “oxidation by ozone is
typically the main reaction for NO2 formation”, and that “the
reaction rate is essentially instantaneous”, they also noted that
“the total amount of NO2 conversion is limited by how quickly the
plume entrains surrounding air”.  Therefore, the OLM algorithms may
somewhat overstate the on-road and near-road NO2 concentrations (all
other factors being equal) by not accounting for the possibility of
limited entrainment of O3-laden air in locations with limited mixing,
such as street canyons.  However, given the rapid formation of NO2 in
the presence of ozone, we feel that incorporating transport time into a
model formulation would be unlikely to substantially affect model
predictions in this application. 

 

The location of the peak NO2 concentration in this study were derived
from objective modeling predictions, not a priori assumptions.  OLM’s
characterization of NO2/NOX ratios as a function of downwind distance is
documented in a series of sensitivity tests by MACTEC (2004).  The
concentration of NO2 is a function of both the photostationary balance
of equation (2) within the plume and the decreasing concentrations of
the dispersing plume.

Despite certain limitations in the OLM approach identified above, the
model predictions show good agreement with measured values off the
roadway.  And as discussed below, the onroad model predictions compare
favorably with reported measurement values at other locations. 

(9)	Comment:  API/AECOM commented that “there are many other sources
of NO that are already competing for ambient ozone”… “unless the
ambient monitor is located in an area similar to the source, it is
unrealistic to assume that all of the monitored ozone is available to
convert NO to NO2” at the source location.

Response:  We agree that the modeling methodologies could be
strengthened by enhanced resolution of O3 concentrations.  However, the
scale over which the reactions occur is critical.  Monitors should
represent the average concentrations to which the plumes are exposed in
traveling from a source to the receptor, not the O3 concentrations
immediately adjacent to source, where the noted titration will reduce
the O3 concentrations below that available elsewhere in the domain. 

The exception would be where the receptor is immediately adjacent to the
source, e.g., the on-road and near-road microenvironments.  In these
cases, the OLM algorithms may somewhat overstate the on-road and
near-road NO2 concentrations (all other factors being equal) by not
accounting for the possibility of limited entrainment of O3-laden air in
locations with limited mixing, such as street canyons.  However, for
this study the overall model performance appears to be reasonable, both
when compared to literature values and when compared to regional
observations (see comments responses 8 through 10 below).

(10)	Comment: AAM (September 14, 2009 memo) commented that, “In
addition to mis-characterizing the spatial distribution of NO2, AERMOD
mis-characterizes the temporal distribution, as shown in Figure 8-7 of
the REA where the modeled peak NO2 from the morning rush hour occurs
about three hours earlier than the peak in the monitored concentrations
in Atlanta.”  They further comment “maximum concentrations occur
under conditions of minimum dispersion where the impact of turbulence
and heat generated by traffic will be greatest. Since the EPA review
acknowledges that AERMOD has not been rigorously evaluated for line
source applications and the algorithm is simplistic compared to other
line source models that account for turbulence and other traffic effects
in greater detail, its predictions of maximum roadway impacts are
suspect.”

	

	API/AECOM had similar comments on the diurnal concentration profile
stating “AERMOD is somewhat successful in simulating the afternoon dip
in NO2 concentrations associated with increased atmospheric ventilation,
but the model substantially overestimates (with about 50% to 70% over
prediction) and prematurely times the early morning and evening peaks.
The overestimates are probably associated with an overestimation of
ambient ozone and the overly conservative treatment of atmospheric
reactions that limit peak NO2 concentrations.”

Response:  The discrepancy between modeled and observed peak
concentration when considering the diurnal profile is less likely due to
the model parameterizations employed than to limitations in the
underlying hourly vehicle activity data from TDM modeling outputs.  The
activity data were based on TDM modeling outputs from the regional
planning organization and considered the best available information,
although somewhat limited in specification of the temporal profile.  
Furthermore, the OLM mechanism should not be related to any morning
overpredictions since the ambient ozone concentrations are generally low
at that time.  The uncertainty regarding this input variable and
magnitude of impact on estimated concentrations was acknowledged by EPA
in section 8.12.1.4 of the REA.

We agree that with equivalent emissions, maximum concentrations will
occur under conditions of minimal dispersion, such as low wind speeds
and mixing heights.  During these times, typically early morning,
effects of vehicle induced turbulence could be significant.  However, in
the present study, over-predictions occur primarily during the afternoon
traffic peak, when ambient turbulence should dominate any traffic
induced values and ambient temperatures are high, minimizing buoyancy
effects.  Thus any model over-predictions at these times are less likely
to be due to the parameterized traffic impacts on dispersion.

As discussed above, the cited EPA review notes that AERMOD is commonly
used for area source applications, which was the way roadway links were
characterized for this study; and input parameters were selected to
account for roadway effects. 

(11)	Comment: AAM (September 14, 2009 memo) commented that the data in
the “REA shows in Figures 8-6 and 8-7 that the AERMOD-estimated
concentrations substantially overestimate measured NO2 concentrations in
Atlanta particularly at the upper percentiles of the distribution. 
Figure 8-8 also indicates that AERMOD overestimates the maximum on-road
concentrations compared to the ratio method used in Chapter 7, which as
shown above, itself overestimates maximum on-road exposures.”

A few other commenters made similar comments.  For example, ACSBPP and
API concluded that AERMOD may overestimate NO2 concentrations,
particularly in the upper percentiles of the distribution.  These
commenters generally concluded that, by relying on the AERMOD dispersion
model, EPA may be overestimating the risks from current ambient
concentrations and the contributions to risk from on-roadway and
near-roadway exposures. 

Response:  Figure 8-8 compares on-road to non-road receptor
concentration ratios generated by using the concentrations estimated by
AERMOD with concentration ratios generated using the measurement data
extracted from the extant literature (i.e., the empirical method).  This
figure does not compare on-road concentration predictions.  Figure 8-8
shows that AERMOD generally predicts both higher and lower (i.e., more
variable) concentration ratios than the empirical method.  The accuracy
of neither method could be demonstrated with direct observations, due to
the lack of available on-road, or even near-road, measurement data in
the modeling domain.  But as noted in the REA, given the greater number
of receptors modeled by AERMOD, the AERMOD approach may better represent
the variability in NO2 concentration ratios than the empirical method. 

Table 8-7 presents a comparison of on-road NO2 concentration predictions
between the two methods.  The comparison suggests that the AERMOD
on-road concentration predictions are higher than those of the empirical
method by about 10 to 15 ppb at each percentile level. EPA notes however
that the accuracy of neither method could be demonstrated with direct
observations, due to the lack of available on-road measurement data in
the modeling domain.

Figures 8-6 and 8-7 show that in the off-road environment, the high end
of the concentration distribution is overestimated at the exact location
of one monitor (ID 130890002), but shows very good agreement at the
exact locations of the other two (IDs 130893001 and 131210048).  For
example, Table B-37, which presents the data used to construct Figure
8-6, shows that for the latter two monitors, the discrepancies between
the AERMOD predictions and the measured values for percentiles 90
through 99 is less than 10 ppb.  In addition, discrepancies between
observed and predicted concentrations are even less when the envelope of
all receptors within 4 km of the monitoring location is considered. 

 (12)	Comment: AAM (September 14, 2009 memo) comment that “measured
concentrations in on-road and near-road studies documented in the ISA
and summarized in the previous section demonstrate that there are no
valid measurements of NO2 exposures as high as the upper percentiles of
exposure predicted by AERMOD.  The REA refers to the 0.548 ppm maximum
NO2 concentration in Riedecker et al. (2003) to support the upper end of
the AERMOD predictions, but, as Riedecker et al. admits, it is not a
valid measurement.  There is additional evidence in the literature that
microscale monitoring will not identify unmonitored “hot spots” of
exposure to motor vehicle pollutants.  The South Coast Air Quality
Management District has carried out two studies that compared motor
vehicle air toxic exposures at microscale sites in Los Angeles suspected
of being unmonitored “hot spots” with exposures at current
monitoring sites.  In both cases, the exposures at the anticipated hot
spots were similar to the exposures at the fixed neighborhood-scale
monitoring sites.”

Response: As noted above, when comparing predicted concentrations,
either from AERMOD or the empirical method, to any measurement data, it
is important to consider differences in averaging times.  For example,
Riediker et al. (2003) measured NO2 inside North Carolina Highway Patrol
cars on duty in Raleigh and determined mean concentrations of about 42
ppb with a range of 2 – 548 ppb for measurements of varying durations,
with an average sampling duration of about 9 hours.  As described above,
the authors did not state the maximum concentration of 548 ppb was
“invalid” but as an outlier.  According to Riediker et. al. (2003),
invalid measurements resulting from “laboratory or handling problems
were excluded from the analysis.”  Note that the roadside measurements
from the same study have a mean concentration of about 50 ppb with a
range of 13 – 212 ppb (about an 8 hour averaging time).  The present
study in Atlanta estimated hourly average on-road concentrations of 43
± 25 ppb (mean and standard deviation) and a maximum value of 556 ppb
(REA Table 8-7).  Even without considering study differences (e.g.,
averaging times, times-of-year, times-of-day included), the mean and
maximum AERMOD on-road concentration predictions are comparable with the
mean and maximum measured roadside concentrations reported by Riediker
et al (2003).

In another study CARB (2003) measured NO2 concentrations inside school
buses on urban and suburban/rural routes in Los Angeles with average
commute time of approximately 85 minutes and found commute average
concentrations of about 70 ppb (urban; range 34 – 120 ppb) and 45 ppb
(Rural/suburban; range 23 – 68 ppb).  The mean and standard deviation
of the on-road concentrations from the Atlanta exposure assessment are
also consistent with this measurement study.

EPA disagrees with the commenters about the relevance of the MATES
studies (II and III) referred to by the commenter to the exposure
assessment presented in the REA.  First, the MATES studies were designed
estimate risk associated with longer-term exposures to air toxics, not
short-term 1-hour exposures to nitrogen dioxide.  In fact, NO2 was not
measured in either of these studies.  The ambient monitoring was
conducted in neighborhoods potentially downwind of important emissions
sources; no measurements were made on-roadways.  Comparisons were made
between concentrations measured at neighborhood microscale monitors to
fixed site monitors located several miles away.  The studies do not
compare on-road, near-road, and away from road concentrations along a
specific transect line.  The study monitoring objectives were to not
only assess potential mobile source influence to neighborhood air
quality, but also considered influence from other localized sources. 
For example, in the MATESII study (SCAQMD, 2000) only 3 of 14 microscale
monitoring sites used for their “hot-spot” analysis were selected
“because of influence and proximity to major mobile sources.

Conclusions reported by SCAQMD (2000) are consistent with the AAM
comments in that “the monitoring at each of the 14 microscale sites
did not register significantly higher levels of any toxic air
contaminants.”  SCAQMD also added that “it cannot be concluded that
‘hot spots’ do not exist at other locations.”  The latter
statement is made because of extreme limitations noted in the microscale
monitoring program and to caution readers against drawing particular
conclusions based on the data.  SCAQMD notes that “the intent to
investigate a number of different sites, with available resources,
limited the power of the microscale study to detect localized
disparities in air toxic levels.  The microscale study should therefore
be regarded as more of a “pilot study” than as a study to
definitively address possible differences in community air pollutant
exposures within the South Coast Air Basin.  These factors should be
taken into consideration to avoid possible over-interpretation of the
results.”  EPA notes that in the summary statements made for one of
the microscale monitors placed to capture potential mobile source
influences, “measured concentrations indicate more on-road activity at
Montclair than at Fontana” (SCAQMD, 2000).

The MATESIII study does not refer to the monitoring of “hot spots”
at all.  The selection of monitoring sites “was done to ensure
sufficient resolution to monitor representative concentrations of
varying land use types and characterize spatial gradients in the
Basin” (SCAQMD, 2008).  The authors also note “[a]s in MATES II, due
to the limited number of mobile monitoring platforms, each microscale
site study lasted for a shorter duration than the overall study for
approximately two to several months.”  This suggests the same caution
should be applied regarding the over-interpretation of the results as
recommended for the MATESII study.  It was not apparent to EPA how many
of the five microscale sites used in this study were placed in areas
expected to receive significant mobile source influence, however SCAQMD
found statistically significantly higher concentrations of mobile-source
related pollutants at two of their microscale monitors compared with
corresponding measurements at paired fixed site monitors.  For example,
24-hour average concentrations of benzene, ethylbenzene, toluene, and
xylene were 5 to 6 times as higher at the Santa Ana site than at the
Anaheim site (SCAQMD 2008, page 5-8).  The other three microscale
monitors may have been sited to capture localized emissions other than
mobile sources, given high concentrations noted for hexavalent chromium,
1,3-butadiene, and manganese composition in PM10.  While limited in
direct relevance to the REA, in general, the overall trend in both MATES
studies support the importance of mobile source influence to ambient
concentrations. 

 (13)	Comment:  AAM (September 14, 2009 memo) commented on the approach
used to estimate in-vehicle and near-road concentrations used to
estimate exposures.  Specifically they state that based on the AERMOD
on-road to non-road receptor concentrations, “multiplicative factors
from 1 to as high as 10 to 30 inappropriately were being used to
estimate near-road exposures from the estimates of ambient
concentrations.”  The commenter argued that “multiplicative factors
as high as 30 are clearly suspect compared to the data in the literature
on in-vehicle NO2 exposures”  They also question the use of the same
APEX proximity factor for both in-vehicle and near-road
microenvironments given the sharp reduction in NO2 concentrations with
distance from a roadway.

 receptor at a distance ≥ 100 meters from the road.  Given the greater
diversity of the AERMOD modeled receptor locations compared with the
literature derived data (a total of 11 studies using 36 roadway sites),
we expected there would be greater variability and hence a greater range
of ratios in the AERMOD derived ratios.

The ratios used for the APEX modeling were calculated using the AERMOD
predicted hourly concentrations.  Ratios having a 1-hour averaging time
are most appropriate for this APEX application since the PROX factor is
applied to a 1-hour ambient concentration.  EPA expected that there
would be greater variability in the 1-hour ratios, likely generating a
distribution of PROX factors having a greater range of values than
observed with the longer averaging time.  EPA notes that there are no
data reported in the extant literature that have on-road/non-road
concentrations or ratios of 1-hour averaging time.  In addition to the
seasonal influences observed (summer/not-summer), EPA also stratified
the ratios by the time-of-day based on analysis of this as a potential
influential factor.  Three time ranges were selected to further stratify
the ratios.  Lognormal distributions were then fit to each of the six
stratifications, with lower and upper bounds of the distributions
approximated by using the 5th and 95th percentiles (Appendix B Table
B-42).  In probabilistic exposure analyses, setting bounds on a fitted
distribution is a commonly used method to control against exceptional or
potentially unrealistic values.  Note in REA Table 7-34, the minimum and
maximum observed on-road ratios were at the 2.3-7.8th and 94.4-97.8th
percentiles, respectively of a lognormal fitted distribution, supporting
the selection of the 5th and 95th percentiles as reasonable lower and
upper bounds.

In evaluating the ratio distributions, clearly the daytime (6AM-7PM)
ratio distributions are distinct and higher than the nightime ratio
distributions (7PM-6AM), confirming that stratification by at least two
time-of-day ranges was appropriate.  Again, seasonal differences were
consistent with that of the ratios derived from the empirical data,
confirming that seasonal stratification was appropriate.  EPA
acknowledges that it is possible that there are other potential
influential factors that could be considered in further stratifying and
applying the PROX factors, but given the limited time and resources
allocated to perform the assessment, additional analyses were not
performed.

When considering (1) the absence of observed on-road concentrations of
1-hour averaging times, (2) having no on-road concentration measurements
within the modeling domain, (3) the reasonable agreement in the AERMOD
concentration predictions with available measurement data, and (4) the
reasonable agreement in the AERMOD on-road concentration predictions
with simulated on-road concentrations using the empirical ratios, EPA
believes that the PROX factor distributions generated are an adequate
representation of the relationship between on-road and non-road
concentrations, even with upper bounds of the distributions that extend
to values of 9 to 30.  Again, EPA is not aware of any 1-hour measurement
data derived from a study conducted in a similar urban environment
across an entire year that could confirm or refute the level of the
upper bound used for each of the ratio distributions.

EPA used the same on-road PROX factor distributions to calculate the
near-road microenvironmental concentrations.  The near-road
microenvironment generally refers to locations in close proximity to
vehicle emissions such as sidewalks, bus stops, and parking garages,
where concentrations would likely be similar to those on roadways.  Not
accounting for decay/dispersion of the on-road concentration with such
short distances from a roadway was considered by EPA as an appropriate
assumption.  EPA notes that the maximum concentration does not always
occur on the roadway, it can occur at a distance from the road (e.g.,
Beckerman et al., 2008).  Therefore, there may be instances where the
near-road concentrations are either greater than or less than that
observed on the road.  In the absence of robust measurement data to
inform the development of a potential adjustment factor to approximate
on-road exposure concentrations that deviate above or below nearby
on-road concentrations, EPA assumed a relationship of unity.

EPA recognizes that when using these ratio distributions to estimate
in-vehicle or near-road exposures, there may be instances where these
ratios are unsuitably applied to an exposure event, given other
potential influential factors mentioned above that are not considered in
the exposure calculation.  This is an important uncertainty that could
contribute to both over- and under-estimates in exposure concentrations.
 We acknowledged that accurately modeling the upper percentiles of the
distribution is of concern given that the exposure metric of interest is
exposures above selected benchmark levels.  EPA qualitatively summarized
the overall impact of these and other input data uncertainties in the
REA (section 10.3.2.1) indicating that there is the “possibility that
we are over-predicting upper percentile NO2 exposures” and therefore
over-predicting the number of days with exceedances, particularly the
highest benchmark levels.  Given the duration of the exposure assessment
(i.e., an entire year), this uncertainty is likely to have a greater
impact on the number of days per year an individual was estimated to
experience an exposure above a benchmark rather than the estimated
number/percent of persons exposed above a particular benchmark level,
though the magnitude of which remains uncertain for either exposure
metric.

(14)	Comment: AAM (September 14, 2009 memo) commented that a study by
Chock (1977) “demonstrated that the turbulence and heat generated by
the traffic had a significant effect on the on-road and near-road wind
and concentration fields”.  They further add “[t]hese effects limit
concentrations that can build up on and near roadways under adverse
ambient meteorology and are not included in AERMOD.” 

Response:  We agree that road-induced turbulence and buoyancy effects
will dominate in the early morning hours where stability is likely high
and mixing heights and temperatures are low.  In these cases, the local
mechanical turbulence will overwhelm the ambient turbulence and lead to
the effects explicitly designed to be sampled by Chock. 

As noted above, on-road vehicle induced turbulence and plume rise were
addressed in the selection of input parameters for initial horizontal
and vertical dispersion and emission release heights for on-road, as
documented in the REA (section 8.4.3.3).  Additionally, to mitigate
morning over-predictions, the mechanical mixing height was raised in
cases where AERMOD-defined mixing heights were judged to be unreasonably
low, as discussed in the REA.

(15)	Comment: API/AECOM (September 14, 2009 memo) commented on the
importance of developing a database that “provides information on NO2
to NOx emission ratios” suggesting that “the estimation of the NO2
to NOX ratio is critical in modeling peak hourly NO2 concentrations.

Response:  We agree that knowledge of the initial NO2/NOX ratio is
limited and that the ratio can be variable.  Although we note that
EPA’s new mobile source emissions model, MOVES, can provide separate
estimates of NO, NO2, and NOX emissions.  

We also note that the importance of the accuracy of the initial ratio
depends on the degree of conversion of NO to NO2 predicted by the model.
 That is, in cases where more complete conversion is predicted, the
sensitivity of the final NO2/NOX ratio to its initial value is
minimized.  In the present exposure assessment where the dispersion
model predicts a high rate of conversion, this effect of this on the
output concentration will be small.

Furthermore, Carslaw (2005) attributes much of the observed increase in
NO2/ NOX emissions ratios to the increased use of diesel particulate
filters (DPF) in London, UK.  Even if this supposition is correct, it is
unlikely to be relevant to the present modeling results.  For
comparison, diesel vehicles make up over fifty percent of new light duty
vehicles in the EU (Dieselnet, 2008) whereas the North American diesel
passenger vehicle market share is less than 1 percent (Haight, 2003). 
Regardless, we agree that, generally, improved characterization of this
ratio would be advantageous in future modeling efforts.

(16)	Comment:  One commenter drew attention to a potential error that
causes underestimation of NO2 impacts using the modeling software AERMOD
(CAAPCOA).  They provided the following description this error and other
modeling shortcomings that will affect the models used to inform the
selection of a short-term NO2 standard:

“The use of AERMOD with a more refined Conversion Tier (e.g., OLM
[Ozone Limiting Method] or PVMRM [Plume Volume Molar Ratio Method])
produces lower 1-hour NO2 impacts than less refined methods do. 
However, California air district modelers have discovered that an error
exists in the multi-source, combined-plume application of AERMOD (the
OLMGROUP keyword) that biases the results low.  This means that 1-hour
NO2 levels calculated using the more refined methods may be
underestimating the impact.  Furthermore, the AERMOD model requires
newer MET [meteorological] data sets that are more complicated and
require data inputs that may not be readily available for all areas in
California.  Consequently, not all air districts have converted to the
AERMOD model.”

