Notice: This opinion is subject to formal revision before publication in the
Federal Reporter or U.S.App.D.C. Reports. Users are requested to notify
the Clerk of any formal errors in order that corrections may be made
before the bound volumes go to press.
United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
–————
No. 03-1003 September Term, 2003
Filed On: July 30, 2004
BLUEWATER NETWORK,
PETITIONER
v.
ENVIRONMENTAL PROTECTION AGENCY AND
CHRISTINE TODD WHITMAN, ADMINISTRATOR, U.S. ENVIRONMENTAL
PROTECTION AGENCY,
RESPONDENTS
INTERNATIONAL SNOWMOBILE MANUFACTURERS ASSOCIATION,
INTERVENOR
Consolidated with
03-1004, 03-1005, 03-1249
–————
BEFORE: EDWARDS, SENTELLE, and TATEL, Circuit Judges.
ORDER
Upon consideration of respondents’ motion to clarify re-
mand, and the opposition thereto, it is
ORDERED that the motion to clarify remand be granted.
It is
FURTHER ORDERED that the opinion filed June 1,
2004, be amended as follows:
2
Page 34, delete the last paragraph of the opinion, and
insert in lieu thereof:
For the reasons set forth above, we grant in part
and deny in part each of the two petitions for review.
We hereby vacate the NOx standard on the ground
that EPA lacks statutory authority to regulate NOx
emissions under § 213(a)(4). We hold that the
Agency acted within its statutory authority in pro-
mulgating the CO and HC standards under
§ 213(a)(3) and (a)(4), respectively. However, we
remand the CO and HC standards for EPA to
clarify the analysis and evidence upon which the
standards are based. Specifically, we direct EPA to
clarify (1) the statutory and evidentiary basis of the
Agency’s assumption that the standards must be
sufficiently lenient to permit the continued produc-
tion of all existing snowmobile models, and (2) the
analysis and evidence underlying the Agency’s con-
clusion that advanced technologies can be applied to
no more than 70% of new snowmobiles by 2012. We
uphold the CO and HC standards in all other re-
spects.
Per Curiam
FOR THE COURT:
Mark J. Langer, Clerk
BY:
Michael C. McGrail
Deputy Clerk
Notice: This opinion is subject to formal revision before publication in the
Federal Reporter or U.S.App.D.C. Reports. Users are requested to notify
the Clerk of any formal errors in order that corrections may be made
before the bound volumes go to press.
United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued April 12, 2004 Decided June 1, 2004
No. 03-1003
BLUEWATER NETWORK,
PETITIONER
v.
ENVIRONMENTAL PROTECTION AGENCY AND
CHRISTINE TODD WHITMAN, ADMINISTRATOR, U.S. ENVIRONMENTAL
PROTECTION AGENCY,
RESPONDENTS
INTERNATIONAL SNOWMOBILE MANUFACTURERS ASSOCIATION,
INTERVENOR
Consolidated with
03-1004, 03-1005, 03-1249
On Petitions for Review of an Order of the
Environmental Protection Agency
James S. Pew argued the cause for petitioners Bluewater
Network and Environmental Defense. With him on the
briefs was Jennifer R. Kefer.
Bills of costs must be filed within 14 days after entry of judgment.
The court looks with disfavor upon motions to file bills of costs out
of time.
2
Eric B. Wolff argued the cause for petitioner International
Snowmobile Manufacturers Association. With him on the
briefs were Stuart A. Drake and Granta Y. Nakayama.
Stephen E. Crowley and Kent E. Hanson, Attorneys, U.S.
Department of Justice, argued the cause for respondents.
With them on the brief were John C. Cruden, Deputy Assis-
tant Attorney General, and Michael J. Horowitz, Attorney,
U.S. Environmental Protection Agency.
James S. Pew and Jennifer R. Kefer were on the brief for
intervenors Bluewater Network and Environmental Defense.
Stuart A. Drake, Granta Y. Nakayama, and Eric B. Wolff
were on the brief for intervenor International Snowmobile
Manufacturers Association.
Before: EDWARDS, SENTELLE, and TATEL, Circuit Judges.
Opinion for the Court filed by Circuit Judge EDWARDS.
EDWARDS, Circuit Judge: In November 2002, the Environ-
mental Protection Agency (‘‘EPA’’ or ‘‘Agency’’) issued a final
rule establishing emissions standards for snowmobiles and
certain other ‘‘nonroad’’ vehicles. See Control of Emissions
From Nonroad Large Spark–Ignition Engines, and Recre-
ational Engines (Marine and Land–Based), 67 Fed. Reg.
68,242 (Nov. 8, 2002). The snowmobile standards at issue in
this case – promulgated under § 213 of the Clean Air Act
(‘‘CAA’’ or ‘‘Act’’), 42 U.S.C. § 7547 (2000) – regulate emis-
sions of three pollutants: carbon monoxide (‘‘CO’’), hydrocar-
bons (‘‘HC’’), and oxides of nitrogen (‘‘NOx’’).
The CO standard was adopted under § 213(a)(3). Under
this provision, EPA must regulate CO and certain ozone-
precursor emissions from a category of engines if, and only if,
the Agency finds that such emissions ‘‘cause, or contribute to’’
CO or ozone concentrations in more than one area that has
failed to attain the relevant national ambient air quality
standard (‘‘NAAQS’’). Where the Agency makes such a
finding – as it did for snowmobiles with respect to CO
3
emissions – it must adopt standards reflecting ‘‘the greatest
degree of emission reduction achievable’’ through the applica-
tion of technology that ‘‘will be available,’’ taking cost and
other factors into account.
EPA regulated HC and NOx emissions under § 213(a)(4),
which is directed at pollution problems other than CO and
ozone. This provision authorizes EPA – upon making certain
findings – to adopt such standards as the Agency ‘‘deems
appropriate,’’ again based on technology that will be available
and taking cost and other factors into account. Of crucial
importance for this case, § 213(a)(4) only permits regulation
of ‘‘emissions not referred to in’’ § 213(a)(2), which expressly
mentions emissions of CO, volatile organic compounds, and
NOx.
The Agency based its standards on the expected application
of two ‘‘advanced’’ technologies to snowmobiles: direct injec-
tion two-stroke engines and four-stroke engines. EPA esti-
mated that compliance with the final phase of its standards –
effective in 2012 – would require the use of these engines in
70% of all new snowmobiles. The Agency found that broader
application would not be possible by 2012, because of resource
constraints on manufacturers and the magnitude of the in-
vestment required to apply the technologies to the wide
variety of snowmobile models on the market.
Petitioner International Snowmobile Manufacturers Associ-
ation (‘‘ISMA’’) challenges EPA’s authority to promulgate the
standards. ISMA argues that EPA lacks authority to issue
the CO standard, because the Agency’s finding that snowmo-
biles contribute to CO pollution in more than one area that
has failed to attain the NAAQS is based on an impermissible
interpretation of the statute and is arbitrary and capricious.
ISMA claims, in addition, that the statute bars EPA from
regulating HC and NOx emissions under § 213(a)(4), because
those emissions are ‘‘referred to’’ in § 213(a)(2).
Petitioners Bluewater Network and Environmental De-
fense (collectively ‘‘Bluewater’’) challenge what they consider
to be the excessive leniency of the standards. Bluewater’s
principal claim is that EPA’s determination that advanced
4
technologies cannot be applied to all new snowmobiles by
2012 is premised on an impermissible interpretation of the
statute and is arbitrary and capricious. Bluewater also raises
a host of other challenges to the regulation, including the
claim that EPA improperly refused to base its standards on
the application of catalyst technology.