	The commenter urges EPA to address this potential model error prior to
establishing a 1-hr NO2 standard.

Response:  EPA was aware of this error while conducting the NO2 exposure
assessment.  The version of AERMOD that was used in our analysis had
been corrected for this error.  

(17)	Comment: In their October 22, 2008 memo, the ACC commented that the
exposure assessment results for Philadelphia (1st draft REA) were very
different from results presented for Atlanta (Final REA).  They
specifically noted “there are substantially higher numbers of
exceedances and numbers of individuals exposed to repeated exceedances
for Atlanta compared to Philadelphia”.  The commenter requested
discussion or comparisons of the differences between the two locations.

	

Response: EPA considered performing the exposure modeling in five
locations described in the Scope and Methods Plan (i.e., Atlanta,
Detroit, Philadelphia, Phoenix, and Los Angeles) however there were
limited data, time, and resources available to perform exposure modeling
in all five locations.  Following the CASAC review of the 1st draft REA
and considering comments made on that document, EPA judged the model
evaluations to be an important factor in selecting a location for an
improved exposure assessment.  Given the availability of personal
exposure measurements within a similar time frame as the exposure
modeling and a complementary quantitative risk assessment to be
performed in the same location, EPA elected to focus the exposure
modeling effort on Atlanta, GA rather than continue with the modeling
that was initiated for the 1st draft REA (i.e., Philadelphia County).

The exposure modeling approach used for Philadelphia in the 1st draft
REA (and documented in Appendix B-3 of the final REA) is not directly
comparable to the exposure modeling performed for Atlanta (REA, Chapter
8).  Therefore, a direct comparison of the exposure results obtained for
each location would be unreasonable.  As stated in REA Appendix B,
section B-1, “due to differences in the approach used in the
Philadelphia analysis, the results are not directly comparable to the
Atlanta case-study.”  For example, one of the most significant
differences between the two assessments was how the emissions from minor
roadways were addressed.  In Atlanta minor roadways were modeled as area
sources using AERMOD, considered by EPA an important improvement to the
modeling approach used for this assessment.  In Philadelphia minor roads
were not modeled as emission sources, but considered as part of the
unaccounted concentration when comparing the modeled concentrations
considering all other emission sources at the ambient monitor locations.
 It is possible that these differing model approaches used were related
to another important difference: the derived distributions of in-vehicle
and near-road proximity factors.  In Atlanta, the proximity
distributions were stratified by season and hour-of-day, having a
time-weighted geometric mean of about 2.7 (REA Appendix B, Table B-42). 
In Philadelphia, the proximity distributions were stratified only by
hour-of-day, having a time-weighted geometric mean of about 1.4 (REA
Appendix B, Figure B-11).  These examples were just two of several
important differences designed to improve the Atlanta assessment for the
2nd draft and final REA, building upon what was learned from the 1st
draft approach used for Philadelphia.  No improvements or adjustments
were made to the exposure modeling approach used for Philadelphia, thus
results for that assessment should only be considered as preliminary.

(18)	Comment:  Some commenters (e.g., API) noted that indoor NO2
concentrations can be higher than the levels that EPA is considering for
an hourly NO2 NAAQS due to indoor emission sources.  These commenters
conclude that, given that people spend the majority of their time
indoors, the Agency should consider NO2 exposures indoors when
evaluating whether the proposed NAAQS would have a meaningful effect on
risk from NO2.  

	Response:  We agree that indoor concentrations of NO2 can be elevated
relative to ambient concentrations, particularly in indoor environments
containing sources of NO2 such as gas stoves (ISA, section 2.5.5). 
Indoor sources were considered as part of the exposure assessment
presented in the REA.  Even when indoor sources were included in
modeling, the majority of exposures to peak NO2 concentrations (i.e.,
1-hour concentrations at or above 100 ppb) were attributable to
roadway-associated sources (REA, Figures 8-17 and 8-18).  

 

(15)	

 

Responses to Significant Comments on the Adequacy of the Current
Standard

(1)	Comment: CASAC agreed that, based on the available information, the
current NO2 standard is not requisite to protect public health with an
adequate margin of safety and that revisions to the standard are
appropriate.  Their letter to the Administrator on the final REA (Samet,
2008) stated that “CASAC concurs with EPA’s judgment that the
current NAAQS does not protect the public’s health and that it should
be revised.”  In supporting adoption of a more stringent NAAQS for
NO2, CASAC generally relied on the assessment of the scientific evidence
presented in the ISA, the results of assessments presented in the REA,
and the conclusions of the policy assessment chapter of the REA.    

A number of public commenters also called for revising the current
standard.  This included environmental groups (e.g., CAC, EJ, EDF, NRDC,
GASP); medical/public health organizations (e.g., AACPR, ALA, AMA, ATS,
NAMDRC, NACPR, ACCP); State, local, and tribal agencies and
organizations (e.g, NACAA, NESCAUM, agencies in CA, IA, IL, MI, MO, NC,
NM, NY, TX, VA, WI, and tribes including NTAA, Fon du Lac); and a number
of individual commenters.  These commenters generally concluded that the
current NO2 standard needs to be revised and that a more stringent
standard is needed to protect the health of sensitive population groups.
 In supporting this conclusion, these commenters typically relied upon
the evidence and information presented in the proposal and on CASAC’s
recommendation.  

Response:  We generally agree with these commenters’ conclusions
regarding the adequacy of the current standard.  The Administrator’s
conclusions regarding adequacy are discussed in more detail in section
II.E.3 of the final rule.    

(2)	Comment:  Some industry commenters (e.g., AAM, API, Dow, INGAA,
UARG), one State commenter (INDEM), and Roger McClellan expressed
support for retaining the current annual standard alone.  In supporting
this view, these commenters generally concluded that available evidence
supports a judgment that the current standard provides adequate
protection of public health.  They also typically concluded that the
available evidence and information is not sufficient to support revision
of the standard.  These commenters generally relied upon their judgments
regarding the scientific evidence and exposure/risk information and on
the uncertainties associated with that evidence and information, as
discussed above (section II) in more detail.  For example, the UARG
stated that “EPA has failed to demonstrate that the present NO2 NAAQS
is no longer at the level requisite to protect public health with an
adequate margin of safety.”  The INGAA stated that “…EPA should be
compelled to retain the current standard and defer a decision on a new
short-term standard until the science is more clearly defined.”  Roger
McClellan commented that “there is a substantial body of scientific
information to under-gird a policy judgment to re-affirm the present
annual standard set at 0.053 ppm NO2 measured at area-wide monitoring
sites.”    

Response:  These comments and EPA’s responses are discussed in detail
in section II.E.2 of the final rule.  The Administrator’s conclusions
on the adequacy of the current standard are described in section II.E.3
of the final rule.    

(3)	Comment: The Oklahoma Independent Petroleum Association (OIPA)
pointed out that “since 1990, NO2 emissions have decreased 35% despite
the 63% increase in the gross domestic product, a 45% increase in the
vehicle miles traveled, a 21% increase in population, and a 20% increase
in energy consumption.  Between 2001 and 2007, NO2 emissions decreased
by 20%.  In addition, EPA anticipates that nitrogen oxide (NOx)
emissions will decrease substantially over the next 20 years as a result
of the ongoing implementation of mobile source emissions standards.” 
As a result of these and similar statements by commenters, several
groups concluded that a more stringent NO2 standard is unnecessary
because the current standard is sufficient to protect against both long-
and short-term exposures (e.g., API, Dow, INGAA, VADOT, MODNR, INDEM). 

Response:  As noted by the commenters, NOX emissions have decreased and
are expected to continue to decrease as a result of ongoing
implementation of mobile source emissions standards.  However, the Clean
Air Act (section 109) requires that primary NAAQS “shall be ambient
air quality standards the attainment and maintenance of which in the
judgment of the Administrator, based on such criteria and allowing an
adequate margin of safety, are requisite to protect the public
health.”  Given this requirement, the primary NAAQS are set such that,
in the Administrator’s judgment, public health is protected with an
adequate margin of safety in locations that meet the standards.  As
discussed more fully in the final rule (section II.E.3), the scientific
evidence and the exposure/risk information support the Administrator’s
conclusion that important NO2-related public health risks are present in
some locations with ambient NO2 concentrations below those allowed by
the current annual standard and that a new standard is needed to provide
the requisite degree of protection for public health.  

Comments on A New Short-Term NO2 Primary Standard

This section discusses comments received on EPA’s proposed 1-hour
standard.  Some commenters provided comments on the cost or economic
impact of monitoring, implementation, or compliance associated with the
proposed NO2 NAAQS.  As noted in section I.B of the preamble, the Clean
Air Act bars consideration of costs in setting the NAAQS, and
accordingly EPA has not considered costs, including the costs or
economic impact of monitoring, implementation or compliance, in revising
the NO2 NAAQS. 

A.	Indicator

CASAC and some public commenters addressed the issue of the indicator
for the standard (AAM, API, Dow, and MODNR).  All of these commenters
endorsed the proposal to continue to use NO2 as the indicator for
ambient NOx, though some commenters expressed the need for special
considerations when using NO2 as a surrogate for NOx.  

(1)	Comment: API concluded that the proposed regulatory language would
change the denomination of the standard from one regulating NO2 to one
regulating oxides of nitrogen, with compliance to be determined by
measurement of NO2 in the ambient air.  API recommends that EPA clearly
state that this language is not a substantive change and merely reflects
the continuation of the Agency’s regulation of NO2 (and only NO2) as
an indicator for all oxides of nitrogen. 

Response:  We agree that no substantive change is being made with regard
to the indicator and that NO2 is the indicator for the oxides of
nitrogen.  This is indicated in the regulatory text for 50.11 in the
final rule.  

(2)	Comment: Some commenters (e.g., AAM) asserted that the tendency to
overestimate NO2 concentrations with the Federal Reference Method (FRM)
should be considered with regard to margin of safety, and that EPA
should develop an FRM not prone to positive interference. 

	Response:  EPA is required to set the NAAQS at a level requisite to
protect public health “allowing a margin of safety”.  Thus, EPA
takes “margin of safety” into consideration in setting the NAAQS.  
EPA separately considers the accuracy and precision of measurement
methods in determining FRM and FEM requirements. The Administrator
believes that the continued use of the chemiluminescence FRM is
appropriate for comparison to the NAAQS. The issue of interference in
FRM measurements and the discussion of the development of alternative
methods that could be used in determining NO2 concentrations are
discussed in section III.A.1 of the preamble to the final rule. 
Although the FRM can be subject to positive artifacts resulting in
varying degrees of overestimation of NO2 concentrations, this
overestimation should be minimized for monitors sited in urban
locations, particularly near NOX sources such as roads.  Additionally,
the ISA concluded that for monitors sited in urban locations near
roadways or other sources of NOX emissions, the overestimation of NO2
concentrations is typically less than 10% (ISA, section 2.3). 

Averaging time

(1)	Comment:  CASAC endorsed the establishment of a new standard with a
1-hour averaging time.  CASAC stated the following in their comments on
the proposal (Samet, 2009b): 

In reviewing the REA, CASAC supported a short-term standard for NO2 and
in reviewing the proposal, CASAC supports the proposed one-hour
averaging time in EPA's proposed rule. 

The rationale offered by CASAC in support of a new 1-hour standard was
generally the same as that put forward in the final REA and the
proposal.  Specifically, that rationale considered the available
scientific evidence, which supports a link between 1-hour NO2
concentrations and adverse respiratory effects, and air quality
information presented in the REA, which suggests that a 1-hour standard
can protect against effects linked to short-term NO2 exposures while an
annual standard would not be an effective or efficient approach to
protecting against these effects.  

A number of public commenters also endorsed the establishment of a new
standard with a 1-hour averaging time.  These included a number of State
agencies and organizations (e.g., NACAA, NESCAUM and agencies in CA, IL,
NM, TX, VA); environmental, medical, and public health organizations
(ACCP, ALA, AMA, ATS, CAC, EDF, EJ, GASP, NACPR, NAMDRC, NRDC); and a
number of individual commenters.  The supporting rationale offered by
these commenters often acknowledged the recommendations of CASAC and the
Administrator’s rationale as discussed in the proposal. 

Response:  We agree with these comments on the need for a new 1-hour
standard.  Comments on averaging time are discussed in section II.F.2.b
of the final rule.  The Administrator’s final decision on averaging
time is discussed in section II.F.2.c of the final rule. 

(2)	Comment:  Though many industry commenters recommended not revising
the current annual standard (see above), several of these groups did
conclude that if a short-term standard were to be set, a 1-hour
averaging time would be appropriate (e.g., CPA, Dow, NAM, PAW, UPA).  

Response:  As discussed above, industry commenters who disagreed with
setting a new 1-hour standard generally based this conclusion on their
interpretation of the scientific evidence and their conclusion that this
evidence does not support the need to revise the current annual
standard.  These comments are discussed in detail in section II.E.2 of
the final notice.  The Administrator’s conclusions regarding averaging
time are discussed in section II.F.2.c of the final notice.    

(3)	Comment:  Two state commenters recommended that further studies be
conducted to determine whether the 24-hr standard is as protective of
human health as the 1-hr standard (e.g., NYDOH, SDDENR).  One of these
commenters specifically noted that epidemiological studies do not
provide sufficient evidence that distinguishes effects between 1-hour
and 24-hour exposure periods (SDDENR).

Response:  The Administrator’s conclusion that available scientific
evidence supports setting an NO2 standard with a 1-hour averaging time
is described in detail in section II.F.2.c of the final notice.  In
addition, EPA notes that the court-ordered schedule for this review
requires a notice of final rulemaking to be signed by January 22, 2010.

Form 

(1)	Comment:  Many commenters emphasized that, for a standard reflecting
the maximum allowable NO2 concentration anywhere in an area (i.e., the
approach adopted in the final rule, section II.F.4.d), a form should be
chosen that provides regulatory stability.  To this end, CASAC favors a
3-year average of the 98th percentile (or 7th or 8th highest) of the
distribution of annual 1-hour daily maximum concentrations. 
Specifically, CASAC commented that “the 98th percentile is preferred
by CASAC for the form, given the likely instability of measurements at
the upper range and the absence of data from the proposed two-tier
approach.”  

A number of other commenters also recommended a form based on the 3-year
average of the 98th percentile (or 7th or 8th highest) of the
distribution of annual 1-hour daily maximum concentrations (e.g., Dow,
SDDENR, NCDENR, API, PAW, VADOT, ACC, ExxonMobil, INDEM, NESCAUM, AQRL,
and IPAMS).  These commenters typically argue that the 98th percentile
form will provide a more stable statistic and, therefore, will better
protect against unusual events and the possibility that concentrations
measured by near-road monitors will be highly variable and responsive to
small changes in monitor placement.

Response:  EPA agrees with the comments that a 98th percentile form is
appropriate.  Her consideration of comments on form and her rationale
for her final decision are discussed in detail in sections II.F.3.b and
II.F.3.c, respectively.  

(2)	Comment:  Several commenters recommended either a 99th percentile
form or a more stringent form.  For example, some commenters recommended
a “no exceedance” form, claiming that such a form would provide
“more coherence in the protection for a standard that targets peak
exposures than a form based in the 98th or 99th percentile” (ALA, EJ,
EDF, NRDC).  These same commenters noted that if EPA is unwilling to
accept a “no-exceedence” form, they strongly favor a 99th percentile
standard over a 98th percentile standard.  They pointed out that a 98th
percentile form would allow for as many as 21 exceedances in a
three-year period, and concluded that this would be unacceptable. 
Alternatively, GASP recommends using a form allowing one exceedence per
year over a three-year period, arguing that this will prevent several of
the top measurements that contribute to adverse health effects from
being dismissed.  Several commenters prefer the 3-year average of the
99th percentile (or 4th highest) of the distribution of 1-hour daily
maximum concentrations (e.g., TXCEQ, HLI, RHAMC, MODNR, NMED, NESCAUM).

Response:  The Administrator recognizes that the public health
protection provided by the 1-hour NO2 standard is based on the entire
standard including the level of the standard (see below), in conjunction
with the averaging time and form of the standard.  In light of her
decision to set a new 1-hour standard that reflects the maximum
allowable NO2 concentration anywhere in an area (see section II.F.4.d of
the final rule), she concludes (consistent with the advice of CASAC)
that an appropriate consideration with regard to form is the extent to
which specific statistics could be unstable in locations where maximum
NO2 concentrations are expected, such as near major roads.  When
considering alternative forms for the standard, the Administrator notes
that an unstable form could result in areas shifting in and out of
attainment, potentially disrupting ongoing air quality planning without
achieving public health goals.  Given the limited available information
on the variability in peak NO2 concentrations near important sources of
NO2 such as major roadways, and given the recommendation from CASAC that
the potential for instability in the 99th percentile concentration is
cause for supporting a 98th percentile form, the Administrator judges it
appropriate to set the form based on the 3-year average of the 98th
percentile of the annual distribution of 1-hour daily maximum NO2
concentrations.  This decision is discussed in section II.F.3.c of the
final rule.  

(3)	Comment:  Some commenters stated a preference for expressing the
form in terms of the nth highest concentration rather than in terms of a
percentile.  These commenters believe that using the nth highest form
would increase transparency (e.g., Dow, AQRL).  AQRL noted that “it
can provide an unambiguous determination of a ‘design value’ in the
face of missing data if concentrations are high.  The percentile form
does not provide that certainty if data are missing.”

Response:  We disagree that a percentile form does not provide certainty
regarding the determination of the design value.  As discussed above,
and in the final rule (section II.F.3), the Administrator has determined
it appropriate to set a form based on the 3-year average of the 98th
percentile of the annual distribution of 1-hour daily maximum NO2
concentrations.  The interpretation of this standard, including
requirements for when data are missing, is described in Appendix S to 40
CFR part 50 and in section IV of the preamble to the final rule.  

Approach and level

Comments on the approach to setting the standard 

(1)	Comment:  We received a number of comments on the most appropriate
approach to setting the 1-hour NO2 standard.  In their comments on the
proposal, CASAC was split regarding the most appropriate approach to
setting the 1-hour standard.  In their letter to the Administrator on
the proposal (Samet, 2009b), CASAC stated the following: 

There was a split view on the two approaches among both CASAC and CASAC
panel members with a majority of each favoring the Agency's proposed
two-tiered monitoring network because they thought this approach would
be more effective in limiting near-roadway exposures that may reach
levels in the range at which some individuals with asthma may be
adversely affected. Other members acknowledged the need for research and
development of near-road monitoring data for criteria pollutants in
general but favored retention of EPA's current area-wide monitoring for
NO2 regulatory purposes, due to the lack of epidemiological data based
on near-roadway exposure measurements and issues related to implementing
a near-road monitoring system for NO2.

As indicated in their letter, the majority of CASAC Panel members
favored the proposed approach of setting a 1-hour standard that reflects
the maximum allowable NO2 concentration anywhere in an area and linking
such a standard with a 2-tiered monitoring network that would include
both near-road and area-wide monitors.  The recommendation of these
CASAC Panel members was based on their conclusion that the proposed
approach would be more effective than the alternative at limiting
near-roadway exposures to NO2 concentrations that could adversely affect
asthmatics.  

In contrast, the minority of CASAC Panel members expressed support for
the alternative approach of setting a 1-hour standard that reflects the
allowable area-wide NO2 concentration.  These CASAC Panel members
concluded that there would be important uncertainties associated with
the proposed approach.  Specifically, they noted that the key U.S. NO2
epidemiologic studies relied upon area-wide NO2 concentrations and that
this introduces uncertainty into the use of these studies to inform a
standard that reflects the maximum allowable NO2 concentration anywhere
in an area.  As a result of this uncertainty, CASAC Panel members who
favored the alternative approach noted that “it would be better to set
the standard on the same area-wide monitoring basis as employed in the
epidemiologic studies upon which it now relies” (Samet, 2009b).  These
CASAC Panel members also noted uncertainties associated with identifying
appropriate monitoring sites near roads (see below and section III.B.2
of the final notice for more discussion of monitoring comments).  

Consistent with the views expressed by the majority of CASAC members, a
number of commenters concluded that the most appropriate approach would
be to set a 1-hour standard that reflects the maximum allowable NO2
concentration anywhere in an area and to couple that standard with a
requirement that monitors be placed in locations where maximum
concentrations are expected, including near major roads.  This view was
expressed by some State and local agencies (e.g., in CA, IA, NY, TX, WA,
WI), by the majority of environmental organizations (e.g., CAC, EDF, EJ,
GASP, NRDC), by the ALA, and by a number of individual commenters. 
Several additional medical and public health organizations (ACCP, AMA,
ATS, NADRC, NACPR) did not explicitly express a recommendation regarding
the approach though these organizations did recommend that, in setting a
1-hour standard, particular attention should be paid to NOX
concentrations around major roadways.  In support of their
recommendation to adopt the proposed approach and to focus monitoring
around major roads, these commenters generally concluded that a primary
consideration should be the extent to which the NO2 NAAQS protects
at-risk populations that live and/or attend school near important
sources of NO2 such as major roads.  As such, these comments were
typically consistent with the rationale discussed by the Administrator
in the proposal in support of setting a 1-hour standard that reflects
the maximum allowable NO2 concentration anywhere in an area.  

	

Consistent with the view of the minority of CASAC Panel members, several
industry groups and an individual commenter (Roger McClellan) noted
uncertainties associated with using epidemiologic studies to inform
decisions on a standard reflecting the maximum allowable NO2
concentration anywhere in an area.  Specifically, Roger McClellan stated
the following: 

In considering the results of the six epidemiological studies that are
at the core of the EPA assessment, it is important to recognize all of
them used NO2 concentration data from central area monitors as a
surrogate for personal exposure. These data are not applicable to
establishing standards for near roadway monitors. 

Similar concerns with the proposed approach were expressed by several
industry groups.  

Response:  As described in detail in section II.F.4.c of the final
notice, the Administrator agrees with the majority of CASAC Panel
members and other commenters who concluded that the most appropriate
approach to setting the 1-hour NO2 standard is to set a standard
reflecting the maximum allowable NO2 concentration anywhere in an area. 
Her consideration of the uncertainties associated with this approach, as
stated by the minority of CASAC Panel members and industry commenters,
is also discussed in the final notice (sections II.F.4.b and II.F.4.c). 


(2)	Comment:  Some industry groups concluded that EPA is inappropriately
combining decisions on the standard with decisions on the monitoring
network.  Specifically, AAM states the following with regard to the
proposed approach 

[T]he effective stringency of a NAAQS depends on more than simply the
numerical level selected.  It also depends on the sensitivity of the
monitoring system, and the ability of that system to determine
accurately the ambient levels of the pollutant in a consistent manner at
monitoring stations across the country.  By affixing the stringency of a
primary NAAQS to a monitoring system that does not now exist and will be
subject to varying interpretation by regional, state and local
officials, EPA is in effect delegating the final decision on the NAAQS
standard to those officials, contrary to statutory requirements that
judgments concerning such standards reside with the Administrator.  

Response:  We disagree with this characterization of the standard.  As
discussed in detail in the final notice (section II.F.4.c), the
Administrator is setting a 1-hour standard reflecting the maximum
allowable NO2 concentration anywhere in an area.  It follows from this
standard that monitors should be placed in locations where maximum NO2
concentrations are expected to occur.  As discussed in the final notice
(section II.A.2 and II.F.4.c), available information supports the
conclusion that these maximum concentrations will occur around major
roads in many areas.  Therefore, the Administrator is requiring monitors
within 50 m of major roads (see section III of the final notice).  In
situations where maximum NO2 concentrations are expected to occur in
locations other than near major roads (e.g., near multiple smaller roads
and/or stationary sources), the Regional Administrator can also require
monitors in these locations.  EPA establishes criteria for the method of
monitoring and for quality assurance/quality control (QA/QC).  Any
hourly data from monitors meeting FRM/FEM and QA/QC requirements will be
considered in making designations. 

(3)	Comment:  Several industry and state commenters recommended that EPA
select its alternate monitoring proposal of using area-wide monitors to
be used in conjunction with a new short-term NO2 standard (e.g., EMA,
SCCC, NMED).  In their recommendation of the alternate monitoring
network, NMED noted that the use of area-wide monitors is more
applicable for measuring ambient air concentrations than near-road
monitors since few people live or work as close to major roadways as
near-road monitoring stations would be required to be located.  Other
commenters (e.g., Exxon) also concluded that monitor siting requirements
should allow for consideration of population.  

Response:  As noted in the proposal and the final notice (section
II.F.4.c), millions of people in the United States live, work and/or
attend school near important sources of NO2, including major roads.  In
addition, people commuting on major roads can be exposed to elevated NO2
concentrations. As described in the preamble, the Administrator
concludes that short-term elevated NO2 concentrations, including those
occurring on or near roads, pose public health risks.  The Administrator
further concludes it is appropriate to locate monitors where maximum
concentrations are expected to occur, and that generally it is
appropriate to locate monitors near major roads.  States are to consider
population in the site selection process if a state identifies multiple
acceptable candidate sites where maximum hourly NO2 concentrations are
expected to occur. 

   

(4)	Comment:  AQRL commented that EPA’s alternative monitoring
approach should not be chosen and implemented if one of the principal
reasons for the proposed 1-hr NO2 standard is to develop the near-road
monitoring network. 

Response:  The rationale supporting a 1-hour standard reflecting the
maximum allowable NO2 concentration anywhere in an area is discussed in
detail in section II.F.4.d.  This rationale does not include the
development of a near-road monitoring network.  

(5)	Comment:  Several commenters (e.g., AAM, API) concluded that it is
not appropriate to use the information discussed in the proposal on the
NO2 concentration gradient around roads to inform decisions on the
1-hour standard.  AAM concluded that the focus of the proposed approach
on NO2 concentrations around major roadways is not justified (and would
result in a standard that is more stringent than necessary) because the
REA and the proposal overstate the extent to which NO2 concentrations
near roads are higher than NO2 concentrations farther away from the
road.  AAM used data from the existing NO2 monitoring network as the
basis for their conclusion that “roadside monitors are not measuring
high NO2 concentrations.”  Specifically, they claim that their
analyses suggest the following: 

Many of the presently located urban monitoring sites are already located
within 50 meters of a major roadway. 

These monitors are not measuring the high concentrations that are of
concern to EPA. 

Consequently, there does not appear to be a need to initiate a massive
deployment of new NO2 monitors within 50 m of major roadways. 

Other commenters also noted that the highest concentrations of NO2 may
be in urban canyons, near stationary sources, or near airports (e.g.,
NESCAUM).    

Response:  With regard to the general point made by several commenters
that maximum NO2 concentrations may not always occur near major roads,
we note that, although the Administrator concluded in the proposal that
maximum NO2 concentrations in many areas are likely to occur around
major roads, she also allowed for situations where this is not the case.
 Specifically, she proposed to set a 1-hour NO2 standard that reflects
the maximum allowable NO2 concentration anywhere in an area, regardless
of where that maximum concentration occurs.  Therefore, the proposed
approach to setting the standard would be expected to limit the maximum
NO2 concentrations anywhere in an area even if in some areas, as is
contended by AAM, those maximum NO2 concentrations do not occur near
roads.

With regard to the analyses submitted by AAM, we agree that there is
uncertainty associated with estimates of roadway-associated NO2
concentrations (see REA, sections 7.4.6 and 8.4.8.3 for detailed
discussion of these uncertainties) and in identifying locations where
maximum concentrations are expected to occur.  However, we note that the
Administrator’s conclusions regarding the relationship between NO2
concentrations near roads and those away from roads rely on multiple
lines of scientific evidence and information.  Specifically, the
administrator relied in the proposal on the following in drawing
conclusions regarding the distribution of NO2 concentrations across
areas:

Monitoring studies discussed in the ISA and REA that were designed to
characterize the NO2 concentration gradient around roads, which
indicated that NO2 concentrations near roads are elevated compared to
concentrations in the same area but away from the road 

Air quality and exposure analyses presented in the REA which estimate
that NO2 concentrations on roads could be 80% higher than away from the
road, on average across locations, and that roadway-associated exposures
account for the majority of exposures to NO2 concentrations at or above
100 ppb

In contrast, to support their conclusions about the NO2 concentration
gradient, AAM relies largely on an analysis of existing NO2 monitors
that focused on 6 locations with a total of 42 monitors.  While this
analysis does provide information on NO2 concentrations at different
distances from roads, the existing NO2 monitoring network was not
designed to characterize the spatial gradients in NO2 concentrations
surrounding roadways.  Rather, concentrations of NO2 measured by
existing monitors may likely reflect contributions from a combination of
mobile and stationary sources, with one or the other dominating
depending on the proximity of the monitors to these sources.  

Specifically, one example used by the commenters to support their
conclusion that NO2 concentrations near roads are not elevated is that
of St. Louis, where the monitors closer to a road measured lower NO2
concentrations than a monitor farther away from the road.  However, a
closer look at the monitor classified as being away from the road
provides a potential explanation for these observations and illustrates
the limitations of relying exclusively on the existing monitoring
network to characterize the NO2 concentration gradient around roads. 
The monitor used (ID 295100086) is characterized by EPA’s Air Quality
System (AQS) as having a “high concentration” monitoring objective,
is 133 m from a major road, and is surrounded by 35 NOX emission sources
within 10 km having median emissions of about 17 tons per year (see REA,
Appendix A, Tables A-7 and A-8).  Thus, the single site used in St.
Louis for the ≥ 100 m bin may not necessarily reflect concentrations
at a site distant from a roadway, is likely not exclusively impacted by
roadway emissions, and would likely not be useful in determining the
relationship between near-road NO2 concentrations and those away from
the road.  The REA identified and discussed situations such as this as
an important uncertainty when estimating on-road concentrations (REA,
section 7.4.6, page 135).

Given the considerations above, we conclude that the analysis submitted
by AAM, which does not consider relevant lines of evidence and
information other than their analysis of existing monitors, does not
appropriately characterize the relationship between NO2 concentrations
near roads and those away from roads.  

Comments on standard level 

Comment:  In commenting on the proposal, CASAC discussed both the
proposed range of standard levels (i.e., 80-100 ppb) and the alternative
range of standard levels (i.e., 50-75 ppb).  Though, as discussed above,
they were split on which approach should be adopted for setting the
1-hour standard, CASAC did express the consensus conclusion that, if the
Agency finalizes a 1-hour standard in accordance with the proposed
approach (i.e., standard level reflects the maximum allowable NO2
concentration anywhere in an area and a monitoring network is
established that includes monitors near major roads), then it is
appropriate to consider the proposed range of standard levels from 80 to
100 ppb.  Specifically, the CASAC letter to the Administrator on the
proposal (Samet, 2009b) stated the following with regard to the proposed
approach:  

[T]he level of the one-hour NO2 standard should be within the range of
80-100 ppb and not above 100 ppb. In its letter of December 2, 2008,
CASAC strongly voiced a consensus view that the upper end of the range
should not exceed 100 ppb, based on evidence of risk at that
concentration. The lower limit of 80 ppb was viewed as reasonable by
CASAC; selection of a value lower than 80ppb would represent a policy
judgment based on uncertainty and the degree of public health protection
sought, given the limited health-based evidence at concentrations below
100 ppb. 

A number of State and local agencies and organizations also expressed
support for setting the level of the 1-hour NO2 standard within the
proposed range of 80 to 100 ppb, though only a few of these State and
local agencies (e.g., in CA, IA, MI, NY, TX) made this recommendation in
conjunction with a recommendation to focus monitoring near major roads
and other important sources of NO2.  

Response:  As is discussed in section II.F.4.d of the final rule, the
Administrator has judged it appropriate to set a standard level of 100
ppb.  Her rationale, which incorporates her consideration of the above
comments, is discussed in detail in this section.  

(2)	Comment:  A number of environmental organizations (e.g., CAC, CSE,
EDF, EJ, GASP, NRDC) and medical/public health organizations (e.g.,
ACCP, ALA, AMA, ATS, NACPR, NAMDRC) supported setting a standard level
below 80 ppb for a standard that reflects the maximum allowable NO2
concentration anywhere in an area.  Several of these groups recommended
a standard level of 50 ppb.  This recommendation was typically based on
the commenters’ interpretation of the epidemiologic and controlled
human exposure evidence as well as the NO2 exposure and risk
information.  

Some of these commenters noted that the 98th percentile area-wide NO2
concentration was below 80 ppb in the location of a single key U.S.
epidemiologic study (i.e., 50 ppb in study by Delfino).  Given this,
commenters concluded that the standard level should be set at 50 ppb. 
Their comments on the monitoring network generally favored a requirement
to place monitors near major roads and, therefore, these commenters
appeared to favor a standard level as low as 50 ppb and to recommend
that such a standard level reflect the maximum allowable NO2
concentration anywhere in an area.  In their comments, the ALA, EDF, EF,
and NRDC stated the following: 

Considering the Delfino study alone on EPA’s terms, that is, focusing
on the 98th percentile of the 1-hour daily maximum concentrations, EPA
reports a concentration of 50 ppb where asthma symptoms were observed.
Based primarily on this study, EPA concluded in the REA that it was
appropriate to set the lower end of the range at 50 ppb, which
corresponded to the lowest-observed effects level of airway
hyperresponsiveness in asthmatics. To provide the strongest public
health protection, we therefore urge the level of the standard be set at
50 ppb. 

In some cases, the same commenters also appeared to recommend setting a
standard level below 50 ppb because mean area-wide NO2 concentrations
reported in locations of key U.S. epidemiologic studies are below this
concentration.  Specifically, with regard to the key U.S. epidemiologic
studies, these commenters (e.g., ALA, EDF, EJ, NRDC) stated the
following:

These studies clearly identify adverse health effects such as emergency
room visits and hospital admissions for respiratory causes at
concentrations currently occurring in the U.S. Mean concentrations for
all but two of these studies are about or below 50 ppb, suggesting that
the standard must be set below this level to allow for a margin of
safety. 

Some of these commenters also concluded that the proposed range of
standard levels fails to provide a margin of safety for effects reported
in controlled human exposure studies.  

Response:  The Administrator’s decision to set a standard level of 100
ppb is discussed in detail in section II.F.4.d of the final rule.  The
rationale for this decision includes consideration of the above
comments.  Specific comments on standard level are discussed in section
II.F.4.c of the final rule.  

(3)	Comment:  Several environmental and public health groups (e.g., ALA,
EDF, EJ, NRDC) concluded that the REA analyses of NO2 exposures and
risks suggest that “only an hourly standard of no more than 50 ppb
would protect against harm from peak exposures.”  

	Response:  We disagree with this characterization.  As discussed in
sections II.C and II.F.4.d of the final rule, the analyses of different
potential alternative standards in the REA are based on the current
monitoring network, which contains primarily area-wide monitors.  The
Administrator concludes that the results of exposure and risk analyses
provide support for limiting area-wide NO2 concentrations to no higher
than 100 ppb.  The Administrator is setting a new 1-hour standard that
reflects the maximum allowable NO2 concentration anywhere in an area. 
Therefore, when considering the REA analyses in the context of such a
standard, the Administrator considered available information on the
relationship between the higher NO2 concentrations around roads and
area-wide NO2 concentrations.  Specifically, as described in detail in
section II.F.4.d, she noted that a standard level of 100 ppb reflecting
the maximum allowable NO2 concentration anywhere in an area would be
expected to limit area-wide NO2 concentrations to approximately 50 to 75
ppb.  Therefore, she concluded that a standard level of 100 ppb, for a
standard that reflects the maximum allowable NO2 concentration anywhere
in an area, is consistent with conclusions based on the NO2 exposure and
risk information.  

(3)	Comment:  Several industry groups (e.g., AAM, Dow, NAM, NPRA) and
Roger McClellan concluded that, if EPA does choose to set a new 1-hour
standard, the level of that standard should be above 100 ppb.  As a
basis for this recommendation, these groups typically emphasized
uncertainties in the scientific evidence.  Specifically, these
commenters typically concluded that available epidemiologic studies do
not support the conclusion that NO2 causes reported health effects. 
This was based on their assertion that the presence of co-pollutants in
the ambient air precludes the identification of a specific NO2
contribution to reported effects.  Some industry groups also concluded
that epidemiologic studies should not be used to inform decisions on the
standard level because these studies were based on area-wide, rather
than near-road, NO2 concentrations.  As a result, these commenters
recommended that a 1-hour standard should be based on the controlled
human exposure evidence and that, in considering that evidence, EPA
should rely on the meta-analysis of NO2 airway responsiveness studies
conducted by Goodman et al., (2009) rather than the meta-analysis
included in the final ISA.  They concluded that in relying on the ISA
meta-analysis, EPA has inappropriately relied on a new unpublished
meta-analysis that has not been peer-reviewed, was not reviewed by
CASAC, and was not conducted in a transparent manner.  

Response:  EPA recognizes the uncertainties in the scientific evidence
that are discussed by these industry commenters; however, we strongly
disagree with their conclusions regarding the implications of these
uncertainties for decisions on the NO2 NAAQS.  These comments, and
EPA’s responses, are discussed in detail in section II.E.2 of the
final rule and above (section II) in this Response to Comments Document.
 The Administrator’s conclusions on standard level are described in
section II.F.4.d of the final rule.  

(4)	Comment:  Some commenters argued that EPA should not adjust the
standard level to account for postulated near-road concentrations
because, due to differences in concentration gradient, it would
establish a level that is arbitrary and variable across locations (e.g.,
NPRA, NYSDEC, AAM).  AAM further argued that EPA’s  information about
the concentration gradient was not sufficiently precise or “factual”
to inform EPA’s decision making. 

Response:  As described in section II.F.4.d of the final rule, EPA has
not adjusted the standard level to account for near-road concentrations.
 Rather, the Administrator has considered the extent to which specific
standard levels would be expected to protect against exposure to the
distribution of NO2 concentrations across an area and, therefore,
against the array of respiratory effects that have been linked to
short-term NO2 exposures.  These considerations are described in detail
in section II.F.4.d of the final rule.  EPA disagrees with AAM’s
suggestion that information about the concentration gradient should be
disregarded in reviewing the standard.

(5)	Comment:  One commenter (NYDOH) concluded that too much weight has
been placed on clinical studies and that more emphasis should be placed
on epidemiologic studies in setting the standard level.  

Response:  As discussed in detail in section II.F.4.d of the final rule,
the Administrator has considered both the controlled human exposure and
epidemiologic evidence in setting the standard level.  Both lines of
evidence provide information on the extent to which specific NO2
standards would be expected to protect against the range of respiratory
effects that have been linked to short-term exposures to NO2.    

(6)	Comment:  INGAA pointed out that the related health-based standards
(e.g., Short-term Exposure Level (STEL) or ceiling Permissible Exposure
Level (PEL) from Occupational Safety and Health Association (OSHA) or
the National Institute of Occupational Safety and Health (NIOSH)) are
higher than the proposed NO2 NAAQS.  They claimed that personnel have
not been negatively affected by NO2 exposures at their facility.  The
commenter also observed that “when comparing the relative margin of
OSHA or NIOSH standards to the NAAQS for other gases, the proposed NO2
NAAQS is more stringent.  For example, the OSHA NO2 PEL (ceiling) is 5
ppmv [parts per million by volume] (i.e., 5000 ppbv [parts per billion
by volume]) and the NIOSH STEL is 1 ppmv (1000 ppbv).  These levels are
one to two orders of magnitude higher than ranges being considered by
EPA for the 1-hour NO2 NAAQS.  By comparison, the CO [carbon monoxide]
NAAQS 1-hour average standard is 35 ppmv, compared to a NIOSH STEL of
200 ppmv, a factor of 5.7.  This appears to indicate a more conservative
approach for the proposed NO2 NAAQS.”

Response:  The CAA requires that air quality criteria and NAAQS be
reviewed periodically.  After consideration of the latest scientific
knowledge, the NAAQS must be set at levels that are, in the
Administrator’s judgment, requisite to protect public health,
including the health of sensitive subpopulations, with an adequate
margin of safety.  The Administrator has considered this requirement in
setting a new 1-hour NO2 standard as described in section II.F.4.  In
contrast, occupational standards, such as those mentioned by the
commenter, are established based on different legal requirements and may
not be updated in the same manner as the NAAQS.  Therefore, there is no
expectation that the ratio of the NAAQS to an occupational standard
would  be a basis for selecting a NAAQS.  

(7)	Comment: NESCAUM argues that “while ozone rapidly oxidizes NO to
NO2, there is usually insufficient ozone to produce substantially
elevated levels of NO2 from this reaction in near-roadway exposure
scenarios during morning rush hours, nighttime, and at least half of the
year (during the non-ozone season).”