We grant in part and deny in part each of the two petitions
for review. First, we hold that EPA acted within its statuto-
ry authority in promulgating the CO and HC standards under
§ 213(a)(3) and (a)(4), respectively. Accordingly, we reject
ISMA’s challenges to those standards. However, we agree
with ISMA that EPA lacks authority to regulate NOx emis-
sions under § 213(a)(4), because such emissions are ‘‘referred
to’’ in § 213(a)(2). We therefore vacate the NOx standard.
In response to Bluewater’s petition, we remand the CO and
HC standards for EPA to clarify the analysis and evidence
upon which the standards are based. Specifically, we direct
EPA to clarify (1) the statutory and evidentiary basis of the
Agency’s assumption that the standards must be sufficiently
lenient to permit the continued production of all existing
snowmobile models, and (2) the analysis and evidence under-
lying the Agency’s conclusion that advanced technologies can
be applied to no more than 70% of new snowmobiles by 2012.
We reject Bluewater’s remaining claims.
I. BACKGROUND
A. Factual Background
The snowmobile industry is relatively concentrated, with
four manufacturers producing 99% of all snowmobiles, or
‘‘sleds,’’ sold in the United States. These manufacturers offer
various types of sleds designed for different applications –
including high-performance trail riding, high-performance off-
trail riding, mountain riding, touring, and entry-level riding –
with multiple engine models available for each type. As a
result, most of the major manufacturers offer 30 to 50 differ-
ent engine-snowmobile model combinations. High-
performance models, with very high power-to-weight ratios,
dominate current sales. See 67 Fed. Reg. at 68,273.
5
The vast majority of snowmobiles now on the market use
carbureted two-stroke engines. In comparison with four-
stroke engines, carbureted two-stroke engines generally are
simpler in design and have lower manufacturing costs. They
also burn an air-fuel mixture that is comparatively rich in
fuel. This makes them less fuel-efficient than four-stroke
engines, but gives them a higher power-to-weight ratio, allows
them to start more easily in cold weather, and permits them
to run at cooler temperatures (which reduces engine wear) –
all important advantages for snowmobiles. See 65 Fed. Reg.
76,797, 76,803–04 (Dec. 7, 2000) (advance notice of proposed
rulemaking).
Because of their design characteristics, carbureted two-
stroke engines emit comparatively high levels of CO and HC,
see id., both of which can contribute to harmful air pollution.
Elevated CO levels can cause a number of health problems
associated with reduced delivery of oxygen to the body’s
tissues, including impairment of visual perception, work ca-
pacity, manual dexterity, learning ability, and performance of
complex tasks. 67 Fed. Reg. at 68,245. HC emissions can,
inter alia, cause visibility impairment (or ‘‘haze’’) due to fine
particulate matter (‘‘PM’’) pollution; specifically, HC emis-
sions contain fine PM and can also contribute to the forma-
tion of ‘‘secondary’’ fine PM in the atmosphere. Id. at 68,254.
Like virtually all internal combustion engines, snowmobile
engines emit volatile organic compounds (‘‘VOCs’’) – most of
which are hydrocarbons – and NOx. VOCs and NOx are the
primary precursors of ground-level ozone, which can cause a
number of severe respiratory problems. 65 Fed. Reg. at
76,798. Ground-level ozone is formed through a complex
chemical reaction of VOCs and NOx in the atmosphere.
Because this reaction occurs only in the presence of heat and
sunlight, elevated ground-level ozone concentrations are pri-
marily a warm-weather phenomenon. See id.
B. Statutory Context
Recognizing the significant and growing role of unregulated
emissions from ‘‘nonroad’’ engines in causing air pollution,
Congress enacted § 213 of the Clean Air Act as part of the
6
1990 amendments to the Act. See Pub. L. No. 101–549,
§ 222, 104 Stat. 2399, 2500–02 (1990) (codified at 42 U.S.C.
§ 7547); see also S. REP. No. 101–228, at 103–04 (1989)
(discussing the policy rationale for regulating nonroad engine
emissions). Section 213 authorizes EPA to set emissions
standards for ‘‘nonroad engines and vehicles,’’ a broad group-
ing including farm and construction equipment, lawn and
garden equipment, airport service equipment, marine engines,
and recreational vehicles such as off-road motorcycles, all-
terrain vehicles, and snowmobiles.
Under § 213’s multi-step scheme, EPA must first complete
a study to determine whether emissions from nonroad en-
gines ‘‘cause, or significantly contribute to, air pollution which
may reasonably be anticipated to endanger public health or
welfare.’’ 42 U.S.C. § 7547(a)(1). Based on the results of
this study, EPA must then determine whether emissions of
CO, NOx, and VOCs from new and existing nonroad vehicles
or engines collectively are ‘‘significant contributors to ozone
or carbon monoxide concentrations in more than 1 area which
has failed to attain the national ambient air quality standards
[NAAQS] for ozone or carbon monoxide.’’ 42 U.S.C.
§ 7547(a)(2). (For convenience, we refer to areas which have
failed to attain the NAAQS as ‘‘nonattainment areas.’’)
If EPA makes a finding of significant contribution for
nonroad engines under § 213(a)(2), the Agency is required
under § 213(a)(3) to promulgate standards for those individu-
al ‘‘classes or categories’’ of new nonroad engines whose
emissions, in EPA’s judgment, ‘‘cause, or contribute to’’ CO
or ozone concentrations in more than one CO or ozone
nonattainment area. 42 U.S.C. § 7547(a)(3). These stan-
dards must
achieve the greatest degree of emission reduc-
tion achievable through the application of tech-
nology which [EPA] determines will be available
TTT, giving appropriate consideration to the cost
of applying such technology within the period of
time available to manufacturers and to noise,
7
energy, and safety factors associated with the
application of such technology.
Id. In setting these standards, EPA is directed to ‘‘first
consider standards equivalent in stringency to standards for
comparable [onroad] motor vehicles or engines (if any)’’ regu-
lated under § 202 of the Act. Id.
Section 213(a)(4) provides EPA with an alternative basis of
regulatory authority. Under this provision, if EPA deter-
mines that nonroad engine emissions ‘‘not referred to’’ in
§ 213(a)(2) ‘‘significantly contribute to air pollution which
may reasonably be anticipated to endanger public health or
welfare,’’ EPA may promulgate emissions standards it
‘‘deems appropriate’’ for individual categories or classes of
nonroad engines that EPA determines ‘‘cause, or contribute
to, such air pollution.’’ 42 U.S.C. § 7547(a)(4). These stan-
dards likewise must be based on technology that EPA deter-
mines will be available, giving appropriate consideration to
cost, noise, energy, and safety factors. Id.
C. Regulatory History
In November 1991, EPA completed the ‘‘Nonroad Engine
and Vehicle Emission Study’’ called for in § 213(a)(1). Based
on this study, EPA made a final determination pursuant to
§ 213(a)(2) that emissions from nonroad engines significantly
contribute to CO and ozone concentrations in more than one
nonattainment area. 59 Fed. Reg. 31,306, 31,307 (June 17,
1994). EPA further found, under § 213(a)(4), that PM emis-
sions from nonroad engines significantly contribute to air
pollution that may reasonably be anticipated to endanger
public health or welfare. See id. at 31,318.