Response:  While EPA acknowledges that the extent to which NO2
concentrations near roads are elevated will vary with ozone
concentrations, we disagree with the conclusion that NO2 concentrations
near roads are not usually elevated substantially.  As discussed in the
final rule (sections II.A.2 and II.F.4), elevated NO2 concentrations do
occur on and around roads and people who live, work, and/or attend
school near major roads, or commute in vehicles on major roads, can be
exposed to these elevated concentrations.  For example, based on NO2
monitoring studies the ISA stated that NO2 concentrations in heavy
traffic or on freeways “can be twice the residential outdoor or
residential/arterial road level,” that “exposure in traffic can
dominate personal exposure to NO2,” and that “NO2 levels are
strongly associated with distance from major roads (i.e., the closer to
a major road, the higher the NO2 concentration)” (ISA, sections 2.5.4,
4.3.6).  In addition, a considerable fraction of the population resides,
works, or attends school near major roadways and these populations are
likely to have increased exposure to NO2 (ISA, section 4.4).  Based on
data from the 2003 American Housing Survey, approximately 36 million
individuals live within 300 feet (~90 meters) of a four-lane highway,
railroad, or airport (ISA, section 4.4).  Furthermore, in California,
2.3% of schools with a total enrollment of more than 150,000 students
were located within approximately 500 feet of high-traffic roads (ISA,
section 4.4).  

(8)	Comment:  Harris County Public Health and Environmental Services
(HCPHES) noted that EPA’s REA “concludes that it is appropriate to
focus on studies that evaluated NO2 health effect associations using
both single- and multi-pollutant models.  However, HCPHES believes that
insufficient study and attention has been given to such synergistic
effects.  Because of this insufficiency, HCPHES believes that EPA should
include an adequate margin of safety and set the one-hour standard
sufficiently below the upper limit of 100 ppb leaning more towards the
bottom end of the proposed range.”

Response:  The Administrator’s decision on standard level, including
her consideration of the requirement to provide an adequate margin of
safety, is discussed in detail in section II.F.4.d of the final rule.  

(9)	Comment:  AASHTO recommended that EPA should not set its NO2
standard below background concentrations of NO2 in order to avoid making
the standard unattainable. 

	Response:  As described in detail in section II.F.4.d of the final
rule, the Administrator has set an NO2 NAAQS that, in her judgment, is
requisite to protect public health, including the health of sensitive
subpopulations, with an adequate margin of safety.  In setting such a
standard, the Administrator is prohibited from considering such factors
as projections of areas that will or will not attain the standard. 
However, EPA notes that we are unaware of any locations with
nonanthropogenic background concentrations that would exceed the
standard. 

Comments on the annual standard   

(1)	Comment:  With regard to an annual standard, CASAC and a number of
public commenters (e.g., NACAA, NESCAUM; agencies from States including
CA, IN, MO, NC, NY, SC, TX, VA; tribal organizations including Fon du
Lac and NTAO; environmental/medical/public health groups including ACCP,
ALA, AMA, ATS, CAC, EDF, EJ, GASP, NACPR, NAMDRC, NRDC; and industry
groups including AAM, API, Dow, INGAA, UARG) agreed with the proposed
decision to maintain an annual standard, though their recommendations
with regard to the level of that annual standard differed (see below).  

In their comments on the final REA, CASAC recommended “retaining the
current standard based on the annual average” based on the “limited
evidence related to potential long-term effects of NO2 exposure and the
lack of strong evidence of no effect” and that “the findings of the
REA do not provide assurance that a short-term standard based on the
one-hour maximum will necessarily protect the population from long-term
exposures at levels potentially leading to adverse health effects”
(Samet, 2008).  A number of State agencies and organizations and
industry groups also recommended maintaining the current level of the
annual standard (i.e., 53 ppb).  This recommendation was based on the
conclusion that, while some evidence supports a link between long-term
NO2 exposures and adverse respiratory effects, that evidence is not
sufficient to support a standard level either higher or lower than the
current level.  

In contrast, some environmental organizations and medical/public health
organizations as well as a small number of States (e.g., ALA, EDF, EJ,
NRDC, and organizations in CA) recommended setting a lower level for the
annual standard.  These commenters generally supported their
recommendation by pointing to the State of California’s annual
standard of 30 ppb and to studies where long-term ambient NO2
concentrations have been associated with adverse respiratory effects
such as impairments in lung function growth.  

Response:  EPA’s response to these comments, and the Administrator’s
conclusions on the annual standard, are discussed in section II.G of the
final rule.  

(2)	Comment: Several petroleum industry commenters (API, UPA, IPAMS,
PAW, UARG) claimed that EPA did not provide a rationale for changing the
units of the annual standard.  These commenters assert that the proposal
to express the level as 53 ppb is different from the current standard
which also expresses the level in terms of micrograms/m3 (100
micrograms/m3).  These commenters claim that by shifting to ppb, the
stringency of the standard is increased for higher elevations (i.e., at
sea level 53 ppb = 100 microgram/m3, but at 5,000 ft, 53 ppb = 84
micrograms/m3).  PAW specifically stated that “the change in the
annual standard from the current mass and volume concentration to a
strictly volume concentration represents a significant and unjustified
tightening of the standard for high altitude areas, which includes many
of the areas in which the oil & gas industry operates in the western
states.” 

Response:  The issue raised in the comments is attributed to the
mathematical conversion between ppb (a mixing ratio, by volume) and
micrograms per cubic meter (µg/m3), which is a mass-based metric. 
According to the ideal gas law (PV=nRT) the number of molecules (n) are
proportional to their volumes (V) which in turn is proportional to their
pressures (P).  Since the total air pressure, and hence the
concentration of air, decreases with altitude, a constant mixing ratio
will not translate into a constant concentration.  Typically, when
converting from ppb to µg/m3, scientists standardize the process by
using a standard pressure and temperature, e.g., 1 atm and 25 C,
respectively, in the conversion calculations.  This standardization
results in the apparent difference in mass concentration provided as an
example by the commenter. 

EPA notes that “parts per” concentration, such as parts per billion
(ppb) or parts per million (ppm), etc., are mixing ratios, and are very
common units for measuring gas phase species.  Further, the use of
mixing ratios is widespread for expressing the relative amounts of
species at various altitudes.  The chemiluminescence FRM for the NO2
NAAQS provides in-situ concentration information, in a mixing ratio by
volume (e.g., ppb), that is compensated for both temperature and
pressure.  As a result, chemiluminescence FRM-produced data, which
measures concentrations from a known volume of air, is not subject to
variation due to ambient temperature or pressure (e.g., altitude). 

The parenthetical in the existing standard was intended to be
informative, indicating the mass-based measurement corresponding to the
level of the standard at standard temperature and pressure.  The
parenthetical did not create a second standard, allowing compliance with
either measurement depending on ambient temperature or pressure, and
therefore deletion of the parenthetical is not a change to the standard.
 Since the FRM is not subject to variation due to ambient temperature or
pressure, the ppb level is the only measurement relevant to attainment.
However, EPA recognized that the presence of the parenthetical in the
standard could be confusing, and determined it was appropriate to delete
it to clarify the standard. 

Technical Issues with Monitoring Requirements

Comments on near-road monitor siting requirements 

(1)	Comment:  Some state agencies (e.g., AASHTO, NYSDOT, and WIDNR)
recommended that the maximum horizontal distance be increased from the
proposed 50 meters to as much as 200 meters, noting that a 50 meter
limit would be infeasible for many locations and could create a serious
safety hazard for monitoring staff and the general public.  In contrast,
several environmental, and industry groups (e.g., GASP and AQRL)
supported the proposed range, and in GASP’s case, supported reducing
the maximum distance to 30 meters or less in order to reduce the
potential variation in roadway distance between monitoring sites and
subsequent measurements of near-road NO2 concentrations. 

Response:  EPA’s decision on the siting criteria for NO2 monitors
requires, among others, that the near-road NO2 monitors be sited within
50 meters from the roadway.  EPA discusses the basis for this decision
in detail in section III.B.7 of the preamble to the final rule. 

(2)	Comment:  A limited number of comments were received recommending
changes to the requirements for the vertical height of the NO2 monitor
inlet manifold.  Two state organizations (NESCAUM and NYSDEC) noted that
the proposed 2-7 meter vertical range could generate significant
variation in the concentrations reported from monitors located at
different heights.  NESCAUM recommended that the height range be reduced
to 2.5-3.5 meters.  NYSDEC commented that the proposed vertical height
range may not be practical for many sites, especially those located in
urban areas.  In addition, the ALA, EJ, EDF, and NRDC noted that: 

[T]he lower end of the proposed height of 2 to 7 meters appears to
capture the highest NO2 concentrations, and more accurately represents
human exposure at the breathing zone.  Additional monitors at other
relationships to these sources may be needed for research purposes, but
they should be in addition to those designed to establish the peak
exposures for NAAQS compliance purposes. 

Response:  EPA’s decision on the siting criteria for NO2 monitors
requires, among others, that the near-road NO2 monitors inlet manifolds
be placed between 2 to 7 meters above the ground.  EPA discusses the
basis for this decision in detail in section III.B.7 of the preamble to
the final rule. With regard to the ALA, EJ, EDF, and NRDC comment on
needing additional monitors for research, EPA does not believe that this
particular research issue is appropriate to address through this
rulemaking, but rather through ongoing and future research that further
characterizes the near-road environment.  For the purposes of supporting
the revised 1-hour NAAQS, EPA believes that the siting criteria are
sufficient and appropriate. 

(3)	Comment:  Several groups that support the proposed two-tiered
monitoring network requested modifications to the proposed criteria for
determining where a near-road monitor should be located.   Some
environmental/public health organizations (ALA, EJ, NRDC, EDF) commented
that “Near-road monitor placement should be determined not only by the
highest AADT [annual average daily traffic] volumes in a given
[core-based statistical area] CBSA, but also by the highest heavy-duty
truck volumes.”  The NYDOH, IADNR, and some industry groups (e.g., EEI
and SRNS) recommend making population a prominent consideration when
deciding where to locate monitors.  HCPHES recommended that near-road
monitoring stations should take into consideration the location of other
major mobile and point sources for NO2 emissions, such as airports,
seaports, and power plants, or vulnerable population groups like
children, when determining the appropriate locations for near-road
monitors.  

Response:  EPA has clarified that the selection criteria should include
the consideration of localized factors when identifying locations of
expected maximum concentrations.  Near-road sites shall be selected by
ranking all road segments within a CBSA by AADT and then identifying a
location or locations adjacent to those highest ranked road segments,
considering fleet mix, roadway design, congestion patterns, terrain, and
meteorology, where maximum hourly NO2 concentrations are expected to be
highest and the siting criteria can be met in accordance with 40 CFR
Part 58 Appendix E.  EPA discusses the basis for this decision in detail
in section III.B.6 of the preamble to the final rule.  

(4)	Comment:  Many state agencies commented that EPA needed to provide
greater guidance to state and municipal agencies about where to site
near-road monitors.  The CAC recommended that state discretion over the
location of near-road monitors be limited “such that gaming is not
tolerated.”

Response:  EPA agrees that guidance will be useful in aiding the
implementation of a near-road network.  As noted in the section III.B.5
of the preamble to the final rule, EPA plans to assist state agencies in
the network implementation process, particularly through guidance
documentation that the Agency intends to prepare and share with state
and local agencies early enough in the overall implementation process to
be useful to states in developing their NO2 monitoring network. 
Further, EPA encourages state and local monitoring agencies to include
or cooperate with transportation officials in the development of the
near-road network. EPA believes that state and local monitoring agencies
may benefit greatly from working with federal, state, and/or local
transportation officials during the development of and long-term
maintenance of a near-road network.  

EPA believes that the commenter’s concern that the state discretion
allowed in the placement of near-road monitors may lead to installations
in locations unlikely to violate the NAAQS, will be avoided due to the
process outlined in 40 CFR §58.10 under which a state’s annual
monitoring network plan must be made available for public inspection or
comment and approved by the EPA Regional Administrator.  This process
will permit scrutiny of the monitor siting locations and assist in
identifying and siting monitors in appropriate locations within CBSAs.

(5)	Comment:  CASAC advised that siting monitors based on traffic counts
alone might miss locations where maximum NO2 concentrations would occur.
 CASAC recommends that EPA should use model results from Congestion
Mitigation and Air Quality Land Use Regression modeling as well as
Gaussian plume models and emissions inventories to determine where
roadside monitors should be located.  

Response:  EPA does not intend for AADT counts to be the sole basis for
choosing a near-road site.  EPA has clarified rule language to reflect
that required near-road NO2 monitoring stations shall be selected by
ranking all road segments within a CBSA by AADT and then identifying a
location or locations adjacent to those highest ranked road segments,
considering fleet mix, roadway design, congestion patterns, terrain, and
meteorology, where maximum hourly NO2 concentrations are expected to
occur and the siting criteria can be met in accordance with 40 CFR Part
58 Appendix E. This issue is discussed in section III.B.6 of the
preamble to the final rule.  EPA also notes that air quality models,
which were noted by the CASAC panel member to be considered for use in
near-road site selection, are tools that EPA believes will be useful,
and likely used by some states to inform where near-road sites need to
be placed

 

(6)	Comment:  EPA received a number of comments recommending using
different population thresholds, or a different approach in requiring
near-road monitors.  Several commenters (e.g., NESCAUM, NYSDOT, NYSDEC)
recommended using a higher population threshold than that proposed, for
example, NYSDEC suggested the population threshold be increased to as
much as 2.5 million.  Conversely, CAC suggested that the proposed
threshold may not be low enough.  Other commenters (e.g., AASHTO and
HCPHES) recommended that EPA use an AADT threshold (ranging from 100,000
to 250,000 AADT), in lieu of a population threshold, to require
near-road monitoring.  Another commenter, San Joaquin Air Pollution
Control District suggested the using a combination of population and
AADT thresholds to require near-road monitors.  

Response:  EPA’s decision is to require one near-road NO2 monitor in
CBSAs with a population greater than or equal to 500,000 persons and a
second near-road monitor is required in CBSAs with a population greater
than or equal to 2,500,000 persons, or in any CBSAs with one or more
road segments with an Annual Average Daily Traffic (AADT) count greater
than or equal to 250,000.  EPA discusses the basis for this decision in
detail in section III.B.2 of the preamble to the final rule. 

(7)	Comment:  EPA received comments suggesting that EPA require
monitoring at sites non- near-road locations.  One commenter (Center on
Race, Poverty, and the Environment) recommended that EPA account for
populations in rural areas, such as those impacted by agricultural
emissions.  The Fond du Lac Band of Lake Superior Chippewa and the NTAA
suggested that EPA expand monitoring coverage into rural areas and
tribal lands.  Other commenters (ATS, AMA, NAMDRC, NACPR, ACCP, and the
Swinomish Tribe) also recommended that EPA consider sources other than
mobile sources, such as large stationary sources.  

Response:  As discussed in section II.F.4.d of the final rule, the
Administrator has set a new 1-hour NO2 standard that reflects the
maximum allowable NO2 concentration anywhere in an area.  In her
judgment, this new standard will protect public health, including the
health of sensitive populations such as asthmatics, with an adequate
margin of safety.  It follows from this standard that monitors should be
placed in locations where maximum NO2 concentrations are expected.  The
Administrator has judged that, in many areas, these maximum
concentrations are likely to occur around major roads (final rule,
sections II.A.2 and II.F.4.d).  Therefore, she is requiring NO2 monitors
within 50 meters of major roads.  In situations where maximum NO2
concentrations are likely to occur in locations other than near a major
road (e.g., near stationary point or area sources, whether agricultural
or industrial (such as those noted by the commenters), the Regional
Administrator has the authority to require additional monitors in these
locations.  Further, the Administrator has also recognized that
susceptible and vulnerable populations, which include asthmatics and
disproportionately exposed groups, (as discussed in sections II.B.4 and
II.F.4.d of the preamble to the final rule) are at particular risk of
NO2-related health effects.  The Administrator is therefore requiring
the Regional Administrators, working in collaboration with states, to
site forty monitors in appropriate locations, focusing primarily on
protecting such susceptible and vulnerable communities.  This decision
is discussed in detail in section III.B.4.

(8)	Comment:  Two tribal commenters ( Fond Du Lac Band of Lake Superior
Chippewa and the National Tribal Association) recommended that more
monitoring stations be required for rural or tribal roadways since EPA
recognized Native Americans as an at-risk population group. 

Response:  As discussed in section II.F.4.d of the final rule, the
Administrator has set a new 1-hour NO2 standard that reflects the
maximum allowable NO2 concentration anywhere in an area.  In her
judgment, this new standard will protect public health, including the
health of sensitive populations such as asthmatics, with an adequate
margin of safety.  It follows from this standard that monitors should be
placed in locations where maximum NO2 concentrations are expected.  The
Administrator has judged that, in many areas, these maximum
concentrations are likely to occur around major roads (final rule,
sections II.A.2 and II.F.4.d).  Therefore, she is requiring NO2 monitors
within 50 meters of major roads.  In situations where maximum NO2
concentrations are likely to occur in locations other than near a major
road (e.g., near multiple smaller roads and/or stationary sources), the
Regional Administrator will have the authority to require additional
monitors in these locations.  EPA notes that it is not in a position to
require monitoring by tribes or on tribal land.  However, EPA has and
will continue to work with tribes in conducting ambient monitoring on
tribal lands.

(9)	Comment:  One state agency, NYSDEC requested that CBSAs with
populations larger than 2.5 million be allowed to have more than two
near-road monitors.

Response:  EPA’s decision is to require, at a minimum, one near-road
NO2 monitor in CBSAs with a population greater than or equal to 500,000
persons and a second near-road monitor in CBSAs with a population
greater than or equal to 2,500,000 persons, or in any CBSAs with one or
more road segments with an Annual Average Daily Traffic (AADT) count
greater than or equal to 250,000.  EPA discusses the basis for this
decision in detail in section III.B.2 of the preamble to the final rule.
EPA also notes that states always have the ability to conduct additional
monitoring above the minimum monitoring requirements at their
discretion.

(10)	Comment:  Some public health and environmental groups (e.g., ALA,
CAC, and NRDC) suggested that near-road monitors should be required to
be located on the downwind side of the target road.  Conversely, several
commenters (AQRL and NYSDEC) suggested that such a requirement may be
over restrictive and not necessary.  AQRL commented that siting monitors
in downwind locations would not be feasible for all locations and EPA
should allow upwind monitoring locations so long as upwind monitors are
used in conjunction with air dispersion modeling.

Response:  EPA is not going to finalize a requirement that near-road
sites must be climatologically downwind of the target road segment
because of the additional limitations this introduces to finding
potential site candidates in exchange for what may be a small increase
in the opportunity to monitor peak NO2 concentrations.  EPA discusses
the basis for this decision in detail in section III.B.7 of the preamble
to the final rule.

(11)	Comment:  Several commenters (e.g., CASAC, ACCP, ALA, AMS ATS, EDF,
NACPR, NAMDRC, NRDC) made comments that the proposed network should form
the basis for a broader near-road monitoring network that would
encompass other pollutants as well, i.e., be a multi-pollutant near-road
monitoring network.  These organizations’ recommendations were based
on the argument that it makes little sense to only monitor NO2
concentrations near roads when it is well-understood that many other
pollutants, hazardous to the public’s health, are emitted by on-road
vehicles along with NO2.

Response:  The scope of this rulemaking pertains to the NO2 NAAQS and
the monitoring associated with it.  However, EPA agrees that
multi-pollutant monitoring is desirable, and believes that the sites
required in the near-road component of the NO2 network design could also
be suitable for monitoring other pollutants including carbon monoxide,
particulate matter (especially ultra-fine particulate matter), air
toxics, and black carbon.

(12)	Comment:  Comments were submitted, particularly from state agencies
(e.g., Michigan, Mississippi, and Tennessee), noting that one of  the
challenges to developing a near-road monitoring network was the
potential safety hazard of having a monitoring station located within 50
meters of busy roadways.  

Response:  EPA notes that in all instances of field work, safety is a
top priority.  EPA believes that the safety issue raised in public
comments is something that is an important component of near-road siting
logistics, and that the near-road network can be safely implemented
while still serving its intended purpose of providing data on the
expected maximum NO2 concentrations that occur within 50 meters of
heavily trafficked roads. EPA discusses this issue in further detail in
section III.B.7 of the preamble to the final rule.

(13)	Comment:  The Clean Air Council (CAC) recommended the “EPA not
permit alternative monitoring systems on a state-by-state basis if they
would hinder uniformity to the extent that public understanding and
accountability of government would be degraded.”  In contrast, the
Southeast Michigan Council of Governments (SEMCOG) suggested that,
rather than requiring a near-road monitoring network, EPA should enable
states to select monitoring locations based on each state’s specific
needs and issues as part of the state implementation plan (SIP) process.


Response:  EPA believes that it must use, and has chosen, a balanced
approach in setting monitoring requirements that recognizes the need to
have as nationally consistent monitoring network as possible, while
allowing an appropriate amount of flexibility so that states are able to
comply with monitoring requirements. EPA believes that the near-road
site selection process must include certain considerations such as AADT,
fleet mix, roadway design, congestion patterns, terrain, and meteorology
to identify where maximum hourly NO2 concentrations are expected occur. 
While states have flexibility in evaluating what specific thresholds or
criteria are appropriate for their particular situations with respect to
the metrics listed above, they are to make all effort to meet the
primary objective of locating the site in a location of expected
concentration.  This flexibility is necessary when considering the
reality of, for example, safety or site access, and other logistical
issues that are inherent to installing a monitoring network that can
differ on a case-by-case basis. EPA discusses issues of site selection
and siting criteria in further detail in sections III.B.6 and III.B.7 of
the preamble to the final rule.

(14)	Comment:  The Spokane Regional Clean Air Authority (SRCAA) noted
that it might not be necessary for EPA to require an extensive near-road
monitoring network if a smaller network can develop a strong statistical
correlation between particular variables (i.e. traffic density, vehicle
mix) and above-standard NO2 concentrations.  

Response:  EPA believes that the currently available data, and the
number of variables that influence the concentration and behavior of NO2
in the near-road environment, including fleet mix, roadway design,
congestion patterns, terrain, and meteorology, do not allow the
development of a “strong,” and nationally applicable, statistical
correlation that would allow the avoidance of what the commenter states
is an “extensive” near-road network. (EPA estimates that
approximately 126 near road monitors will be needed.) The reasoning
behind the size and extent of the near-road monitors required in the
two-tier network design, with respect to SRCAA’s comments, are
discussed in detail in sections III.B.1 and III.B.2 of the preamble to
the final rule..

	 

(15)	Comment:  Several environmental and public health groups (e.g.,
ALA, EDF, EJ, NRDC) pointed out that “only 58 of 489 total NO2
monitors are sited in areas of expected peak concentrations.”

Response: The relative current state of the NO2 network is detailed in
the NOx Network Review and Background document (Watkins and Thompson,
2009), which is posted in the docket.  This document provides
information and summaries on what EPA believes the current monitoring
network, according to state provided meta-data, is addressing.

Comments on area-wide monitor requirements 

A number of comments were received, particularly from state agencies, in
support of maintaining the existing area-wide monitoring network. 
Comments specific to the siting or area-wide monitors tended to focus on
either (1) the need for more area-wide monitors than the proposed 52 or
(2) the need to make sure that all area-wide monitors were measuring
ambient NO2 concentrations rather than NO2 concentrations near major NO2
emissions sources.

Comment:  Several industry groups and state agencies (e.g., AAM, EMA,
INDEM, and NYSDOT) provided specific comments that they preferred the
alternative monitoring approach of using area-wide monitors to measure
1-hour NO2 concentrations due to the high cost of implementing,
operating, and maintaining a near-road monitoring network and the
uncertain benefits such a network would provide.  In addition, one
commenter recommended the area-wide approach because it provided greater
flexibility for monitors to be located in places where more people are
likely to be exposed to high concentrations of NO2.

Response:  The Administrator has set a new 1-hour NO2 standard that
reflects the maximum allowable NO2 concentration anywhere in an area. 
In her judgment, this new standard will protect public health, including
the health of sensitive populations such as asthmatics, with an adequate
margin of safety.  It follows from this standard that monitors should be
placed in locations where maximum NO2 concentrations are expected.  The
Administrator has judged that, in many areas, these maximum
concentrations are likely to occur around major roads (final rule,
sections II.A.2 and II.F.4.d).  Therefore, the Administrator has
concluded that a two-tier network design composed of (1) near-road
monitors which would be placed in locations of expected maximum 1-hour
NO2 concentrations near heavily trafficked roads in urban areas and (2)
monitors located to characterize areas with the highest expected NO2
concentrations at the neighborhood and larger spatial scales (also
referred to as “area-wide” monitors) are needed to implement the
1-hour NO2 NAAQS.  The rationale for this decision is discussed in
section III.B.1 of the preamble to the final rule.

Comment:  Several environmental, public health, and tribal organizations
(ALA, EDF, EJ, NRDC, NTAA) commented that EPA should require more
monitors than currently proposed.  Specifically, the ALA, EJ, NRDC, and
EDF stated that they “oppose the proposed requirement to retain only
52 air monitors to measure area wide concentrations NO2” In addition,
they commented: 

“[The] EPA should require states and local offices to review inventory
data to identify any potential NO2 hotspots outside of those large
metropolitan areas.  For instance, if a large power plant or any other
source is creating elevated NO2 levels in proximity to homes, schools or
other sensitive sites, in an area of less than one million people, EPA
should consider requiring a monitor.  In particular, certain large
agricultural facilities may emanate high concentrations of NO2 under
certain conditions, such as wet weather.  Many of these facilities are
directly upwind of rural communities, meriting an NO2 monitor.”  

Response:  EPA recognizes a variety of exposure scenarios can occur in
an area such as the ones described by the commenters above.  Therefore,
the final rule authorizes Regional Administrators to require additional
monitors above the minimum required number of monitors in circumstances
such as those described by the commenter above where there is a
likelihood of high concentrations of NO2 that approach or exceed the
NAAQS.  The size and extent of the required area-wide monitors in the
network design are discussed in section III.B.3 of the preamble to the
final rule.  A more detailed discussion of the Regional Administrator
authority and some examples where such authority may be exercised is in
section III.B.4 in the preamble to the final rule.  

Comment:  One state agency provided comments opposing using only an
area-wide network for monitoring NO2 on the grounds that the network
represents an inadequate alternative to the creation of a near-road
monitoring network.  The commenter was critical of how useful dispersion
modeling would be at identifying NO2 concentrations beyond a monitor’s
immediate area.  The commenter claimed there were too many confounding
factors to accurately estimate the NO2 concentrations near roadways and
other NO2 emissions sources from an area-wide monitor.

Response: The Administrator has set a new 1-hour NO2 standard that
reflects the maximum allowable NO2 concentration anywhere in an area. 
In her judgment, using only an area-wide network of NO2 monitors would
not adequately support this new standard because maximum concentrations
of NO2 are likely to occur around major roads (final rule, sections
II.A.2 and II.F.4.d)]. The Administrator has concluded that a two-tier
network design composed of (1) near-road monitors which would be placed
in locations of expected maximum 1-hour NO2 concentrations near heavily
trafficked roads in urban areas and (2) monitors located to characterize
areas with the highest expected NO2 concentrations at the neighborhood
and larger spatial scales (also referred to as “area-wide” monitors)
are needed to implement the 1-hour NO2 NAAQS.  The rationale for this
decision is discussed in section III.B.1 of the preamble to the final
rule.

Comment:  One commenter, Dow, recommended that area-wide monitors should
be located at least 1,000 meters from any major roads or intersections
to ensure that the concentration of NO2 measured is representative of an
area-wide concentration instead of “peak” near-road concentrations.

Response:  EPA believes that Table E-1 of 40 CFR Part 58 Appendix E,
which provides roadway set-back distances for neighborhood and larger
scale sites is appropriate to use to ensure that any NO2 site that is
intended as an area-wide site will be located at a sufficient distance
from any major road.  This issue is discussed section III.B.8.of the
preamble to the final rule.

(5)	Comment:  NYSDEC encourages EPA to develop a table for defining the
minimum distance to the nearest major roadway that can be used to
calculate the area-wide NO2 concentrations, since specific distances are
unlikely to be appropriate for all core based statistical areas (CBSAs).
 

	Response:  EPA believes that existing data in 40 CFR Part 58 Appendix
E, Table E-1, are  appropriate for use in determining the minimum
distance a monitoring site must be from a road with a given annual
average daily traffic count to be considered a neighborhood or larger
spatially representative site.  If a state feels that an increased
minimum distance compared to Table E-1 is appropriate for an area-wide
site in one or more of their particular CBSAs, or on a case-by-case
basis, the state can elect to site their area-wide monitors accordingly,
so long as they continue to meet siting and network design requirements
specified in 40 CFR Part 58.

(6)	Comment:  Several environmental and public health groups (e.g., ALA,
EDF, EJ, NRDC) opposed the proposed requirement to retain only 52 air
monitors to measure area wide concentrations NO2.  

	Response:  EPA is not requiring that the current NO2 network be
reduced; rather, EPA has introduced minimum monitoring requirements
where there was previously no minimum NO2 monitoring requirement.  This
issue is discussed in section III.B.8 of the preamble to the final rule.
Further, the relative current state of the NO2 network is detailed in
the NOx Network Review and Background document (Watkins and Thompson,
2009), which is posted in the docket.  This document provides
information and summaries on what EPA believes the current monitoring
network, according to state provided meta-data, is addressing. 

Comments on the need for a research monitoring network 

A number of commenters, primarily from industry groups and state
agencies recommended that some form of near-road research monitoring
network be utilized before a full-scale near-road monitoring network is
implemented. 

Comment:  A subset of CASAC panel members recommended establishing a
limited roadside monitoring network (20 to 50 monitors in different
cities) for the purpose of informing the development of a national
network.  Specifically, these CASAC member recommended the following: 

A small roadside network could provide more information than a national
network of NOx monitors if the smaller network also monitored other
roadway-associated pollutants, including CO and continuous particulate
matter (PM) with speciation, and if the network also included detailed
meteorological measurements and automatic traffic counters.  They note
that measuring only NO2 could lead to the mistaken impression that air
quality is deteriorating as diesel particulate filters, which can
increase the fraction of NOx that is emitted as NO2, become more common.


If EPA undertakes near-road monitoring, the Agency should seek the
advice of technical experts, such as those in the CASAC Ambient Air
Monitoring and Methods (AAMM) subcommittee. 

Siting criteria should include consideration of such factors as vehicle
mix, location of susceptible or vulnerable populations (e.g., schools,
hospitals), local physical features such as urban street canyons. 

Such a network should be established quickly to capture ongoing changes
in mobile source emissions and should be funded primarily by EPA. 

Response:  The Administrator has judged that a regulatory, two-tier
network design, which includes near-road monitors in urban areas (not a
research network), is necessary to support the intent of the revised
NAAQS to protect against risks associated with exposures to peak
concentrations of NO2 anywhere in an area. This rationale is discussed
in section III.B.1 of the preamble to the final rule. EPA believes that
it will be useful to provide guidance to ensure consistent
implementation of the monitoring network and has committed to engage the
CASAC AAMM subcommittee on this effort, and discusses this issue in
section III.B.5 of the preamble to the final rule.  In regard to
measuring other roadway-associated pollutants, the scope of this
rulemaking pertains to the NO2 NAAQS and the monitoring associated with
it. However, EPA agrees that multi-pollutant monitoring is desirable,
and believes that the sites required in the near-road network design
could be suitable for other pollutants including carbon monoxide,
particulate matter (especially ultra-fine particulate matter), air
toxics, and black carbon.  

Comment:  Some industry commenters (e.g., CE, EEI, NAM, NPRA, NMC, OIPA,
POH, SRNS, SCCC, and SCMA) provided support for a research monitoring
network due to perceived weaknesses in the scientific basis for
implementing an extensive near-road monitoring network.  Generally
indicating that a research network would allow EPA to gain a better
understanding of near-road peak NO2 concentrations and their impact on
public health before requiring the development of a large and expensive
monitoring network.  For example, NAM commented:

If EPA concludes that NO2 near roadway monitoring is required, the NAM
recommends at a minimum, EPA should consider conducting a near roadway
"special purpose" monitoring network program.  Conducting such a near
road monitoring program would allow EPA to collect necessary data that
can be used to better understand the health impacts associated with
short term NO2 exposures.

Response:  The administrator has judged that a regulatory, two-tier
network design, which includes near-road monitors in urban areas (not a
research network), is necessary to support the intent of the revised
NAAQS to protect against risks associated with exposures to peak
concentrations of NO2 anywhere in an area.  EPA believes that the
existing near-road research provides a sufficient base of information to
implement an appropriately designed near-road monitoring network, and
the collective experience that exists in the ambient monitoring
community will allow for successful implementation of that network. EPA
discusses the basis and rationale for the two-tier network design, which
include near-road monitors for regulatory purposes in section III.B.2 of
the preamble to the final rule. 

Comment:  Some state groups (e.g., AASHTO, ILEPA, MIDEQ, MSDEQ, NACAA,
NCDENR, NESCAUM, NMED, SCDHEC, SJAQMD, and VDOT) focused their support
for a research monitoring program by identifying a need to develop more
experience siting near-road monitors.  Creating greater certainty about
where and how near-road monitors ought to be used could reduce
variations in measurement results driven solely by differences in
monitor siting.  These comments are exemplified by NACAA’s comment:

 

…a major new network – particularly one that is inherently
complicated and untried – should not be rolled out without the benefit
of an effective near-road monitoring research program that can address
many of the relevant data questions, and inform the specific siting
requirements of the rule.

Response:  The Administrator has judged that a regulatory, two-tier
network design, which includes near-road monitors in urban areas (not a
research network), is necessary to support the intent of the revised
NAAQS to protect against risks associated with exposures to peak
concentrations of NO2 anywhere in an area.  EPA believes that the
existing near-road research provides a sufficient base of information to
implement a appropriately designed near-road monitoring network, and the
collective experience that exists in the ambient monitoring community
will allow for successful implementation of that network. EPA discusses
the basis and rationale for the two-tier network design, which include
near-road monitors for regulatory purposes in section III.B.2 of the
preamble to the final rule.

(4)	Comment:  The CASAC panel members who support the alternative
approach, which is coupled to a monitoring network that includes only
area-wide monitors, stated that roadside monitors will likely detect
extreme concentrations of NO2 at certain times (e.g., heavy traffic,
idling truck near monitor) and more effort is needed to characterize
these measurements and their relationship to traditional area-wide
measurements.

Response: EPA believes that some urban areas that will have both
area-wide monitors (whether required or not) and near-road monitors may
be able to provide data to increase the understanding or possibly
quantify the relationship (at least for that particular area) between
near-road and area-wide concentrations.  Intense investigation of the
roadway gradient has been, and will continue to be best served by
research studies that utilize transect monitoring and/or saturation
monitoring around target road segments. 

(5)	Comment: The CASAC panel members who support the alternative
approach, which is coupled to a monitoring network that includes only
area-wide monitors, stated that the siting of the near-road monitors
will require consideration of many factors as the monitors will be
sensitive to the microscale environment of the location.  Because the
monitors will capture concentrations at the extreme end of the range,
the data from any two given monitors may not be directly comparable if
the characteristics of their sites vary greatly.

Response: EPA understands that near-road sites will be microscale in
nature, and concentrations at the site will be subject to factors
including AADT, fleet mix, roadway design, congestion patterns, terrain,
and meteorology.  As a result, EPA would expect site to site differences
corresponding to variability in the above factors.  However, the intent
of the revised primary NO2 NAAQS is to protect against the maximum
allowable NO2 concentration anywhere in an area, which includes ambient
air on and around roads.  The required near-road monitoring sites are
intended to measure the maximum expected concentrations, considering the
above factors.  

(6)	Comment:  The CASAC panel members who support the alternative
approach, which is coupled to a monitoring network that includes only
area-wide monitors, stated that siting monitors based on traffic count
alone may miss locations where maximum concentrations of NO2 occur.  EPA
should investigate model results from Congestion Mitigation and Air
Quality (CMAQ) Land Use Regression modeling as well as Gaussian plume
models and emissions inventories to determine where roadside monitors
should be located.

Response: As noted in section III.B.6 of the preamble to the final rule,
EPA does not intend for AADT counts to be the sole basis for choosing a
near-road site.  EPA understands that there are other factors that can
influence which road segment in a CBSA may be the actual location where
the maximum NO2 concentrations could occur.  These factors include
vehicle fleet mix, roadway design, congestion patterns, terrain, and
meteorology.  When states identify their top-ranked road segments by
AADT, EPA intends for states to evaluate all of the factors listed
above, which influence where the location of expected maximum NO2
concentration may occur, when evaluating a pool of candidate near-road
monitoring sites.  Further, EPA expects that modeling will be a tool
used by states to assist in informing where near-road monitoring sites
should go.

(7)	Comment:  The CASAC panel members who support the alternative
approach, which is coupled to a monitoring network that includes only
area-wide monitors, commented that diesel particle filters increase the
fraction of NO2 in NOx and may lead to higher measured concentrations of
NO2 in places where vehicles are actually cleaner.

Response:  EPA notes that the required near-road NO2 monitors are to be
sited  in locations that are expected to measure maximum NO2
concentrations.  The issue of whether the fleet, or a component of the
fleet, travelling on the target road is relatively dirty or clean is
part of consideration of fleet mix that states are to consider when
identifying near-road site locations.  

(8)	Comment:  Many state and industry commenters (e.g., NACAA, NESCAUM,
many states, Consumers Energy, EA, EEI, NAM) recommended that a
near-road research monitoring program be implemented before a full-scale
near-road monitoring network is required.  Specifically, EA recommended
that a near-road monitoring network not be required until enough
near-road monitoring data was available to ensure that EPA’s proposed
NO2 level is not set below background NO2 levels.  

Response:  The EPA believes that is has sufficient understanding of the
near-road environment from existing research to propose a near-road
network that is appropriate for regulatory use.  This issue is discussed
in issue in section III.B.2 of the preamble to the final rule.  In
addition, as discussed in section II.F.4.d of the preamble to the final
rule, the Administrator has set the level of the 1-hour standard at 100
ppb to protect public health with an adequate margin of safety.  Her
decision is based on consideration of available health evidence,
exposure/risk information, the advice of CASAC, and input from the
public.

Justification of meteorological measurements

Most commenters did not mention the issue of meteorological measurements
but the comments received on the subject were unanimously against
requiring monitoring stations to include three-dimensional anemometry
equipment.

Comment:  CASAC Panel members and a number of state commenters (e.g.,
AKDEC, NCDENR, SCDHEC, and WIDNR) questioned the need for meteorological
measurements at the NO2 monitoring sites.  These commenters called into
question the ability of three-dimensional anemometry to provide the
microscale meteorological information EPA desires if the monitors are to
be located close to the ground and other structures, and if they are
averaging measurements over hour long periods of time. SCDHEC commented
that the recording of air turbulence data at monitor stations ought to
be encouraged but not required unless EPA were willing to fully fund the
purchase, installation, and operation of the necessary equipment.  These
commenters concluded that the requirement to include equipment for
measuring meteorological characteristics would be an unnecessary
expense.

Response:  EPA has chosen not to finalize a requirement for
meteorological measurements, including three-dimensional anemometry, at
near-road NO2 monitoring stations, but does encourage states to perform
meteorological monitoring to better characterize the behavior of NO2
cocentrations in the near-road environment.  EPA discusses its rationale
for this decision in section III.B.9 of the preamble to the final rule.

Monitoring technology 

Commenters (CASAC Panel members, some state agencies, and some industry
groups) who expressed an opinion regarding monitoring techniques
generally agreed that the proposed chemiluminescence approach is
appropriate.  However, some groups (e.g., CASAC and industry)
recommended that EPA promote the development of additional methods for
measuring true NO2 that would not be subject to positive interference. 

(1)	Comment:  Several commenters (e.g., MODNR, SRNS, Teledyne API, and
UARG) recommended that EPA encourage the development of alternative
monitoring technologies, which includes the photolytic-chemiluminescence
method or cavity ring-down spectroscopy, and to permit the use of these
technologies for monitoring if the new technologies provide comparable
or better monitoring results than existing FRM/FEM [Federal Reference
Method/Federal Equivalent Method].  NCDENR urged EPA to maximize the
flexibility in the choice of FRM and FEM technology available to
demonstrate compliance so long as the technology provides comparable
results.  

Response:  The FRM/FEM program ensures that only evaluated and approved
methods may be used for comparison to the NAAQS (see 40 CFR  Part 53). 
EPA only approves a particular method for use in measuring oxides of
nitrogen once it has been shown to meet requirement set forth in 40 CFR
Part 50 Appendix F, with testing and data provided usually provided by
instrument vendors. The use of FRM/FEMs and furthering method
technologies is discussed in section III.A.1 of the preamble to the
final rule.

(2)	Comment:  One industry commenter, Aerodyne, expressed concern that
current NO2 monitoring equipment and siting protocols will not be
suitable for use in near-road applications without proper
reconfiguration. The commenter notes that the method by which the
chemiluminescence FRM determines NO2 concentrations, using alternating
measurements of NO and NOX and subsequently using the difference to
indicate NO2, may not be appropriate because this technique does not
possess the time response required to accurately track the plumes of
combustion gas emitted by passing vehicles.  The commenter notes that
these plumes “tend to be short-lived, with durations on the order of 3
to 30 seconds.” As a result, the commenter asserts that the values
reported by chemiluminescence FRMs “…can be quite inaccurate because
of the inability of the alternating channel monitors to properly average
these plumes.” The commenter goes on to note that this issue
“…begs the question of whether any of this affects reported 1 hour
average NO2 concentrations, the subject of the proposed regulations.”