On December 7, 2000, EPA issued a final finding under
§ 213(a)(3) that emissions from the category of large spark-
ignition (‘‘large-SI’’) engines and the lesser included category
of land-based recreational vehicles (which includes snowmo-
biles, offroad motorcycles, and all-terrain vehicles) each con-
tribute to ozone and CO concentrations in more than one
nonattainment area. 65 Fed. Reg. 76,790, 76,791 (Dec. 7,
2000). EPA also found, pursuant to § 213(a)(4), that the
8
large-SI and land-based recreational vehicle categories each
contribute to PM air pollution that may reasonably be antici-
pated to endanger the public health or welfare. Id. EPA
indicated that it would consider further whether it should
regulate snowmobile emissions of ozone precursors, because
ozone is less of a concern during cold weather, when snowmo-
biles are used. Id.
On November 8, 2002, EPA promulgated a final rule estab-
lishing emissions standards for large-SI engines and land-
based recreational engines, including snowmobiles. 67 Fed.
Reg. 68,242 (Nov. 8, 2002) (‘‘Final Rule’’). Only the snowmo-
bile exhaust emissions standards are at issue in this case.
The Final Rule established a CO emissions standard for
snowmobiles pursuant to § 213(a)(3). In discussing the need
for this standard, EPA referred to its December 2000 finding
that the land-based recreational vehicle category – in which
snowmobiles are included – contributes to CO concentrations
in more than one CO nonattainment area. Id. at 68,248.
EPA further found that, even when considered separately
from other land-based recreational vehicles, snowmobiles con-
tribute to such pollution. Id. at 68,248–49. The Agency
determined that regulation of snowmobile emissions of ozone
precursors under § 213(a)(3) was inappropriate, because
snowmobiles are operated in cold weather and are therefore
unlikely to contribute to ozone pollution. See Summary and
Analysis of Comments (‘‘SAC’’) at II–24, V–31, reprinted in
Joint Appendix (‘‘J.A.’’) 92, 110; see also 66 Fed. Reg. 51,098,
51,154 (Oct. 5, 2001) (notice of proposed rulemaking).
EPA promulgated HC and NOx emissions standards under
§ 213(a)(4). The Agency found that snowmobile emissions
contribute significantly to haze in a number of relatively
pristine protected areas, known under the CAA as ‘‘Class I’’
areas, including at least eight national parks. 67 Fed. Reg. at
68,252–54. This phenomenon is the result of increased am-
bient concentrations of fine PM. Id. EPA offered two
grounds for regulating snowmobile HC emissions as a means
9
of controlling fine PM pollution. First, HC emissions them-
selves contain fine PM and contribute to the formation of
secondary fine PM in the atmosphere. Id. at 68,254. Sec-
ond, EPA determined that HC emissions provide a good
proxy for regulating fine PM emissions from snowmobiles,
because the technologies for reducing HC emissions also
reduce PM emissions and direct regulation of PM is more
difficult. Id. Although the rule is unclear on this point, EPA
appeared to base its authority to regulate NOx emissions
under § 213(a)(4) on its finding that NOx contributes to the
formation of secondary PM. See id. at 68,254 n.30, 68,255.
EPA based its emissions standards on two ‘‘advanced’’
technologies that it determined would be available for snow-
mobiles in the foreseeable future: (1) direct injection (‘‘DI’’)
two-stroke engines, which replace air-fuel carburetion with
direct injection of fuel into the cylinder, and (2) four-stroke
engines. Id. at 68,272. The Agency predicted that DI two-
stroke engines could reduce HC emissions by 70–75% and CO
emissions by 50–70%. Four-stroke engines could reduce HC
emissions by 70–95%, and could reduce CO emissions by 50–
80% for low-power applications and 20–50% for high-power
applications. Id. EPA did not view either technology as
obviously superior. DI two-stroke engines would likely pro-
duce lower CO emissions than comparably powered four-
stroke engines, but four-stroke engines would yield greater
reductions in HC emissions. Four-stroke engines would like-
ly produce more pure power, whereas DI two-stroke technol-
ogy might be preferable for applications requiring a powerful,
but lighter and more compact, engine. Id.
In setting emissions standards, EPA framed the regulatory
question as ‘‘how broadly [these] technolog[ies] can be prac-
tically applied across the snowmobile fleet in the near term,
taking into account factors such as the number of engine and
snowmobile models currently available, and the capacity of
the industry to perform the research and development efforts
required to optimally apply advanced technology to each of
these models.’’ Id. at 68,273. EPA concluded that, ‘‘at least
in theory,’’ there was no purely technological barrier to the
application of these technologies to all new snowmobiles by
2012. Id. However, the Agency identified a number of factors
10
that would limit the speed with which such technologies could
be applied to all snowmobiles models, including resource
constraints on manufacturers, the fact that not all manufac-
turers produce their engines in-house, and the design and
development work required to optimize advanced technologies
for each model. Id.
Taking these factors into account, the final rule requires
that snowmobile engines meet successively more stringent
emissions standards in three phases. In Phase 1, manufac-
turers would be required to reduce CO and HC emissions by
30% relative to current baseline emissions. Half of all snow-
mobile sales would have to meet the Phase 1 standards by
model year 2006, and all would have to meet them by model
year 2007. EPA estimated that compliance with the Phase 1
standards would require application of advanced technologies
to approximately 10% of new snowmobiles, with cleaner car-
buretion and other technologies applied to the remainder.
Id. at 68,271.
In Phase 2, effective for the 2010 model year, manufactur-
ers would be required to achieve a 50% reduction in HC
emissions relative to baseline and a 30% reduction in CO
emissions relative to baseline. Id. at 68,273. In Phase 3,
effective for the 2012 model year, manufacturers must achieve
a nominal 50% reduction relative to baseline for both CO and
HC. This standard requires percentage reductions in CO
and HC that together add up to 100%, e.g., 60% for HC and
40% for CO. However, emissions for each pollutant may not
exceed the level permitted under the Phase 2 standards. Id.
at 68,274. EPA predicted that the Phase 2 and 3 standards
would require application of advanced technologies to 50%
and 70% of new snowmobiles, respectively, with less advanced
technologies applied to the remainder. Id. at 68,271, 68,273.
The Phase 3 standards also require engines to meet a NOx
standard (actually a HC v NOx standard), which caps NOx
emissions at or near existing levels. Id. at 68,274.
The three-phase scheme is summarized in the following
table:
11
EPA noted that it believed that it would be feasible at some
point after 2012 to apply advanced technologies to all new
snowmobiles and that catalysts or other exhaust aftertreat-
ment devices might become available at some future time.
The Agency stated that it had considered setting a standard
reflecting application of advanced technologies to 100% of new
snowmobiles, but did not believe that this was feasible by
2012. EPA indicated that in the future it would consider
promulgating more stringent standards to be applied in a
fourth phase. The Agency declined to do so in this rulemak-
ing, however, in order to monitor the development of new
technologies in response to the Phase 1 through 3 standards.
Id.
Bluewater and ISMA filed separate petitions for review of
the snowmobile standards, and the petitions were consolidat-
ed in the instant case.
II. ANALYSIS
A. Standard of Review
Under § 307(d)(9) of the Clean Air Act, we review the
Final Rule to determine, inter alia, whether EPA’s action is
‘‘in excess of statutory jurisdiction, authority, or limitations,
or short of statutory right,’’ or is ‘‘arbitrary, capricious, an
abuse of discretion, or otherwise not in accordance with law.’’
42 U.S.C. § 7607(d)(9); see also Ethyl Corp. v. EPA, 51 F.3d.
1053, 1064 (D.C. Cir. 1995) (noting that review under the
CAA’s ‘‘arbitrary and capricious’’ standard is the same as that
12
required by the Administrative Procedure Act, 5 U.S.C.