Response:   The “alternating measurement” technique utilized by the
chemiluminescence FRM is asserted by the commenter to cause variability
in measuring short duration plumes or ‘bursts’ of NO2.  EPA believes
this variability would be most apparent when there is only one plume, or
a few individual plumes, that may pass the monitor for a short period of
time, but is unlikely to cause a discernable effect in the 1-hour
average.  EPA believes that this particular characteristic of the
chemiluminescence FRM does not warrant a new measurement approach for
near-road monitors because the averaging time of the NAAQS is 1 hour. 
The NAAQS is a 1-hour standard, which EPA believes is a long enough
averaging period where short-term variations, on the order of 3 to 30
seconds will be averaged out, an are not expected to have a discernible
effect.  The commenter states that “It has always been (rightly)
assumed that when reporting 24 hour and /or yearly averages, the short
term upsets observed in the NO2 data have no discernible effect.”  EPA
notes that states calculate 24-hour and annual averages from the 1 hour
averages they acquire each day. 

	The commenter suggests several alternatives to relying on a single
chemiluminescence analyzer at near-road sites.  EPA has expressed its
desire to further the development of alternative methods in determining
NO2 concentrations, which is discussed in section III.A.1 of the
preamble to the final rule, and also is open to alternative applications
of existing methods, all of which would need to follow the process for
submission and approval described in 40 CFR part 53.

(3)	Comment:  Some industry commenters expressed concern that open path
monitors might not be appropriate for use in near-road applications. 

Response:  EPA notes that path integrated optical remote sensing
techniques, also known as open-path methods, are not typically used by
states at State and Local Ambient Monitoring Station (SLAMS) sites. 
However, there are approved open-path methods capable of providing
hourly NO2 concentrations that can be compared to both the annual and
1-hour standard.  EPA believes that an appropriate path length is
between 50 meters to 300 meters, where 50 corresponds to the maximum
distance away from the edge of the nearest traffic lane of a nearby road
segment and 300 meters corresponds to the traditional maximum path
length at micro- or middle scales site.  EPA recognizes that if open
path methods are going to be used in the near-road environment,  issues
regarding path orientation, safety, and optimal path length within the
allowable range will likely be considered in the monitoring guidance
that will need to be developed to assist states in the implementation of
the network design.

(4)	Comment:  One industry commenter, AQRL, requested that EPA set
guidelines for the maximum allowable inlet length and sample residence
time.

Response:  EPA has only chosen to set a maximum residence time of NO2 in
the sample line between the inlet probe and the analyzer of 20 seconds. 
EPA discusses this issue in section III.B.7 of the preamble to the final
rule. 

(5)	Comment:  API commented that, because state and local air monitoring
agencies have little experience in near-road NO2 monitoring, EPA needs
to be prepared to support significant field validation and statistical
analysis to assess monitor performance and dispersion and photochemical
modeling analyses to assist in siting monitors. 

Response:  EPA believes it has systems in place that can provide
“field validation and statistical analysis” of any ambient
monitoring network. Field validation and statistical analysis starts
with the initial testing of instruments to ensure that they are
operating properly and providing precise and unbiased measurements
before monitoring begins. All monitoring organizations are required to
have quality management plans and quality assurance project plans that
provide the necessary elements to help in their achievement of the NO2
data quality goals.  In addition, the quality assurance (QA)
requirements in 40 CFR part 58 Appendix A provide the quality control
samples, the auditing activities and the appropriate statistical
analysis that monitoring organizations and EPA can use to judge whether
the monitoring systems are under control from a data quality standpoint.

(6)	Comment:  Several state commenters (e.g., MODNR and NCDENR) suggest
that the current bias measurement uncertainty criteria used for other
NAAQS will be appropriate and acceptable for use with a 1-hr NO2
standard.

Response: EPA is finalizing the approach to develop data quality
objectives, and is  changing the proposed goal for measurement
uncertainty, to match those criteria of the NCore network and the
historical approach for NO2, where the goals for acceptable measurement
uncertainty for NO2 methods for precision is an upper 90 percent
confidence limit for the coefficient of variation (CV) of 10 percent and
for bias is an upper 95 percent confidence limit for the absolute bias
of 15 percent.

This issue is discussed section III.C of the preamble to the final rule.

(7)	Comment:  A state commenter, IADNR, recommend allowing provisions
for mobile monitoring to determine NO2 exposure to motorists.   

Response:  The use of mobile monitoring may be a very beneficial tool to
assist states in locating candidate near-road NO2 sites; however, EPA
does not believe that mobile monitoring is suitable for regulatory
monitoring.  The network design is predicated on placing monitors in the
location of expected maximum concentration, which is inherently in a
generally fixed location.  Further, much consideration would have to be
given to issues regarding the deviation from the standard practice of
using fixed point (or fixed path-integrated) data for comparison to the
NAAQS. Another issue that would arise would be regarding data
completeness and representation, e.g. at what time and where the data
are valid, and whether the data are representative of a given location. 
EPA is encouraged by the interest in mobile monitoring, but we believe
that such monitoring is currently best suited for non-regulatory
informational purposes.

Timing of monitor deployment

In general, the majority of comments about the timing of monitor
deployment, State Implementation Plan (SIP) development, and
determination of attainment were received from state agencies and
industry groups recommending that EPA keep the proposed current
implementation deadlines as proposed or extend the deadlines further.  A
few environment and public health groups provided comments urging EPA to
move the implementation deadlines forward. [who were the commenters
among the states, industry and enviros?]

Comment:  A lage majority of state and industry groups who made comments
on network implementation recommended that the deadline for installing
near-road monitors remain January 1, 2013, or that EPA provide the
maximum amount of time possible.  They claimed that the current January
1, 2013 deadline provides insufficient time to plan and implement such a
complicated network of monitors.  SJAQMD went further,  recommending
that EPA allow for schedule relief if a state can demonstrate that
monitor installation delays are the result of delays in obtaining
permits or rights-of-way for the near-road monitors.  Conversely, the
environmental and public health groups who made comments on network
implementation recommended that urged “EPA to seek Congressional
funding for an expanded network and to set a deadline for deployment of
no later than January 1, 2012.”  

Response:  EPA recognizes the need to aid state agencies in the network
implementation process, particularly through guidance documentation that
will be developed in partnership between EPA and various stakeholders
including NACAA and the states.  Further, EPA has modified the timelines
for implementation by changing the date by which state and, when
appropriate, local air monitoring agencies shall provide a plan for
deploying monitors in accordance with required network design from July
1, 2011 to July 1, 2012.  However, EPA is finalizing the date by which
state and, when appropriate, local air monitoring agencies shall
establish the required NO2 monitoring network as January 1, 2013, as was
proposed.  The basis and rationale for this decision is discussed in 
section III.B.5 of the preamble to the final rule.

Other monitoring issues

Many comments about the monitoring network associated with a one-hour
standard were also submitted, including comments on what should be
considered an exceptional event and the appropriate requirements for
using data from monitors for evaluating attainment.

Comment: Some commenters urged EPA to create criteria for defining
exceptional events for near-road monitors that included events like road
construction.

Response:  The Exceptional Events Rule (EER) and the accompanying
preamble (72 FR 13560, March 22, 2007) set forth and explain the
criteria and procedures that must be used to determine whether an event
qualifies as an exceptional event and the documentation that must be
submitted to support an exceptional events claim.  For an exceptional
event claim to be approved and data affected by such an event to be
excluded from consideration in any NAAQS attainment designations, the
data would need to meet the criteria and procedures established in the
rule.  EPA believes that these criteria and procedures are sufficient to
address any exceptional events claims that may arise for NO2 NAAQS. 
Additions or modifications to the EER are not needed for any exceptional
events claims that may arise for NO2.  In addition, the current
exceptional event data flags available in the AQS--unique traffic
disruption and other-- provide sufficient flexibility to address
near-road exceptional events.  

Comment:  One state commenter, NYSDOT, recommended (if EPA chose to
finalize the two-tier network design which included near-road
monitoring) that “EPA establish national guidance so there is
reasonable uniformity among EPA regions in the implementation of these
provisions.”

Response:  EPA notes that network consistency, from site to site, is
achieved in ambient monitoring networks by the adherence to monitoring
regulations and through the use of guidance. EPA has detailed siting
criteria  as specific as we believe is necessary and appropriate to
develop a consistent monitoring network , while allowing some necessary
flexibility for logistical considerations that will occur on a
case-by-case basis.  Further, EPA recognizes the need to aid state
agencies in the network implementation process, particularly through
guidance documentation that will be developed in partnership between EPA
and various stakeholders including NACAA and the states.  This issue is
discussed in section III.B.5 of the preamble to the final rule.

Comment:  Some commenters recommended that the annual primary standard
design value be calculated by averaging all 1-hour values for the year,
while a third commenter recommended the design value be calculated by
averaging within calendar quarters before averaging across calendar
quarters.

Response:  This issue is discussed in the final notice (Section IV.A).

Comment:  Several comments were made in regard to EPA’s data
completeness test.  For example, three commenters agreed that the
completeness test ought to stay at 75% for each quarter while a fourth
commenter recommended the percentage be increased to 82%.

	Response:  This issue is discussed in the final notice (Section IV.B).

Comment:  A comment from a State agency noted that the proposed 75%
capture requirement was inappropriate and should not be used for a 1-hr
NO2 standard. 

	Response:  This issue is discussed in the final notice (Section IV.B). 

Comment:  A commenter requested that EPA truncate NO2 concentration
measurements rather than rounding them.

	Response: EPA agrees with this comment, as reflected in the final
notice.  

Comment:  An industry group requested that EPA further clarify the
guidelines for situations where incomplete monitoring data could be
considered valid. 

Response:  The rule text identifies factors for the administrator to
consider in determining when to use such data.  No further clarification
is necessary because the variety and uniqueness of such situations
require that they be treated on a case by case basis.

Comment:  One commenter expressed concern that the use of near-road
monitoring would lead to near-source monitoring for other point sources
of NO2 and other pollutant emissions.

Response:  EPA is establishing a two-tier monitoring network, which
includes both near-road and area-wide monitoring, under this rule. EPA
does not believe that near-road monitoring will directly result in
non-near-road monitoring.  However, EPA recognizes that in certain
circumstances, there can be an area or areas of expected maximum
concentrations of NO2 due to non-road, point, or area sources, that may
not be monitored even though a state is fulfilling its minimum
monitoring requirements.  EPA has included a mechanism to deal with such
circumstances by providing the Regional Administrator with the authority
to require additional monitors above those minimally required. This
issue is discussed in section III.B.4 of the preamble to the final rule.


Comment:  A State agency recommended that EPA allow averaging across
monitors for design value calculations and substitution of values from
additional monitors for the primary monitor, for example, if the primary
monitor is taken offline for quality control (QC) during a high
pollutant period.

Response:  This issue is discussed in the final notice (Section IV.B).  

Air Quality Index 

Comment:  Many State agency commenters discouraged the use of near-road
concentrations in calculating the AQI since near-road NO2 concentrations
do not represent exposure of the general public to ambient NO2.  Several
other commenters, including NACAA, expressed concern that using
near-road monitors for calculating the AQI could diminish the value of
the index as a metric for reporting air quality to the general public by
making it sensitive to NO2 concentrations that are present only in small
areas.  CASAC also raised questions about how the AQI would be affected
by roadside monitors.  An additional commenter noted that it would be
difficult to forecast the AQI for regions where near-road monitoring
stations existed, since near-road NO2 concentrations are likely to be
unpredictable.

	Response:  The response to this comment on the AQI is discussed in
section VII of the final rule.  

Comment:  The ALA, EJ, EDF, NRDC support setting the 100 level of the
AQI at 50 ppb and the level of the “moderate” category well below 50
ppb (i.e., 25 ppb).  These commenters disagree with EPA’s proposal to
maintain the breakpoints at the higher end of the AQI scale (i.e., 200
to 500).  They note that the 500 level has not been changed in the last
22 years, and EPA cannot claim that it represents the most up-to-date
scientific evidence without more extensive review.  They also note that
the proposed 200 level for NO2 is too high, relative to the 100 level,
given the relationship between these levels for other pollutants.  As a
result, these commenters recommend that EPA “establish a meaningful
and practical scale of levels of concern and graduated cautionary
statements for both sensitive groups as well as the general
population” based on the current scientific evidence.

In addition, NACAA expressed support for the “EPA’s proposed range
of 0.040 to 0.053 ppm for the AQI value of 50 and 0.360 and 0.370 ppm
for the AQI value of 150.”  But NACAA did express some concern that it
might make for a “confusing outreach message if the AQI level of 50 is
set below the annual standard of 0.053 ppm” since this would mean that
state and local air agencies would “be forecasting moderate air
quality even though an area is meeting the annual standard.”

Response:  With respect to an AQI value of 100, EPA has concluded that
it is appropriate in this case to set this value at 100 ppb NO2, 1-hour
average, the level of the short-term standard.  With respect to an AQI
value of 50, EPA agrees with the NACAA comment which noted it might make
for a confusing outreach message if this value is set below the level of
the annual standard.  This method of structuring the index is
appropriate in the case where a short-term standard is set to protect
against the health effects associated with short-term exposures and/or
an annual standard is set to protect against health effects associated
with long-term exposures.  In such cases, the short-term standard in
effect defines a level of health protection provided against short-term
risks and thus can be a useful benchmark against which to compare daily
air quality concentrations. 

With respect to the breakpoints at the higher end of the AQI scale
(i.e., 200 to 500), EPA acknowledges that these breakpoints have been in
place for a long period of time, and that an AQI value of 200 is higher
for NO2, relative to the 100 level, than it is for the other pollutants
in the AQI.  EPA does not agree that there is new health evidence that
would provide support for changing these levels.  The levels at the
upper end of the AQI are linked to a related program designed to prevent
air pollution emergencies.  Some state and local air agencies are
required to have plans to take actions at these levels, thus it is
important to have a firm basis for revising these levels.  

Significant harm levels (SHL) are those ambient concentrations of air
pollutants that present an imminent and substantial endangerment to
public health or welfare, or to the environment, as established in 40
CFR part 51.151.  The SHL is typically set at the same ambient
concentration of a pollutant as the AQI value of 500.  Appendix L of 40
CFR part 51 includes example emergency episode plans as part of the
Prevention of Air Pollution Emergency Episodes program.  This program
requires specified areas to have contingency plans in place and to
implement these plans during episodes when high levels of air pollution,
approaching the SHL, are in danger of being reached or have been
reached.  The Appendix L example links the AQI with emergency episodes. 
 SEQ CHAPTER \h \r 1  AQI values of 200, 300 and 400 are the basis for
the example Alert, Warning and Emergency episode levels included in 40
CFR part 51, Appendix L.  In this guidance, the Alert level corresponds
to the breakpoint between the Unhealthy and Very Unhealthy categories,
and the Warning level corresponds to the breakpoint between the Very
Unhealthy and Hazardous categories.  The Hazardous category ranges up to
the SHL.

    

 

Comments on the Process for Reviewing the NO2 Primary NAAQS

A number of comments were received recommending changes to the current
rulemaking process, either for NAAQS in general or the NO2 NAAQS in
particular.  

Comment:  A number of industry groups and state agencies recommended
that the proposed near-road monitoring network be considered in a
separate rulemaking from the proposed new 1-hr NO2 standard, because it
represents a dramatic shift in the method of monitoring ambient air
quality from previous NAAQS.

Response: EPA notes that separating the monitoring issues from the NAAQS
revision is not consistent with recently finalized or planned NAAQS
reviews.  EPA has promulgated monitoring rules as part of, or
simultaneously with the PM and Pb rules, and has proceeded to include
the monitoring rule as part of this primary NO2, as well as the primary
SO2, the secondary NOx/SOx, and CO revisions.  Monitoring issues were
handled separately from the 2008 ozone NAAQS revision, however EPA
intends that future ozone NAAQS reviews will integrate monitoring into
the action.

Comment:  Some commenters (e.g., NACAA, NESCAUM, Roger McClellan) noted
that EPA had originally intended to publish an advance notice of
proposed rulemaking (ANPR) and that because no ANPR was published, the
public and CASAC had limited opportunity to comment on the range of
available options.  Other comments expressed disapproval with the
limited amount of time for stakeholders to comment on EPA’s ISA and
REA; the late notifications that EPA provided when comment periods were
extended; and the inclusion of new sections and studies in the final
versions of the ISA and REA, which denied stakeholders the opportunity
to comment on these additions.  

Response:  The Clean Air Act requires EPA to review and to revise, as
appropriate, the NAAQS every five years.  As a result of a lawsuit
challenging EPA’s failure to complete its review of the air quality
criteria for NOX and the NAAQS for NO2, inter alia, EPA agreed in a
consent decree to a schedule for this review of the NO2 primary NAAQS. 
Pursuant to CAA § 113, EPA provided an opportunity for comment on the
consent decree.  This schedule was approved by the United States
District Court for the District of Columbia.

We disagree that EPA’s decision to not publish an ANPR limited the
opportunity for meaningful comment on the rulemaking.  EPA on occasion
uses an ANPR to seeking input from stakeholders before issuing a
proposed rule, but EPA is not required to publish an ANPR.  Moreover,
EPA provided for stakeholder comment throughout the review process,
including 60 days of comment and held two public hearings on the
proposed rule.  To the extent that “new sections and studies” were
added to the final version of the ISA and REA, stakeholders had the
opportunity to comment on this material following the notice of proposed
rulemaking.

Comment:  API commented that EPA did not not adequately address comments
provided by other government agencies on the draft proposed rule. 

Response:  EPA engaged in extensive discussions with other agencies as
part of the interagency review process prior to the publication of the
proposed rule.  Under section 307(d) of the CAA, EPA is under no
obligation to provide written responses to comments provided by other
agencies during the interagency review process.  The scientific
rationale for the standards is clearly provided in the record for the
final rule. 

Comment:  AASHTO requested that if EPA creates a 1-hr NO2 standard more
stringent than 80 ppb, an additional public comment period be created to
allow stakeholders to comment directly on the specific standard level
selected. 

Response:  As described in section II.F.4.d of the final rule, the
Administrator has set the level of the 1-hour standard at 100 ppb.  

Comment:  AGCA commented that EPA’s Regulatory Impact Analysis (RIA)
did not provide enough information to predict which areas of the country
will not be in attainment with a new standard if the proposed near-road
monitoring network is also included.  This lack of information has the
potential to negatively affect AGCA members because it prevents the
organization from assessing the potential impact of a new near-road
monitoring network on the road construction industry.

Response:  We developed the best estimate we were able to develop of
future design values at currently existing monitors, based on emission
inventory projections for 2020, using the best available emission data
and control strategy information.  What we are not able to do is to
credibly estimate specific design values at locations where monitors do
not currently exist, but will in the future.  States will determine
where to place those monitors as they implement the NAAQS.  In addition,
we did not model near-roadway concentrations as a function of traffic
density; this RIA is a national scale analysis.

Interpretation of the Clean Air Act 

Comment: Numerous comments were received from industry groups and state
agencies expressing concern with EPA using near-road measurements of NO2
concentrations to evaluate area-wide attainment with an NO2 NAAQS.  Many
of the commenters note that EPA’s proposed near-road monitoring
network focuses on monitoring roadways likely to have high NO2
concentrations without requiring any demonstration that large and
adjacent populations are exposed to similarly high concentrations. 
Consequently, the commenters argue that EPA has proposed a monitoring
network using rationale that does not reflect the purpose of the CAA and
NAAQS, which is to protect the public from exposure to unhealthy
concentrations of pollutants found within “ambient air” rather than
just “air”.

Response:  We note that ambient air includes the outdoor air present
around roads as well as the air that intrudes from outdoors into homes,
buildings, or vehicles on/near those roads.  As discussed in the final
rule, people who spend time on or around major roads (e.g., because they
live there, go to school there, or commute in vehicles on major roads)
can be exposed to elevated concentrations of ambient NO2.  

Comment:  A number of industry groups provided comments maintaining that
EPA may have overstepped its regulatory authority with its proposed 1-hr
NO2 standard.  Comments included claims that the health effects from
short-term exposure to high concentrations of NO2 fail to constitute the
“adverse health effects” called for in the CAA.  Other comments
noted that EPA is authorized by the CAA to set its NAAQS at levels
intended to protect sensitive subpopulations, not the most sensitive
individual.  In addition, some commenters noted that NAAQS are not
intended to eliminate all risks.   

Response:  The issues raised by these commenters are addressed in
sections II.E.2, II.E.3, II.F.4 in the final rule.  

Comment:  Roger McClellan provided some general comments on the role of
CASAC and concluded that-- 

In offering recommendations for a specific upper bound for the Standard,
the Committee moved beyond advising on the science informing the policy
judgments inherent in setting the Standard, to taking on a role reserved
for the Administrator – the setting of the Standard. The Clean Air Act
wisely delegates the setting of the Standard to the Administrator and
calls for a Clean Air Scientific Advisory Committee, not a Clean Air
Standard Setting Committee.

Response:     Section 109(d)(2)(B) of the Clean Air Act, and the charter
of CASAC, provide that CASAC shall review the air quality criteria and
the NAAQS and “recommend to the Administrator any new national ambient
air quality standards and revisions of existing criteria and standards
as may be appropriate.”   EPA believes that CASAC fulfilled this duty
and provided helpful advice as a committee on the science issues raised
by this review.  Of course, the Administrator cannot, and did not,
delegate the setting of the standard to CASAC.