§ 706(2)(A)). Our review under the ‘‘arbitrary and capri-
cious’’ standard is narrow and does not permit us to substi-
tute our policy judgment for that of the Agency. Rather, we
are principally concerned with ensuring that EPA has ‘‘exam-
ine[d] the relevant data and articulate[d] a satisfactory expla-
nation for its action including a rational connection between
the facts found and the choice made,’’ that the Agency’s
‘‘decision was based on a consideration of the relevant fac-
tors,’’ and that the Agency has made no ‘‘clear error of
judgment.’’ Motor Vehicle Mfrs. Ass’n v. State Farm Mut.
Auto. Ins. Co., 463 U.S. 29, 43 (1983) (internal citations and
quotation marks omitted).
We review EPA’s interpretations of the CAA under the
standard established by Chevron U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 467 U.S. 837 (1984), and its
progeny. Under this standard, we first employ the tradition-
al tools of statutory construction to determine whether Con-
gress has spoken to the precise question at issue. Id. at 842–
43 & n.9. ‘‘If the intent of Congress is clear, that is the end
of the matter; for the court, as well as the Agency, must give
effect to the unambiguously expressed intent of Congress.’’
Id. at 842–43. However, where the statute is ambiguous and
the Agency has acted within its delegated authority, we will
defer to the Agency’s interpretation if it is reasonable. Mo-
tion Picture Ass’n of Am., Inc. v. FCC, 309 F.3d 796, 801
(D.C. Cir. 2002) (citing Chevron, 467 U.S. at 843–44). Such
deference is appropriate only where the agency acts pursuant
to an express or implied congressional delegation of authority
to regulate in the area at issue and the Agency’s action has
the ‘‘force of law.’’ Id. (citing United States v. Mead Corp.,
533 U.S. 218, 226–27 (2001)).
B. EPA’s Authority to Regulate CO Under § 213(a)(3)
Section 213(a)(3) requires EPA to regulate CO emissions
from an individual ‘‘class or category’’ of nonroad engines if,
and only if, the Agency first determines that such emissions
‘‘cause, or contribute to’’ CO concentrations in more than one
area that has failed to attain the NAAQS for CO. 42 U.S.C.
13
§ 7547(a)(3). In the Final Rule, EPA based its authority to
issue the snowmobile CO standard on two such ‘‘contribution’’
findings: the December 2000 finding for the land-based re-
creational vehicle category in which snowmobiles are includ-
ed, and the Final Rule’s finding for snowmobiles considered
as a separate category.
ISMA challenges EPA’s authority to issue the CO standard
under § 213(a)(3) on three grounds, arguing (1) that EPA’s
contribution findings are premised on an impermissible inter-
pretation of the ‘‘cause, or contribute to’’ language in
§ 213(a)(3); (2) that the snowmobiles-only CO contribution
finding in the Final Rule is arbitrary and capricious; and (3)
that EPA improperly grouped snowmobiles with land-based
recreational vehicles for purposes of its December 2000 con-
tribution finding. We reject each of these claims and, accord-
ingly, uphold EPA’s authority to promulgate the CO stan-
dard.
1. EPA’s Interpretation of ‘‘Contribute’’
The relevant portions of the Act read as follows:
(a) Emissions standards
(1) The Administrator shall conduct a study of
emissions from nonroad engines and nonroad
vehicles (other than locomotives or engines used
in locomotives) to determine if such emissions
cause, or significantly contribute to, air pollution
which may reasonably be anticipated to endan-
ger public health or welfare. Such study shall be
completed within 12 months of November 15,
1990.
(2) After notice and opportunity for public hear-
ing, the Administrator shall determine within 12
months after completion of the study under
paragraph (1), based upon the results of such
study, whether emissions of carbon monoxide,
oxides of nitrogen, and volatile organic com-
pounds from new and existing nonroad engines
or nonroad vehicles (other than locomotives or
engines used in locomotives) are significant con-
tributors to ozone or carbon monoxide concen-
14
trations in more than 1 area which has failed to
attain the national ambient air quality standards
for ozone or carbon monoxide. Such determina-
tion shall be included in the regulations under
paragraph (3).
(3) If the Administrator makes an affirmative
determination under paragraph (2) the Adminis-
trator shall, within 12 months after completion
of the study under paragraph (1), promulgate
(and from time to time revise) regulations con-
taining standards applicable to emissions from
those classes or categories of new nonroad en-
gines and new nonroad vehicles (other than
locomotives or engines used in locomotives)
which in the Administrator’s judgment cause, or
contribute to, such air pollution. Such standards
shall achieve the greatest degree of emission
reduction achievable through the application of
technology which the Administrator determines
will be available for the engines or vehicles to
which such standards apply, giving appropriate
consideration to the cost of applying such tech-
nology within the period of time available to
manufacturers and to noise, energy, and safety
factors associated with the application of such
technology. In determining what degree of re-
duction will be available, the Administrator shall
first consider standards equivalent in stringency
to standards for comparable motor vehicles or
engines (if any) regulated under section 7521 of
this title, taking into account the technological
feasibility, costs, safety, noise, and energy fac-
tors associated with achieving, as appropriate,
standards of such stringency and lead time. The
regulations shall apply to the useful life of the
engines or vehicles (as determined by the Ad-
ministrator).
(4) If the Administrator determines that any
emissions not referred to in paragraph (2) from
new nonroad engines or vehicles significantly
contribute to air pollution which may reasonably
be anticipated to endanger public health or wel-
15
fare, the Administrator may promulgate (and
from time to time revise) such regulations as
the Administrator deems appropriate containing
standards applicable to emissions from those
classes or categories of new nonroad engines
and new nonroad vehicles (other than locomo-
tives or engines used in locomotives) which in
the Administrator’s judgment cause, or contrib-
ute to, such air pollution, taking into account
costs, noise, safety, and energy factors associat-
ed with the application of technology which the
Administrator determines will be available for
the engines and vehicles to which such stan-
dards apply. The regulations shall apply to the
useful life of the engines or vehicles (as deter-
mined by the Administrator).
42 U.S.C. § 7547(a)(1)-(4) (also referred to as §§ 213(a)(1),
213(a)(2), 213(a)(3), 213(a)(4)).
We begin by examining EPA’s interpretation of the lan-
guage of § 213(a)(3), because that interpretation provides the
basis for the disputed contribution findings. ISMA argues
that § 213(a)(3) requires a finding that an individual category
of vehicles ‘‘significantly contributes’’ to CO concentrations in
more than one nonattainment area. EPA rejected that read-
ing, stating that § 213(a)(3) ‘‘does not require a finding of
‘significant contribution,’ but merely ‘contribution,’ for indi-
vidual categories of nonroad engines.’’ 65 Fed. Reg. at
76,791; see also 59 Fed. Reg. at 31,309 (explaining the basis
of EPA’s interpretation). We agree with the Agency: Con-
gress clearly did not intend to require an affirmative finding
of ‘‘significant’’ contribution from individual vehicle categories
in order to trigger regulation under § 213(a)(3).
We begin our interpretation of the provision with the
‘‘assumption that legislative purpose is expressed by the
ordinary meaning of the words used.’’ Sec. Indus. Ass’n v.
Bd. of Governors, 468 U.S. 137, 149 (1984) (quoting Russello
v. United States, 464 U.S. 16, 21 (1983)). The ordinary
meaning of ‘‘contribute’’ supports EPA’s reading. As used in
this context, ‘‘contribute’’ means simply ‘‘to have a share in
any act or effect,’’ WEBSTER’S THIRD NEW INTERNATIONAL DICTIO-
NARY 496 (1993), or ‘‘to have a part or share in producing,’’ 3
16
OXFORD ENGLISH DICTIONARY 849 (2d ed. 1989). Standing
alone, the term has no inherent connotation as to the magni-
tude or importance of the relevant ‘‘share’’ in the effect;
certainly it does not incorporate any ‘‘significance’’ require-
ment.
More important, the language and structure of § 213 as a
whole make quite clear that Congress did not intend to
require a finding of ‘‘significant contribution’’ for individual
vehicle categories. Twice in the provision, Congress drew a
distinction between the ‘‘significant contributor’’ finding re-
quired for all new and existing nonroad vehicles, and the
‘‘cause, or contribute to’’ finding for an individual category of
new nonroad vehicles. Under § 213(a)(2), EPA must deter-
mine whether all new and existing nonroad vehicles are
‘‘significant contributors’’ to CO concentrations in more than
one CO nonattainment area. If EPA makes this finding,
§ 213(a)(3) then requires the Agency to set standards for CO
emissions from each individual category of new vehicles that
‘‘cause, or contribute to, such air pollution.’’ This same
distinction between the ‘‘significant contributor’’ finding for
all nonroad vehicles and the ‘‘cause, or contribute to’’ finding
for an individual category of new nonroad vehicles is mirrored
in § 213(a)(4). In addition, § 213(a)(1) expressly directs EPA
to complete a study to determine whether all nonroad vehicles
‘‘cause, or significantly contribute to’’ harmful air pollution.
42 U.S.C. § 7547(a)(1) (emphasis added).
‘‘[I]t is a general principle of statutory construction that
when ‘Congress includes particular language in one section of
a statute but omits it in another section of the same Act, it is
generally presumed that Congress acts intentionally and pur-
posely in the disparate inclusion or exclusion.’ ’’ Barnhart v.
Sigmon Coal Co., 534 U.S. 438, 452 (2002) (quoting Russello,
464 U.S. at 23). The repeated use of the term ‘‘significant’’ to
modify the contribution required for all nonroad vehicles,
coupled with the omission of this modifier from the ‘‘cause, or
contribute to’’ finding required for individual categories of
new nonroad vehicles, indicates that Congress did not intend
to require a finding of ‘‘significant contribution’’ for individual
vehicle categories.
17
This interpretation is consistent with § 213’s structure and
purpose. The ‘‘significant contributor’’ finding required for all
nonroad engines can be understood as a threshold determina-
tion that the overall regulatory program is justified. See 58
Fed. Reg. 28,809, 28,811–13 (May 17, 1993) (discussing the
basis of the significance determination). Emissions from
individual categories of new vehicles will necessarily make a
lesser contribution to air pollution than will emissions from all
new and existing nonroad engines. However, as we noted in
Michigan v. EPA, 213 F.3d 663, 684 (D.C. Cir. 2000), ‘‘unlike
bologna, which remains bologna no matter how thin you slice
it, significant contribution may disappear if emissions activity
is sliced too thinly.’’ Unless each individual vehicle category
that makes some nontrivial contribution to pollution is subject
to regulation, it is unlikely that the ‘‘significant’’ cumulative
effect of emissions from all nonroad vehicle categories can
effectively be controlled.
ISMA’s arguments in support of its position are untenable.
ISMA first contends that § 213(a)(3) somehow incorporates
paragraph (a)(2)’s ‘‘significant contributor’’ language, because
paragraph (a)(3) refers back to paragraph (a)(2) by means of
the phrase ‘‘cause, or contribute to, such air pollution.’’ But
the phrase ‘‘such air pollution’’ clearly refers to ‘‘ozone or
carbon monoxide concentrations in more than 1 area which
has failed to attain’’ the NAAQS, and bears no plausible
relation to the ‘‘significant contributor’’ language in para-
graph (a)(2). See 42 U.S.C. § 7547(a)(2), (3).
ISMA also cites a Senate Committee report referring to
proposed language from a Senate bill that would have re-
quired EPA first to list categories of nonroad engines, and
then to promulgate regulations for each category found to
‘‘contribute significantly’’ to certain pollution. See S. 1630,
101st Cong. § 217 (1989), reprinted in 5 A LEGISLATIVE HISTO-
RY OF THE CLEAN AIR ACT AMENDMENTS OF 1990, at 7906, 8044–45
(Comm. Print 1993) (‘‘LEGISLATIVE HISTORY’’); S. REP. No. 101–
228, at 103–05 (1989). The Senate provision in question was
dropped in favor of a House amendment, see Chafee–Baucus
Statement of Senate Managers, S. 1630, The Clean Air Act
Amendments of 1990, reprinted in 1 LEGISLATIVE HISTORY, at
880, 896–97, and the discussion of the proposed language in
18
the Senate report is therefore irrelevant. The House Com-
mittee report discussing the provision that was ultimately
enacted provides no commentary on § 213(a)(3)’s ‘‘cause, or
contribute to’’ language. See H.R. REP. No. 101–490, at 309–
10 (1990).
Finally, ISMA argues that EPA’s interpretation leads to an
absurd result, because it permits promulgation of national
regulations for a vehicle category on the basis of a finding
that this category makes an insignificant contribution to the
relevant pollution. No such issue is presented in this case,
however, because EPA’s contribution findings, discussed at
greater length below, clearly involve more than a de minimis
contribution.
2. Snowmobiles-Only Contribution Finding
In the Final Rule, EPA found that emissions from snowmo-
biles, considered as a separate category, contribute to CO
concentrations in more than one area designated by EPA as a
‘‘nonattainment’’ area for CO. 67 Fed. Reg. at 68,248–49.
ISMA now challenges that finding, claiming that it lacks
adequate evidentiary support and is premised on an improper
interpretation of the statute. We reject these claims and
uphold EPA’s finding.
EPA’s snowmobiles-only finding is based primarily on
data from the State of Alaska, estimating that snowmobile
operation for engine maintenance, loading, and unloading
contributes 0.3 tons/day of CO emissions to the Fairbanks
nonattainment area and 0.34 tons/day of CO emissions to the
Anchorage nonattainment area. Id. at 68,248. For Fair-
banks, this contribution was equivalent to 1.2% of the total
daily CO inventory for 2001. Id. EPA further found that
there is a snowmobile trail located within the Spokane,
Washington, nonattainment area, and that snowmobile oper-
ation on that trail contributes to CO concentrations there.
Id. The Agency also noted that there are snowmobile trails
located in close proximity to a number of other CO nonat-
tainment areas and that there is evidence – primarily from
Michigan and Alaska – that snowmobiles are ridden on
roads as well as trails. Id. This evidence is adequate to
19
support EPA’s contribution finding, at least as to the An-
chorage, Fairbanks, and Spokane nonattainment areas.