Comment:  NPRA commented that EPA lacks the legal authority to withhold
a designation of “attainment” from districts unless the districts
cannot be classified on the basis of available information as meeting or
not meeting the national air quality standard.  The commenter argued
that, at present, EPA cannot know whether it will lack the necessary
information in 2012.  According to NPRA, “it is clear from the statute
that for NO2 EPA must designate areas as in attainment that have
attaining monitors absent an affirmative showing that such an area is
contributing to downwind nonattainment in another area..”  In
addition, NPRA commented that section 107(d) of the CAA requires that
EPA “consider the recommendations of States with regard to the
designation of areas as attainment, nonattainment and unclassifiable.”

One state agency agreed with EPA’s proposal to designate all areas
currently in attainment with the NO2 standard as “unclassifiable”
until a near-road network has been established and measurements have
been evaluated.

	Response:  (Implementation group) 

(5)	Comment:  Several environmental groups (e.g., ALA, EDF, EJ, NRDC)
noted that the CAA requires that NAAQS be set to protect public health
with an adequate margin of safety without consideration of cost and
that, in setting the NAAQS, EPA must err on the side of protecting
public health.  These groups stated that this includes protecting
sensitive subpopulations and guarding against “potential” health
effects.  

Response:  EPA agrees that the NAAQS must be requisite (i.e., no less
and no more stringent than necessary) to protect the public health,
including the health of sensitive populations, with an adequate margin
of safety.  The Administrator’s consideration of this issue in the
current review is described in section II.F.4.d of the final rule.  

Comments on Implementation 

(1)	Comment:  A minority of CASAC Panel members posed the question of
what would be an effective control strategy to reduce NO2 concentrations
around major roads.  A number of public commenters also expressed the
concern that states have little ability to reduce on-road emissions from
motor vehicles.  Several went on to opine that it is the role of the
federal government to reduce emissions from motor vehicles.  One pointed
out that their state has already implemented measures to reduce on-road
emissions of NOX as part of their plan to attain the ozone standard. 
One commenter opined that states have few options to reduce on-road
emissions beyond federal requirements.  Another commenter suggested that
one option for states would be to adopt California’s mobile source
regulations including those that apply to non-road equipment.  The
commenter is concerned that adoption of California’s non-road
regulations would place a significant burden on the construction
industry.  The commenter is also concerned that states may attempt to
reduce emissions from non-road equipment by imposing operating
restrictions, requiring older equipment to be replaced or mandating
retrofits for older equipment.

Response:  EPA agrees that it is a federal responsibility to implement
regulations that reduce emissions from new light- and heavy-duty
vehicles.  However, states have authority that can be used to reduce
emissions from in-use light- and heavy-duty vehicles.  For example,
states can implement programs to retrofit heavy-duty diesel vehicles to
reduce their NOX emissions.  States can implement regulations to reduce
or eliminate long-duration idling of heavy-duty diesel vehicles.  They
can implement an inspection and maintenance program for light-duty
vehicles.  Additionally, states can implement a wide range of programs
to improve travel efficiency.  States can implement strategies such as
congestion pricing, programs to reduce trips for commuting purposes, and
measures to improve the operational efficiency of the transportation
network.  States can also work with freight shippers to improve the
efficiency of goods movement in an area.

EPA agrees that a number of states have already implemented controls on
on-road sources.  The emissions reductions attributable to those
measures will help such areas to attain the NO2 NAAQS and can be
included in a SIP for NO2.   If such an area was designated
nonattainment for NO2, EPA would expect the state to evaluate mobile
source controls that it had not yet implemented and to evaluate the
potential for controlling emissions from other sources that are
contributing to the nonattainment problem.

With regard to the comments related to the adoption of California’s
on-road and non-road regulations by other states and the possibility
that some states may choose to reduce NOX emissions by requiring older
non-road equipment to be replaced or equipped with retrofits or impose
operating restrictions on non-road equipment, EPA expects each state
with a designated NO2 nonattainment area to develop a SIP that brings
the area into attainment by the applicable deadline and that each state
would evaluate the potential for controlling emissions from all sources
that are contributing to the nonattainment problem.  CAA section 177
allows states to adopt California’s standards that apply to new motor
vehicles or new motor vehicle engines and CAA section 209(e)(2)(B)
allows states to adopt California’s standards that apply to new or
used non-road engines and equipment.  California, and therefore other
states, are prohibited by CAA section 209(e)(1) from promulgating
standards for new construction and farm equipment with engines smaller
than 175 hp and from new locomotives and locomotive engines.  It is
possible that a state could choose to adopt California’s standards
that apply to new on-road and/or other non-road equipment in order to
attain the NO2 NAAQS.  States may also choose to reduce emissions from
non-road equipment through operating restrictions, or requiring older
equipment to be replaced to the extent that such measures are not
federally pre-empted.  EPA would expect that before any state included
any of these types of measures in a SIP for NO2 the state would have
evaluated a wide range of potential control measures and concluded that
such measures are necessary to bring the area into attainment by the
applicable deadline.

(2)	Comment:  Some commenters stated that modeling for permitting of the
short term 1-hr NO2 standard could likely show that stationary sources
violate the standard as opposed to near-road mobile sources.  Others
requested that if EPA requires near-roadway monitoring for NO2, it
should provide clarifying guidance stating that near-road monitors would
not be applicable for NSR or other point source permit applications.  

Response:  The permitting regulations for Prevention of Significant
Deterioration of Air Quality (PSD) require applicants to provide 1 year
of air quality data representative of the air quality in the area that
the proposed new or modified source will impact.  Typically, such
monitoring data is used to represent background air quality levels that
are considered along with modeled concentrations of a pollutant to show
that the source’s predicted impact will not cause or contribute to a
violation of any NAAQS.  We agree that the ambient data collected from
near-road monitoring sites for NO2 may not be representative of
background air quality levels throughout the area of concern.  However,
if used correctly, such data could be useful in assessing the source’s
contribution in specific areas of potentially high ambient NO2
concentrations largely attributable to near-road mobile source
emissions.  EPA will give further consideration to this comment as we
evaluate any NSR guidance and regulations that may be needed to
implement a 1-hour NO2 NAAQS.

(3)	Comment:  API recommends that EPA update the Guidelines on Air
Quality Models to address screening and refined models for modeling peak
short-term concentrations in order to accurately predict short-term NO2
concentrations in diverse environmental settings.  

Response:   The EPA agrees with this comment and intends to review the
need to revise the Guideline on Air Quality Models to address issues
associated with the promulgation of a short-term NO2 NAAQS. 

(4)	Comment:  API also recommends that EPA develop a database of NO2 to
NOX emissions ratios for the variety of sources that would be affected
by any new NO2 standards.  The database could also include methods to
make use of data from existing ambient monitoring networks to provide
realistic background ozone and NO2 concentrations required for the
short-term NO2 modeling framework.  

Response: The EPA will review the need for the development of a database
of NO2 to NOX emissions ratios sources following the promulgation of the
NO2 NAAQS. 

(5)	Comment:  CAC disagrees with EPA’s decision not to impose
non-attainment classifications on areas with measured near-road NO2
concentrations in excess of the new NO2 standard, and urges EPA to
provide a graduated non-attainment classification system for the new
standard.  According to the Council, “a classification system defining
higher levels of non-attainment with increasingly stringent requirements
at those levels is one that allows for finer calibration of air quality
regulatory response defined at the federal level.”  

Response:  As stated in the proposed rule, Section 192(a), of part D, of
the CAA specifically provides an attainment date for areas designated as
nonattainment for the NO2 NAAQS.  Therefore, EPA has legal authority to
classify NO2 nonattainment areas, but the 5 year attainment date
addressed under section 192(a) cannot be extended pursuant to section
172(a)(2)(D).  Based on this limitation, EPA proposed not to establish
classifications within the 5 year interval for attaining any new or
revised NO2 NAAQS. See the preamble to the final rule for consequences
for areas failing to meet the attainment date.  Once EPA makes a
determination that an area has failed to meet its attainment date, the
State will be requested to submit a SIP revision which must show that it
can attain the standard as expeditiously as practicable, but no later
than 5 years from the determination of failure to attain.  

(6)	Comment:  One commenter opined that states that fail to develop SIPs
or meet EPA’s CAA deadlines would be subject to “numerous federal
sanctions” including emissions caps, limiting economic development and
the loss of federal highway transportation dollars.  The commenter
specifically states that a construction ban would impede projects that
would improve municipal water supplies and wastewater treatment.

Response:  CAA section 179 provides EPA with only two sanctions.  These
are the 2:1 offset sanction that applies to new or modified major
stationary sources and the highway funding sanction.  These sanctions
are only applied in specific circumstances, such as when a state has
failed to submit a required SIP or when a SIP revision is disapproved.  
 The CAA no longer contains a sanction specifically aimed at limiting
development.  The CAA Amendments of 1990 do not include the construction
ban that applied in certain situations under the 1977 CAA.  Therefore,
EPA disagrees that, if the available sanctions are imposed in an NO2
nonattainment area, they would result in a de facto construction ban and
they would not delay municipal water or wastewater projects.  It must be
noted that there is no sanction for failing to reach attainment by the
CAA deadline.  For areas that fail to attain by the applicable deadline
CAA section 179(d) requires that such areas submit a revised SIP within
one year after EPA publishes a notice finding that the area failed to
attain.  In the case of NO2, this revised SIP would have to demonstrate
attainment as expeditiously as practicable, but no later than 5 years
after the date of the Federal Register notice in which EPA made the
finding that the area had failed to attain.

(7)	Comment: One commenter pointed out that a new 1-hr standard for NO2
will complicate permitting of small sources that may contribute to local
NO2 emissions.

Response: Section 110(a)(2)(C) of the CAA requires states to regulate
the construction and modification of any stationary source as necessary
to assure that NAAQS are achieved.  Consequently, the addition of a
1-hour NO2 NAAQS would require that the necessary consideration be given
to preventing sources from violating a 1-hour NO2 NAAQS.  We do not,
however, view this as a complication of the permitting process.  In some
instances, a modeling analysis may be required to make the necessary
compliance demonstration, while in other cases modeling may not be
required.  Such decisions as to the level of analysis needed to make the
necessary compliance demonstration must be made by the applicable state
or local agency permitting authority. 

(8)	Comment:  Several industry commenters requested that EPA slow the
timeline for implementing a near-road monitoring network and designating
roadway areas, because they believe EPA lacks significant information
about the implementation and performance of a national, near-road
monitoring network.  

Response:  EPA believes that there is sufficient information to
implement a near-road NO2 network and that the timeline for
implementation will allow EPA to consult CASAC and collaborate with
states in the development of guidance that will assist in the
implementation of a network that is as consistent as possible.  This
issue is discussed in section III.B.5 of the preamble to the final rule.
  

(9)	Comment:  Numerous comments were received, primarily from state
agencies, noting the significant cost of a near-road monitoring networks
and requesting that the federal government pay for a large percentage of
the network’s cost by funding the monitors through Section 103 of the
CAA, rather than section 105.  Many commenters claimed that without
federal funds, and because of the tight budgets faced by state
governments presently, the funding of a near-road monitoring network
would result in less spending on other valuable environmental programs.

Response:  EPA understands the resource concerns of the states and
intends to work with states in identifying available funds and assess
the increased resource needs that may be needed for network
implementation.  This issue is discussed in section III.B.1 of the
preamble to the final rule. 

(10)	Comment:  Several commenters claim that EPA significantly
underestimates the cost of near-road monitors for many locations where
monitors are likely to be installed.

Response:  EPA used direct quotes or its best estimate for the variety
of capital and logistical costs that will be associated with
implementing the network design.  EPA must make its estimates as
nationally applicable as possible, and cannot anticipate the individual
variations in resource needs, or in the actual costs themselves, that
may occur from one location of the country to the next.  EPA intends to
work with states in identifying available funds and consider the
increased resource needs that may be needed for network implementation.

(11)	Comment:  State agencies requested that EPA provide additional
guidelines for the criteria that would be used by Regional
Administrators in determining when additional near-road monitors would
be required or when requests for additional monitors will be granted. 
An additional comment requested that EPA provide further guidance and
oversight of Regional Administrators to ensure the consistent
implementation of any near-road monitoring network across the country. 
State agencies also requested that the Regional Administrator be
required to coordinate with the state government in deciding where
near-road monitors would be placed.

Response:  EPA has provided multiple examples of situations where
Regional Administrator may use their discretion to require additional
monitors above the minimum required in section III.B.4 of the preamble
to the final rule.  In particular, EPA notes that in situations where a
Regional Administrator may consider the need for additional monitoring,
EPA expects that the state and the Regional Administrator would work
together to evaluate quantitative evidence that suggests an area may
warrant additional monitoring.  

(12)	Comment:  AASHTO recommended that state and local air monitoring
agencies be required to coordinate with their respective DOTs when
developing a near-road monitoring plan. 

Response:  Although can not require state and local air agencies to work
with another state or local entity, EPA believes that state and local
air monitoring agencies can greatly benefit from collaboration with
their counterparts in state or local transportation authorities.  Such
collaboration may help in site identification, site access, and ensuring
public and worker safety at near-road sites.

(13)	Comment:  Several State Agencies commented on the interaction
between NOX and O3.  For example, AASHTO requested that EPA clarify how
areas can overcome potential negative impacts on ambient O3 of increased
control of NOX emissions.  In contrast, one State Agency (HCPHES)
commented that a more stringent NO2 standard will help them attain the
O3 NAAQS because NO2 is a precursor to O3.  

Response:  It has long been recognized that reducing NOX emissions will
result in lower concentrations of regional O3 (National Research
Council, 1991).  This has been confirmed in numerous photochemical model
simulations and assorted ambient data analyses over the past two decades
(EPA 2005b; EPA, 2008d).  As a result, efforts to reduce NOX emissions
are the foundation of Federal and State actions to attain the ozone
NAAQS.  However, because of the complex chemistry of ozone formation,
there are some specific instances in which certain NOX reductions can
lead to localized ozone increases.  In these areas, the local planning
process will need to consider the multipollutant aspects of air quality
management and derive the most appropriate set of local control measures
needed to attain multiple NAAQS simultaneously. 

(15)	Comment: One state agency recommended that state and local DOTs
should be required to perform air quality impact analyses for NO2 prior
to road construction projects as well as conduct regular mobile monitor
testing to demonstrate that the construction projects are not causing
violations of the 1-hr NO2 standard.

Response:  Any such requirement for state and local DOT’s to perform
air quality impact analysis prior to road construction or to conduct any
testing would have to be established and carried out under state or
local authority.  

(16)	Comment:  One commenter encouraged EPA to develop a 1-hr NO2
significant impact level (SIL) without which, “any increase in NO2 in
a nonattainment area (or a near non-attainment area) could result in
denial of the proposed permit, even if the increase is not significant
or offset with offsite emission reduction credits” (CAAPCOA).

	Response:  This issue is discussed in section VI.D.2 of the preamble to
the final rule.

(17)	Comment:  The North Carolina Department of Environment and Natural
Resources (NCDENR) supports the use of a weighted annual mean (quarterly
average) for the annual primary standard.

Response:  This issue is discussed in the final notice (Section IV.A).

(18)	Comment:  SJVAPCD recommended that EPA include guidelines for
near-road monitoring that allows for monitoring to stop at locations
where sufficient data has been collected to conclude that measured
concentrations of NO2 are less than 85% of the standard level. 

Response: 40 CFR § 58.14(c),  explains the process by which states may
request approval for discontinuing monitoring at an individual site. 
The discontinuation is subject to Regional Administrator approval, and
requires the satisfaction of multiple requirements which are spelled out
in the regulation text.

(19)	Comment:  CASAC Panel members who supported the alternative
approach, as well as some public commenters (e.g., States) also raised
questions related to how non-attainment areas would be designated if
roadside monitors measure violations.  For example, some commenters
questioned whether an entire county could be out of attainment based on
a roadside monitor. 

	Response:  This comment is addressed in sections V.B and V.C of the
final notice.  

(20)	Comment:  The ALA, EJ, EDF, and NRDC also recommended that EPA
“formulate a rational basis for re-designation based on limited
monitoring data and available information on mobile source emissions
from the relevant roadways.”  Under this approach, EPA is encouraged
to “commit to designating near roadway areas no later than the end of
2013.”  These same groups also commented, “We believe it is
reasonable to…promulgate a nonattainment SIP submittal deadline of no
later than the end of 2015.”  Other environmental groups (e.g., CAC)
also commented that the timing for designations should be accelerated. 

	Response:   Section 110(d)(1)(B) of the CAA requires the EPA to
designate areas as 

attainment, nonattainment or unclassifiable no later than 2 years
following promulgation 

of a new or revised NAAQS (the CAA provides the Agency an additional
third year 

from promulgation should there be insufficient information  

on which to make compliance determinations).  The EPA intends to
finalize initial 

designations for the revised NO2 NAAQS in January 2012.  

A near-roadway monitoring network is not expected to be fully deployed
until January 

2013.  For this reason, EPA will proceed with initial designations using
air quality data 

from the existing, area-wide NO2 monitoring network in order to complete
designations 

by January 2012.  Once the near-roadway network is fully deployed and 3
years of air

quality data are available, the EPA intends to redesignate areas, as
appropriate, based on 

the most recent air quality data from the new monitoring network.  For
both initial 

designations and redesignations, EPA will use monitoring data to
identify violations of 

the standards.  EPA would then consider a variety of other factors in
determining which 

nearby areas contribute to a violation in setting the boundaries.      

The response to commenters’ requests that EPA both shorten and extend
SIP and 

attainment deadlines is provided in the implementation section of the
preamble.  

(21)	Comment:  One commenter (City of NY) recommended that EPA should
develop a screening approach to estimate short-term NO2 concentrations,
and provide source-type specific N02/NOX ratios with the NOX emission
factors. 

Response:  It is EPA's intention to utilize the current guidance and
policies to implement the revised NAAQS for NO2.  However, we will be
reviewing the need to provide additional technical as well as policy
guidance following the promulgation of the NAAQS.  

(22)	Comment:  API recommended that EPA “follow the precedent that was
set for PM2.5 by postponing the use of dispersion modeling to evaluate
short-term ambient NO2 concentrations until appropriate modeling tools
are developed, evaluated, peer-reviewed, and subjected to public comment
and review. 

	Response:  It is EPA's intention to utilize the current guidance and
policies to implement the revised NAAQS for NO2.  However, we will be
reviewing the need to provide additional technical as well as policy
guidance following the promulgation of the NAAQS. 

References

Atkinson, R., “Atmospheric Transformations of Automotive Emissions”.
In Air Pollution, the Automobile, and Public Health, A. Y. Watson, R. R.
Bates, and D. Kennedy, Ed. National Academies Press, ISBN:
0-309-56826-9, (1988). Available at:
http://www.nap.edu/catalog/1033.html

Bailey C.  (2009).  Dispersion Modeling for Mobile Source Applications,
Regional, State, and Local Air Modelers’ Workshop, Philadelphia, PA
–May 14, 2009.   Available at:   HYPERLINK
"http://www.cleanairinfo.com/regionalstatelocalmodelingworkshop/archive/
2009/presentations/06%20Thurs%20AM/Bailey_Dispersion%20Modeling%20for%20
Mobile%20Source%20Applications.pdf" 
http://www.cleanairinfo.com/regionalstatelocalmodelingworkshop/archive/2
009/presentations/06%20Thurs%20AM/Bailey_Dispersion%20Modeling%20for%20M
obile%20Source%20Applications.pdf .

Beardsley M, Warila J, Dolce G, Koupal J.  (2009).  Air Pollution
Emissions from Highway Vehicles: What MOVES Tells Us.  18th Annual
International Emission Inventory Conference, Baltimore, MD, April, 2009.

Beckerman B,  Jerrett M, Brook JR, Verma DK, Arain MA, Finkelstein MM. 
(2008). Correlation of nitrogen dioxide with other traffic pollutants
near a major expressway.  Atmos Environ.  42:275-290.

CARB.  (2003).  Characterizing the Range of Children's Pollutant
Exposure during School Bus Commutes.  Prepared for the California Air
Resources Board under Contract No. 00-322. October 10, 2003.

Carslaw DC.  (2008).  Evidence of an increasing NO2/NOx emissions ratio
from road traffic emissions.  Atmos Environ.  39:4793–4802.

Centers for Disease Control and Prevention (CDC). (2004) Introduction
and approach to causal inference. In: The health consequences of
smoking: a report of the Surgeon General. Atlanta, GA: U.S. Department
of Health and Human Services, Centers for Disease Control and
Prevention, National Center for Chronic Disease Prevention and Health
Promotion, Office on Smoking and Health. Available:
http://www.cdc.gov/tobacco/sgr/sgr_2004/chapters.htm (18 August, 2004).