ISMA levels a number of challenges against EPA’s finding,
each of which falls short. Citing this court’s decisions in
Michigan v. EPA, 213 F.3d 663 (D.C. Cir. 2000), and Appala-
chian Power Co. v. EPA, 251 F.3d 1026 (D.C. Cir. 2001),
ISMA first argues that contribution must be demonstrated
through modeling or analysis showing transport of the emis-
sions to the relevant area. But EPA’s contribution finding in
the instant case is based on emissions within the nonattain-
ment areas, so no question of transport is implicated. Michi-
gan and Appalachian Power Co. are inapposite: Both cases
involved EPA’s attempt to impose regulatory requirements
on upwind areas of a State without the requisite finding that
these specific areas in fact made any measurable contribution
to pollution in downwind areas. See Appalachian Power Co.,
251 F.3d at 1040; Michigan, 213 F.3d at 683–84.
ISMA also contends that EPA was not entitled to rely on
Alaska’s data in making the contribution finding, because
those data are estimates of expected snowmobile emissions,
rather than actual measured emissions. We see nothing
improper in EPA’s reliance on Alaska’s data, which in the
case of Fairbanks were confirmed by a National Research
Council study, see 67 Fed. Reg. at 68,248. Nothing in the
statute requires direct empirical measurements, and nothing
in the record suggests that Alaska’s estimates are otherwise
unreasonable.
Finally, ISMA challenges EPA’s decision to base its contri-
bution finding on whether the areas in question are designat-
ed by EPA as ‘‘nonattainment’’ areas, rather than on whether
the area is currently attaining the NAAQS. The Clean Air
Act imposes numerous requirements for redesignation of an
area from nonattainment to attainment status, including EPA
approval of a state maintenance plan that will ensure attain-
ment of the NAAQS for the decade following redesignation.
See 42 U.S.C. § 7407(d)(3); see also 57 Fed. Reg. 13,498,
13,561–64 (Apr. 16, 1992) (discussing EPA procedures govern-
ing redesignation). Accordingly, attainment of the NAAQS
20
does not automatically result in redesignation. While Fair-
banks has not yet attained the NAAQS for CO, Spokane and
Anchorage are currently attaining the NAAQS for CO but
continue to be designated as ‘‘serious nonattainment’’ areas.
See 66 Fed. Reg. 44,060 (Aug. 22, 2001) (Spokane); 66 Fed.
Reg. 36,476 (July 12, 2001) (Anchorage). The other areas
mentioned in EPA’s snowmobiles-only contribution finding –
Fort Collins, Colorado, and Medford, Oregon – have been
redesignated as attainment areas for CO. See 68 Fed. Reg.
43,316 (July 22, 2003) (Fort Collins); 67 Fed. Reg. 48,388
(July 24, 2002) (Medford).
This issue turns on the meaning of the language in
§ 213(a)(2) referring to emissions contributing to CO concen-
trations in ‘‘more than 1 area which has failed to attain’’ the
NAAQS for CO, 42 U.S.C. § 7547(a)(2) (emphasis added).
The phrase ‘‘has failed to attain’’ – stated in the present
perfect tense – is ambiguous with regard to whether it applies
to an area that failed to attain the NAAQS in the past but is
currently attaining the standard. EPA’s interpretation of the
phrase to refer to an area’s attainment designation status,
rather than whether it currently is attaining the NAAQS, is
reasonable. Current attainment does not demonstrate that
an area will continue in attainment, and areas often experi-
ence ‘‘relapse.’’ See 67 Fed. Reg. at 68,248 n.7. Areas not
yet redesignated ‘‘have failed to attain’’ the NAAQS, and may
reasonably be considered to be at risk of relapse – due, in
part, to any emissions contributing to the relevant pollution –
until the steps necessary for redesignation have been taken.
3. Grouping of Snowmobiles With Land–Based Recre-
ational Vehicles
ISMA also challenges EPA’s decision to group snowmobiles
with the land-based recreational vehicle category for purposes
of the Agency’s December 2000 contribution finding. ISMA
claims that the grouping is unreasonable, because snowmo-
biles are operated in different parts of the country and at
different times of the year than other recreational vehicles.
The relevance of this claim is uncertain in light of our
decision to uphold the snowmobiles-only contribution finding,
21
which appears to provide an adequate basis for promulgating
the CO standards under § 213(a)(3). However, because EPA
purported to rely in part upon the December 2000 finding
when issuing the standards, see id. at 68,248; SAC at II–24,
J.A. 92, we address – and reject – ISMA’s claim.
EPA raises a threshold concern as to whether ISMA’s
challenge to the Agency’s grouping decision is timely. We
find that it is. Although EPA first made the grouping
decision in the December 2000 finding, the Agency reopened
that decision in the rulemaking under review in this case.
‘‘ ‘[W]hether an agency has in fact reopened an issue’ is
dependent upon ‘the entire context of the rulemaking includ-
ing all relevant proposals and reactions of the agency,’ and
not just the agency’s stated intent.’’ Appalachian Power Co.,
251 F.3d at 1033 (quoting Pub. Citizen v. Nuclear Regulatory
Comm’n, 901 F.2d 147, 150 (D.C. Cir. 1990)). More specifi-
cally, ‘‘if an agency’s response to comments ‘explicitly or
implicitly shows that the agency actually reconsidered the
rule, the matter has been reopened.’’’ Id. (quoting PanAm-
Sat Corp. v. FCC, 198 F.3d 890, 897 (D.C. Cir. 1999)). Such is
the case here.
In the notice of proposed rulemaking, EPA directed virtu-
ally all of its discussion of the health effects of CO pollution to
emissions from snowmobiles alone. See 66 Fed. Reg. at
51,105–06. ISMA and Polaris Industries submitted com-
ments, based on a study by a consulting group, challenging
the reasonableness of grouping snowmobiles with other land-
based recreational vehicles for purposes of the CO finding
and requesting that EPA reconsider its December 2000 find-
ing. EPA considered these comments on the merits and
responded, in part, with a new defense of the grouping: i.e.,
that, ‘‘even if [the Agency] did review snowmobile contribu-
tion separately, there is no question that they’’ meet the
contribution criteria. SAC at II–24, J.A. 92. This was not a
case in which parties merely ‘‘comment[ed] on matters other
than those actually at issue [and] goad[ed the] agency into a
reply.’’ See Am. Iron & Steel Inst. v. EPA, 886 F.2d 390, 398
(D.C. Cir. 1989), cert. denied, 497 U.S. 1003 (1990). Rather,
EPA put the basis of its finding in play by offering new
22
evidence linking snowmobiles to CO nonattainment, and the
comments went directly to the basis of EPA’s authority to
issue the snowmobile standards. Most important, EPA’s
response to these comments clearly indicates that it actually
reconsidered – and therefore reopened to challenge – its
grouping decision for purposes of the CO contribution finding.
On the merits, we find nothing improper in this decision.
EPA has discretion to define reasonable ‘‘categories or
classes’’ of vehicles under § 213(a)(3). See Engine Mfrs.
Ass’n v. EPA, 88 F.3d 1075, 1097–98 (D.C. Cir. 1996). In this
case, the Agency reasonably found that engines used in
snowmobiles and other land-based recreational vehicles have
similar characteristics, including combustion cycle, fuel usage
patterns, power rating, and equipment type. See 65 Fed.
Reg. at 76,791; 66 Fed. Reg. at 51,148. EPA ultimately
concluded that snowmobiles should not be grouped with other
land-based recreational vehicles for purposes of regulating
ozone precursors, because snowmobiles are operated in cold
weather and ozone is a warm-weather phenomenon. SAC at
II–24, V–31, J.A. 92, 110. But this has no bearing on the
reasonableness of the grouping for purposes of CO contribu-
tion, because peak CO concentrations typically occur during
the colder months of the year, see id.; 66 Fed. Reg. at 51,105.
ISMA offers no support for its assertion that snowmobiles
and other recreational vehicles are not operated in the same
parts of the country. We therefore uphold as reasonable
EPA’s decision to group snowmobiles with other land-based
recreational vehicles for purposes of the December 2000 CO
contribution finding.
C. EPA’s Authority to Regulate HC and NOx Under
§ 213(a)(4)
EPA promulgated the HC and NOx standards under
§ 213(a)(4), which authorizes the Agency to regulate emis-
sions ‘‘not referred to in paragraph (2)’’ if such emissions
contribute to air pollution which may reasonably be anticipat-
ed to endanger the public health or welfare. Paragraph (2)
directs the Agency to determine ‘‘whether emissions of car-
bon monoxide, oxides of nitrogen, and volatile organic com-
23
pounds from new and existing nonroad engines or nonroad
vehicles TTT are significant contributors to ozone or carbon
monoxide concentrations in more than 1’’ nonattainment area.
42 U.S.C. § 7547(a)(2). ISMA argues that EPA lacks author-
ity to regulate HC and NOx emissions under § 213(a)(4),
because they are referred to in § 213(a)(2). We reject this
claim as to HC, but we agree that EPA exceeded its statutory
authority in issuing the NOx standard. Accordingly, we
hereby vacate the latter standard.
1. HC Standard
The only ‘‘emissions’’ identified in § 213(a)(2) are CO, NOx,
and VOCs. ISMA argues that § 213(a)(2) nevertheless ‘‘re-
fers to’’ HC, because it mentions ozone, of which HC is a
precursor, and VOCs, which consist primarily of HC. We
disagree. First, ozone is not an ‘‘emission,’’ but rather an
ambient pollutant formed through reactions between chemical
precursors in the atmosphere, so paragraph (a)(2)’s reference
to ozone is irrelevant here. Second, VOCs and HC are not
coterminous. While HC and VOCs overlap in parts, they are
distinct: Not all VOCs are hydrocarbons, and not all HC
emissions are VOCs. See SAC at II–26, J.A. 94. It is true
that EPA has in the past used HC as a surrogate for
regulating VOCs, see id., but it did not do so here. Rather,
EPA regulated HC emissions as a means of controlling fine
PM emissions and pollution. The Agency reasonably deter-
mined that regulating HC would control PM pollution both
because HC itself contributes to such pollution, and because
HC provides a good proxy for regulating fine PM emissions.
67 Fed. Reg. at 68,254. Based on the foregoing, we conclude
that EPA has the authority to regulate HC emissions under
§ 213(a)(4).
2. NOx Standard
Section 213(a)(2) expressly and undeniably refers to NOx,
and we therefore conclude that EPA lacks authority to regu-
late NOx emissions under § 213(a)(4). EPA resists this con-
clusion by arguing that the phrase emissions ‘‘referred to in
paragraph (2)’’ means ‘‘emissions of carbon monoxide, oxides
of nitrogen, and volatile organic compounds from TTT nonroad
24
engines TTT [that] are significant contributors to ozone or
carbon monoxide concentrations in more than one [ozone or
CO nonattainment area].’’ SAC at II–26, J.A. 94; see also 65
Fed. Reg. at 76,790. In other words, EPA reads § 213(a)(4)
to allow it to regulate any emissions meeting (a)(4)’s contribu-
tion criteria – including CO, NOx, and VOCs – so long as they
are being regulated for the purpose of addressing pollution
problems other than CO or ozone nonattainment. See id.
The Agency’s interpretation is untenable, for several reasons.
First, EPA’s interpretation runs counter to the plain mean-
ing of the phrase ‘‘emissions TTT referred to in paragraph
(2).’’ The most natural reading of the phrase is ‘‘emissions of
carbon monoxide, oxides of nitrogen, and volatile organic
compounds.’’ See 42 U.S.C. § 7547(a)(2). EPA’s reading, by
contrast, requires a strange contortion of § 213(a)(2)’s lan-
guage – omitting several key words and supplying others.
Had Congress intended the meaning and result which EPA
urges, it would have said so more clearly. We therefore find
EPA’s reading of the language to be implausible on its face.
In addition, the sole discussion of the relevant language in
the legislative history contradicts EPA’s reading. In ad-
dressing the language ultimately enacted, the House Commit-
tee report states:
Paragraph (4) of revised section 213(a) provides
that if the Administrator determines that emis-
sions from nonroad vehicles not specifically
mentioned in paragraph (2) (which lists CO,
VOCs, and NOx) significantly contribute to air
pollution which may reasonably be anticipated
to endanger public health or welfare, the Ad-
ministrator may promulgate such regulations as
he deems appropriateTTTT
H.R. REP. No. 101–490, at 309–10 (1990) (emphasis added).
This confirms that the drafters intended the phrase ‘‘emis-
sions TTT referred to in paragraph (2)’’ to mean simply
emissions of CO, NOx, and VOCs, without regard to whether
such emissions do or do not significantly contribute to CO or
ozone concentrations in more than one nonattainment area.
25
EPA protests that the reading of § 213(a)(4) that we adopt
today leaves a ‘‘gap’’ in its regulatory authority. It precludes
the Agency from addressing pollution problems, other than
CO or ozone nonattainment, caused by CO, VOCs, and NOx
emissions from nonroad engines – for instance, the potential
contribution of NOx emissions to haze or acid deposition. We
agree that § 213 would likely permit more comprehensive
pollution control if Congress had drafted § 213(a)(4) to au-
thorize regulation of these emissions. But Congress has not
done so. Because NOx is an ‘‘emission referred to’’ in
§ 213(a)(2), we hold that EPA lacks statutory authority to
regulate NOx under § 213(a)(4). We therefore vacate the
NOx standard.
D. Scope of Implementation of Advanced Technologies
Having disposed of ISMA’s claims, we now turn to Bluewa-
ter’s primary challenge to the Final Rule. Section 213(a)(3)
requires EPA to set standards that
achieve the greatest degree of emission reduc-
tion achievable through the application of tech-
nology which the Administrator determines will
be available for the engines or vehicles to which
such standards apply, giving appropriate consid-
eration to the cost of applying such technology
within the period of time available to manufac-
turers and to noise, energy, and safety factors
associated with the application of such technolo-
gy.
42 U.S.C. § 7547(a)(3). EPA based its standards on two
advanced technologies: DI two-stroke engines and four-
stroke engines. The Agency estimated that its Phase 3
standards, effective for the 2012 model year, would require
implementation of these technologies in 70% of all new snow-
mobiles. It declined to set standards requiring full imple-
mentation of these technologies, finding that a number of
factors – including the effort required to apply advanced
technologies to the variety of snowmobile models – made this
infeasible within the regulatory lead time.
Bluewater claims that EPA improperly assumed that
§ 213(a)(3) requires the Agency to set standards at a level
26
sufficiently lenient to allow continued production of all exist-
ing snowmobile models. Bluewater appears to read the
provision to instead mandate 100% implementation of technol-
ogies that EPA has determined ‘‘will be available,’’ subject
only to the proviso that the standards must permit continued
production of vehicles satisfying ‘‘basic demand.’’ See Blue-
water Br. at 18–19. Alternatively, Bluewater claims that
EPA acted arbitrarily and capriciously by failing to explain
the basis of its conclusion that manufacturers could not
achieve broader application of advanced technology by 2012.
We reject Bluewater’s statutory challenge as articulated,
but we agree that EPA has failed to explain adequately the
basis of its decision to set the standards as it did. According-
ly, we remand the standards for EPA to clarify the analysis
and evidence upon which the standards are premised.