Chock D.  (1977).  General motors sulfate dispersion experiment:
assessment of the EPA HIWAY model.  JAPCA.  27:39-45.

Delfino, RJ, Zeiger RS, Seltzer JM, Street DH, McLaren CE. (2002). 
Association of asthma symptoms with peak particulate air pollution and
effect modification by anti-inflammatory medication use.  Environ.
Health Perspect. 110:A607-A617.

Dieselnet.  (2008).  EU diesel market share at 53%.  22 February 2008. 
Available at:   HYPERLINK
"http://www.dieselnet.com/news/2008/02acea.php" 
http://www.dieselnet.com/news/2008/02acea.php . 

EPA (2004).  Air Quality Criteria Document for Particulate Matter. 
Accessible at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546

EPA (2005). Guidelines for Carcinogen Risk Assessment.  Accessible at  
HYPERLINK "http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=116283" 
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=116283 .  

EPA (2005a).  Review of the National Ambient Air Quality Standards for
Particulate Matter: Policy Assessment of Scientific and Technical
Information, OAQPS Staff Paper. Office of Air Quality Planning and
Standards, Research Triangle Park, NC.  Available at:   HYPERLINK
"http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221.pd
f" 
http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221.pdf
. 

EPA (2005b).  Technical Support Document for the Final Clean Air
Interstate Rule: Air Quality Modeling, RTP, NC, 285pp. 

EPA (2007a).  Review of the National Ambient Air Quality Standards for
Pb: Policy Assessment of Scientific and Technical Information. OAQPS
Staff paper.  Office of Air Quality Planning and Standards, Research
Triangle Park, NC.  EPA-452/R-07-013.  Available at:
http://www.epa.gov/ttn/naaqs/standards/pb/data/20071101_pb_staff.pdf 
HYPERLINK "http://epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_sp.html"   .

EPA.  (2007b).  Review of the National Ambient Air Quality Standards for
Ozone: Assessment of Scientific and Technical Information. OAQPS Staff
paper.  Office of Air Quality Planning and Standards, Research Triangle
Park, NC.  EPA-452/R-07-007a.  Available at:   HYPERLINK
"http://epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_sp.html" 
http://epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_sp.html .

EPA (2008a).  Integrated Science Assessment for Oxides of Nitrogen -
Health Criteria.  National Center for Environmental Assessment, Research
Triangle Park, NC.  Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=194645.

 

EPA (2008b).  Risk and Exposure Assessment to Support the Review of the
NO2 Primary National Ambient Air Quality Standard. Office of Air Quality
Planning and Standards, Research Triangle Park, NC. Available at: 

  HYPERLINK
"http://www.epa.gov/fedrgstr/EPA-AIR/2009/July/Day-15/a15944.pdf" 
http://www.epa.gov/fedrgstr/EPA-AIR/2009/July/Day-15/a15944.pdf 

EPA (2008c).  Risk and Exposure Assessment to Support the Review of the
NO2 Primary National Ambient Air Quality Standard: Appendices. 
EPA-452/R-08-008b.  

Available at:   HYPERLINK
"http://www.epa.gov/ttn/naaqs/standards/nox/data/20081121_NO2_REA_final_
appendices.pdf" 
http://www.epa.gov/ttn/naaqs/standards/nox/data/20081121_NO2_REA_final_a
ppendices.pdf .

EPA (2008d).  National Air Quality: Status and Trends thought 2007, RTP,
NC, 48pp.

Finlayson-Pitts and Pitts (1999)  Chemistry of the Upper and Lower
Atmosphere, Academic Press, 1999. 

Goodman SN. (1993).  p values, hypothesis tests, and likelihood:
implications for epidemiology of a neglected historical debate. American
Journal of Epidemiology 137:485–496. 

Goodman SN. (1999).  Towards evidence-based medical statistics. 1: the P
value fallacy. Annals of Internal Medicine 130:995–1004. 

Goodman, JE, Chandalia JK, Thakali S, Seeley M.  (2009).  Meta-analysis
of nitrogen dioxide exposure and airway hyper-responsiveness in
asthmatics.    HYPERLINK
"javascript:AL_get(this,%20'jour',%20'Crit%20Rev%20Toxicol.');" \o
"Critical reviews in toxicology."  Crit. Rev. Toxicol.  39:719-742.  

Greenland, S. (1998) Meta-analysis. In: Rothman, K. J.; Greenland, S.,
eds. Modern epidemiology.  Philadelphia, PA: Lippincott Williams &
Wilkins; pp. 643-673.

Health Effects Institute (2003).  Commentary on revised analyses of
selected studies.  In: Revised analyses of time-series studies of air
pollution and health.  Special report.  Boston, MA: Health Effects
Institute; pp. 255-290.

Hesterberg, TW, Bunn WB, McClellan RO, Hamade AK, Long CM, Valberg PA. 
(2009).  Critical review of the human data on short-term nitrogen
dioxide (NO2) exposures: evidence for NO2 no-effect levels.    HYPERLINK
"javascript:AL_get(this,%20'jour',%20'Crit%20Rev%20Toxicol.');" \o
"Critical reviews in toxicology."  Crit. Rev. Toxicol.  39(9):743-781. 

Hill, AB.  (1965).  The environment and disease: association or
causation? Proc. R. Soc. Med. 58: 295-300. 

Howson, C, Urbach P.  (1993).  Scientific Reasoning: The Bayesian
Approach. 2nd ed. Chicago: Open Court Publishing Company.

Institute of Medicine (IOM). (2007) Improving the presumptive disability
decision-making process for veterans. Washington, DC: The National
Academies Press. Available:
http://www.nap.edu/catalog.php?record_id=11908 (11 February, 2008).

Isakov V.  (2009).  Near-Road Modeling Research.  R/S/L Modelers
Workshop, June 2009.  Available at:   HYPERLINK
"http://www.cleanairinfo.com/regionalstatelocalmodelingworkshop/archive/
2009/presentations/06%20Thurs%20AM/Vlad%20RSL2009%205-14-2009.pdf" 
http://www.cleanairinfo.com/regionalstatelocalmodelingworkshop/archive/2
009/presentations/06%20Thurs%20AM/Vlad%20RSL2009%205-14-2009.pdf .

MACTEC.  (2004)  Sensitivity Analysis of PVMRM and OLM in AERMOD. 
Alaska DEC, Contract No. 18-8018-04.  Final Report.  September 2004. 
Available at:   HYPERLINK
"http://www.epa.gov/scram001/7thconf/aermod/pvmrm_sens.pdf" 
http://www.epa.gov/scram001/7thconf/aermod/pvmrm_sens.pdf . 

McCurdy TR.  1994.  Analysis of high 1 hour NO2 values and associated
annual averages using 1988-1992 data. Office of Air Quality Planning and
Standards, Research Triangle Park, NC. Available in Docket A-93-06.

Mortimer, KM, Neas LM, Dockery DW, Redline S, Tager IB. (2002).  The
effect of air pollution on inner-city children with asthma. Eur. Respir.
J. 19:699-705.

National Research Council, (1991). Rethinking the Ozone Problem in Urban
and Regional Air Pollution. National Academy Press, Washington, DC,
500pp. 

OLM/ARM Workgroup. (1998).  Use of the Ambient Ratio Method (ARM) for
Estimating Ambient Nitrogen Dioxide Concentrations.  Draft for Comment. 
May 27, 1998.  Available at:   HYPERLINK
"http://www.dec.state.ak.us/AIR/ap/docs/sitearm.pdf" 
http://www.dec.state.ak.us/AIR/ap/docs/sitearm.pdf . 

Orehek, J, Massari JP, Gayrard P, Grimaud C, Charpin J.  (1976).  Effect
of short-term low-level nitrogen dioxide exposure on bronchial
sensitivity of asthmatic patients.  J Clin Invest, 57:301-307. 

Poole, C.  (2001).  Causal values. Epidemiology 12:139–41.

Riediker M, Williams R, Devlin R, Griggs T, Bromberg P.  (2003). 
Exposure to particulate matter, volatile organic compounds, and other
air pollutants inside patrol cars.  Environ Sci Technol. 37:2084-2093.

Rizzo M.  2008.  Investigation of How Distributions of Hourly Nitrogen
Dioxide Concentrations Have Changed Over Time in Six Cities.  Memorandum
to the Nitrogen Dioxide NAAQS Review Docket EPA-HQ-OAR-2006-0922. 
Available at:   HYPERLINK
"http://www.epa.gov/ttn/naaqs/standards/nox/data/20081124_rizzo.pdf" 
http://www.epa.gov/ttn/naaqs/standards/nox/data/20081124_rizzo.pdf .

Roorda-Knape MC, Janssen NAH, De Hartog JJ, Van Vliet PHN, Harssema H,
Brunekreef B.  (1998).  Air Pollution from traffic in city districts
near major roadways.  Atmos Environ.  32(11)1921-1930.

Rothman, KJ, Greenland S., eds. (1998).  Modern epidemiology. 2nd ed.
Philadelphia, PA: Lippincott-Raven Publishers.

Royall, RM.  (1997).  Statistical Evidence: A Likelihood Paradigm.
London: Chapman and Hall.

Samet, JM. (2008).  Letter to EPA Administrator Stephen Johnson:
“Clean Air Scientific Advisory Committee’s (CASAC) Review Comments
on EPA’s Risk and Exposure Assessment to Support the Review of the NO2
Primary National Ambient Air Quality Standard.” EPA-CASAC-09-003,
December 16. 

Samet, JM. (2009a). CASAC Review of integrated science assessment for
particulate matter (second external review draft, July 2009). U.S.
Environmental Protection Agency. Washington, D.C.. EPA-CASAC-10-001. 

Samet, JM. (2009b).  Letter to EPA Administrator Lisa P. Jackson:
“Comments and Recommendations Concerning EPA's Proposed Rule for the
Revision of the National Ambient Air Quality Standards (NAAQS) for
Nitrogen Dioxide.” EPA-CASAC-09-014, September 9. 

Savitz, DA, Tolo KA, Poole C. (1994).  Statistical significance testing
in the American Journal of Epidemiology, 1970–1990. American Journal
of Epidemiology 139:1047–1052.

SCAQMD.  (2000).  Multiple Air Toxics Exposure Study in the South Coast
Air Basin (MATES-II).  Available at:   HYPERLINK
"http://www.aqmd.gov/matesiidf/matestoc.htm" 
http://www.aqmd.gov/matesiidf/matestoc.htm .

SCAQMD.  (2008).  Final Report: Multiple Air Toxics Exposure Study in
the South Coast Air Basin (MATES III).  Available at:   HYPERLINK
"http://www.aqmd.gov/prdas/matesIII/MATESIIIFinalReportSept2008.html" 
http://www.aqmd.gov/prdas/matesIII/MATESIIIFinalReportSept2008.html .

Schildcrout, JS, Sheppard L, Lumley T, Slaughter JC, Koenig JQ, Shapiro
GG.  (2006).   Ambient air pollution and asthma exacerbations in
children: an eight-city analysis.  Am J Epidemiol.  164:505-517.

Singer BC, Hodgson AT, Hotchi T, Kim JJ (2004).  Passive measurement of
nitrogen oxides to assess traffic-related pollutant exposure for the
East Bay Children’s Respiratory Health Study.  Atmos Environ. 
38:393-403.

US EPA.  (2008).  Emission and Air Quality Modeling Tools for
Near-Roadway Applications.  EPA/600/R-09/001.  Available at:   
HYPERLINK
"http://nepis.epa.gov/Exe/ZyNET.exe/P1003MVX.TXT?ZyActionD=ZyDocument&Cl
ient=EPA&Index=2006+Thru+2010&Docs=&Query=EPA%2F600%2FR-09%2F001&Time=&E
ndTime=&SearchMethod=3&TocRestrict=n&Toc=&TocEntry=&QField=pubnumber%5E%
22600R09001%22&QFieldYear=&QFieldMonth=&QFieldDay=&UseQField=pubnumber&I
ntQFieldOp=1&ExtQFieldOp=1&XmlQuery=&File=D%3A%5Czyfiles%5CIndex%20Data%
5C06thru10%5CTxt%5C00000007%5CP1003MVX.txt&User=ANONYMOUS&Password=anony
mous&SortMethod=h%7C-&MaximumDocuments=10&FuzzyDegree=0&ImageQuality=r75
g8/r75g8/x150y150g16/i425&Display=p%7Cf&DefSeekPage=x&SearchBack=ZyActio
nL&Back=ZyActionS&BackDesc=Results%20page&MaximumPages=1&ZyEntry=1&SeekP
age=x"  http://nepis.epa.gov/  

Vardoulakis, S, Gonzalez-Flesca N, Fisher BEA, Pericleous K.  (2004). 
Spatial variability of air pollution in the vicinity of a permanent
monitoring station in central Paris.  Atmos. Environ.  39:2725-2736.

Watkins, N. and Thompson, R. (2008). NOX Network Review and Background.
Memo to the NO2 NAAQS docket. 

Watson A, Bates R, Kennedy D, ed. (1988).  Air Pollution, the
Automobile, and Public Health, The National Academies Press, 1988.

Westerdahl D, Fruin S, Sax T, Fine PM, Sioutas C.  (2005).  Mobile
platform measurements of ultrafine particles and associated pollutant
concentrations on freeways and residential streets in Los Angeles. Atmos
Environ.  39:3597-3610.

  It should be noted that use of smaller data subsets will markedly
reduce statistical power to detect associations and also reduce the
precision on any findings when the data are stratified by season.

 hÿ

Š

 hÿ



 

/

0

1

2

 hÿ

 hÿ

 hÿ

 hÿ

#2

4

5

g

h

i

ƒ

„

…

‡

ˆ

‰

Š

‹

Œ

¨

©

ª

«



®

â

ã

ä

þ

ÿ

 hÿ

 hÿ

 hÿ

 hÿ

	

.

/

- hÿ

 hÿ

 hÿ

 hÿ

%/

0

1

3

4

=

>

?

Y

Z

[

]

^

_

`

a

b

~



€

ƒ

„

’

“

”

®

¯

°

²

³

´

µ

¶

·

Ó

 hÿ

 hÿ

 hÿ

 hÿ

`

µ

ᴇഀࣆȀь⒆਀ᰇഀࣆȀͰ⒆਀᠀Ó

Ô

Õ

Ö

Ø

Ù

Ý

Þ

ß

ù

ú

û

ý

þ

ÿ

j;

 hÿ

 hÿ

 hÿ

 hÿ

j/

 hÿ

 hÿ

j¸

 hÿ

 hÿ

 hÿ

 hÿ

j¬

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

j

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

- hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÿ

 hÛ

 h	

h	

  h	

 h	

h	

  h	

  h	

 h	

 h	

h	

  h	

  h	

 h	

  h	

 h	

  h	

  h	

# h	

  h	

  h	

 h	

 h	

h	

h	

 h	

h	

  h	

 h	

  h	

h

H* h€

 h

 

퀁萏ː萑ﴰ葞ː葠ﴰ摧ᔜß

 h

 

h¤5

 h

 

 hš*

hús

 h

¤

¥

¦

^

_

¦

§

¨

ª

²

·

Ñ

â

í

 h

_

i

m

摧⾕Ä

ༀ킄ᄂや㟽$␸䠀$葞ː葠ﴰ摧⾕Ä

 h

 h

 h_

쀁ȡ萏ː葞ː摧⾕Ä

萏ː萑ﴰ葞ː葠ﴰ摧僿ó

ༀ킄ᄂや㟽$␸䠀$葞ː葠ﴰ摧⾕Ä

 h

 h

 h³

萏ː萑ﴰ葞ː葠ﴰ摧⾕Ä	㜀$␸䠀$摧⾕Ä

ༀ킄ᄂや㟽$␸䠀$葞ː葠ﴰ摧䰔$

h

摧ᣃ0	㜀$␸䠀$摧ᣃ0	㜀$␸䠀$摧ഗí	㜀$␸䠀$摧洵

 h	

h	

摧ḷð



%

&

q

s

w

®

¯

Û

Ü

 hÀ

hù

ࠀ萏ː葞ː摧㞝sࠀ萏ː葞ː摧织ðఀ萎ː萏֠葝ː葞֠摧ⓢ|

 hrb

H*  hrb

 hrb

 hrb

hù

 hù

hù

hù

h"

hE

 h	

਀&䘋
옍ćŨ㠁؄萏и葞и摧姴ఀ萏ː萑ﴰ葞ː葠ﴰ摧姴᐀

hE

hE

hE

.

h"

ༀ킄ᄂ킄㜂$␸䠀$葞ː葠ː摧ӹÊ

 h	

萏ː葞ː摧ᓖQ

 hÙ

%

☊଀ెᄀや㟽$␸䠀$葠ﴰ摧ᓖQ

 

!

$

'

0

Z

[

Ñ

Ó

Ô

‹

Œ

.%

&

Œ

Ä

ༀ킄㜂$␸䠀$葞ː摧湽

摧湽

萏ː葞ː摧湽

਀&䘋	옍ćŨ㠁؄萏и葞и摧湽

萑ﴰ葠ﴰ摧湽

ༀ킄ᄂや㟽$␸䠀$葞ː葠ﴰ摧ᓖQကŒ

›

œ

§

¨

°

±

»

¼

Â

Ä

Ì

h}n

h}n

h}n

h}n

 h	

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

萏ː萑ﴰ葞ː葠ﴰ摧湽

摧湽

☊଀ๆᄀや㟽$␸䠀$葠ﴰ摧湽

萏ː葞ː摧湽

ༀ킄㜂$␸䠀$葞ː摧湽

฀ꂄ༅ꂄ㜅$␸䠀$葝֠葞֠摧湽

h}n

h}n

h}n

h}n

h}n

h}n

h}n

 h	

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

h}n

萑ː葠ː摧湽

萏ː萑ﴰ葞ː葠ﴰ摧湽

摧湽

萏ː葞ː摧湽

h}n

h}n

H* h"

h}n

h}n

h}n

 h	

h}n

h}n

h}n

h}n

 h

h

摧⎚Þ

摧㩢ó

 h"

h

÷òîæîáîáîáîæîáîáîáîÝÕîÍÅººÅººî¦ÕžÍºº
žºžº“†“†“Ý h"

 h"

  h"

h"

 h"

h"

萏ː萑ﴰ葞ː葠ﴰ摧澮Ä

h"

 h"

 h"

H* h"

 h"

 h"

  h"

 h"

 h"

 h"

! h"

 h"

  h"

h"

 h"

 h"

 h"

h"

 h"

 h"

 h"

 h"

 h"

 h"

h»f

퀁摧暻

萑ː葠ː摧祧é

OJ

QJ

^J

.

/

`

a

.

/

7

:

¢

£

_

b

m

Õ

Ö

 hÛ

 hÛ

. EPA has addressed NAM's concerns in the final notice and this RTC. EPA
has fully satisfied the requirements of the Information Quality Act, and
has acted consistently with EPA's Information Quality Guidelines. 

 There were no upper percentile values less than the 95th percentile
determined as statistically significant.

 Only Los Angeles (2001-2003), Other CMSA/MSA (2001-2003), and Provo
(2004-2006) had as is air quality more similar with a 1-hour 99th
percentile 100 ppb standard level out of a total of 38
location/year-group combinations.  

 When considering the two forms (98th and 99th), the 19 locations, and
two year-groups (2001-2003 and 2004-2006), 61/76=80% needed a downward
adjustment to meet the 50 ppb level. 

 Note that to generate the ratio, the m factor has a value of 1 added to
it.  See REA equation (7-2).

To measure maximum concentrations, the Administrator proposed monitoring
provisions that would require monitors within 50 meters of major roads
and to allow the Regional Administrator to require additional monitors
in situations where maximum concentrations would be expected to occur in
locations other than near major roads (e.g., due to the influence of
multiple smaller roads and/or stationary sources). 

 The most current American Housing Survey (  HYPERLINK
"http://www.census.gov/hhes/www/housing/ahs/ahs.html" 
http://www.census.gov/hhes/www/housing/ahs/ahs.html ) is from 2007 and
lists a higher fraction of housing units within the 300 foot boundary. 
According to Table 1A-6 from that report
(http://www.census.gov/hhes/www/housing/ahs/ahs07/tab1a-6.pdf), out of
128.2 million total housing units in the United States, about 20 million
were reported by the surveyed occupant or landlord as being within 300
feet of a 4-or-more lane highway, railroad, or airport.  That
constitutes 15.6% of the total housing units in the U.S.  Assuming equal
distributions, with a current population of 306.3 million, that means
that there would be 47.8 million people meeting the 300 foot criteria.

 These groups also cited a recent study by Karr et al. (Influence of
Ambient Air Pollutant Sources on Clinical Encounters for Infant
Bronchiolitis. Am. J. Respir. Crit. Care. Med. (2009), Aug. 27, epub
ahead of print).  