1. Statutory Claim
Section 213(a)(3) is a ‘‘technology-forcing’’ provision, requir-
ing EPA ‘‘to project future advances in pollution control
capability TTT [and] to press for development and application
of improved technology rather than be limited by that which
exists today.’’ Husqvarna AB v. EPA, 254 F.3d 195, 201
(D.C. Cir. 2001) (quoting Natural Res. Def. Council v. EPA,
655 F.2d 318, 328 (D.C. Cir. 1981) (‘‘NRDC’’)). However, the
Agency is not to set standards based solely on a determina-
tion that a particular technology ‘‘will be available,’’ as a
matter of pure technological feasibility. Rather, in determin-
ing the maximum emissions reduction achievable, EPA must
also give ‘‘appropriate consideration to the cost of applying
such technology within the period of time available to manu-
facturers and to noise, energy, and safety factors associated
with the application of such technology.’’ 42 U.S.C.
§ 7547(a)(3).
In the Final Rule, EPA found that there was no purely
technological obstacle to the application of DI two-stroke and
four-stroke engines to snowmobiles, which in fact are already
in use in some models. The Agency concluded that, by 2012,
‘‘manufacturers could, at least in theory, apply advanced
technology across essentially their entire product lines.’’ 67
27
Fed. Reg. at 68,273. And it is clear that standards reflecting
across-the-fleet application of such technology would result in
significantly greater emissions reductions than the standards
adopted. However, EPA interprets § 213(a)(3)’s reference to
cost and other factors to permit it to set less stringent
standards: Standards reflecting across-the-fleet implementa-
tion are not ‘‘achievable’’ by 2012, the Agency contends,
because of manufacturers’ resource constraints and the de-
sign and development efforts required to apply new technolo-
gies to the variety of different snowmobile models.
We agree that EPA may rely on cost and other statutory
factors to set standards at a level less stringent than that
reflected by across-the-fleet implementation of advanced tech-
nologies. This court noted in Husqvarna that ‘‘[t]he over-
riding goal of [§ 213] is air quality and the other listed
considerations, while significant, are subordinate to that
goal.’’ 254 F.3d at 200. Nevertheless, as the court empha-
sized in reflecting on very similar language in § 202(l ) of the
CAA, the provision ‘‘does not resolve how the Administrator
should weigh all [the statutory] factors in the process of
finding the ‘greatest emission reduction achievable.’ ’’ Sierra
Club v. EPA, 325 F.3d 374, 378 (D.C. Cir. 2003).
We find nothing in the statute requiring EPA to set
standards at a level of stringency that would require discon-
tinuation of all vehicles other than those satisfying ‘‘basic
demand.’’ Bluewater derives the ‘‘basic demand’’ formula
from our decision in International Harvester Co. v. Ruckel-
shaus, 478 F.2d 615, 640 (D.C. Cir. 1973), which involved
statutory provisions and regulations that are very different
from those at issue in this case. Instead, the relevant
question here is whether, based on ‘‘appropriate consider-
ation’’ of cost and the other factors enumerated in § 213(a)(3),
EPA could reasonably determine that the standards it
adopted reflect the ‘‘greatest degree of emission reduction
achievable.’’ Nothing in § 213(a)(3) prevents EPA from set-
ting standards sufficiently lenient to permit continued produc-
tion of all existing models, so long as EPA’s decision is based
upon appropriate consideration of these factors.
28
2. ‘‘Arbitrary and Capricious’’ Claim
Although we agree with EPA that the statute permits it to
balance technological feasibility against cost and other factors
in setting standards, we conclude that the Agency has not
adequately explained its exercise of that discretion in this
case.
First, EPA expected that manufacturers would apply ad-
vanced technologies on an engine-family-by-engine-family ba-
sis, and concluded that, by 2012, advanced technologies could
be applied to models accounting for roughly 70% of new
snowmobiles, but not the remaining models. See 67 Fed.
Reg. at 68,273. Implicit in this conclusion is an assumption
that no existing models could be eliminated. In other words,
EPA assumed that manufacturers could not discontinue or
replace any of the models – collectively accounting for rough-
ly 30% of new snowmobiles – to which advanced technology
could not be applied by 2012. That assumption, if it is to
stand, must be grounded in ‘‘appropriate consideration’’ of the
relevant statutory factors. EPA argues before this court that
its position was based on cost considerations. However, we
can find nothing in the record indicating that the Agency
evaluated or reached any conclusions as to the cost of discon-
tinuing models to which advanced technology could not be
applied by 2012. Absolute certainty and precision on this
point are not required, but a reasonable explanation clearly is
necessary. Accordingly, we direct the Agency on remand to
clarify the statutory and evidentiary basis of its position.
Second, even assuming that EPA correctly concluded that
no models could be discontinued, the Agency failed to explain
adequately how it arrived at the specific standards adopted.
EPA did articulate several general reasons for its conclusion
that greater emissions reductions could not be achieved by
2012. Most important, EPA found that, because of the wide
variety of snowmobile models, the design and development
work necessary to apply advanced technologies to all models
would require significant time and investment. Id. The
Agency noted that snowmobile manufacturers are ‘‘resource
constrained,’’ and that those relying on external engine sup-
29
pliers would find it more difficult to undertake rapid develop-
ment of new technologies. Id.
This generalized discussion of the limiting factors does not
explain how the Agency arrived at the specific conclusion that
emissions reductions corresponding to application of advanced
technologies to 70% of new snowmobiles were the most that
could be achieved by 2012. The Agency’s explanation of its
reasoning could just as well support standards corresponding
to 30% or 100% application in that time frame. And we find
nothing in the record before us explaining the analysis and
evidence underlying EPA’s conclusions.
We emphasize that we do not view the standards adopted
as facially unreasonable, nor have we found evidence in the
record contradicting the Agency’s ultimate decision. But in
order to determine whether that decision reflects a ‘‘rational
connection between the facts found and the choice made,’’
Motor Vehicle Mfrs. Ass’n, 463 U.S. at 43, a reasonable
explanation of the specific analysis and evidence upon which
the Agency relied is necessary. ‘‘With its delicate balance of
thorough record scrutiny and deference to agency expertise,
judicial review can occur only when agencies explain their
decisions with precision, for ‘[i]t will not do for a court to be
compelled to guess at the theory underlying the agency’s
actionTTTT’ ’’ Am. Lung Ass’n v. EPA, 134 F.3d 388, 392
(D.C. Cir. 1998) (quoting SEC v. Chenery Corp., 332 U.S. 194,
196–97 (1947)).
In defense of its limited explanation for its decision, EPA
refers us to a line of cases in which we have deferred to the
Agency’s predictions that a particular control technology will
be available in the future. See Nat’l Petrochemical & Refin-
ers Ass’n v. EPA, 287 F.3d 1130, 1144 (D.C. Cir. 2002);
Natural Res. Def. Council v. Thomas, 805 F.2d 410, 432–34
(D.C. Cir. 1986); NRDC, 655 F.2d at 333. These cases stand
for the proposition that, ‘‘[i]n the absence of theoretical
objections to the technology, the agency need only identify
the major steps necessary for development of the device, and
give plausible reasons for its belief that the industry will be
able to solve those problems in the time remaining.’’ Nat’l
30
Petrochemical & Refiners Ass’n, 287 F.3d at 1144 (quoting
NRDC, 655 F.2d at 333).
The issue here is different. In this case, EPA’s decision
does not involve a prediction about the development of a
technology that is not yet available. Rather, it concluded that
technology that is currently available could not be applied to
all models within the available lead time. This decision was
apparently based, not on technological obstacles per se, but
rather on the cost and time required to ‘‘optimize’’ advanced
technology for each snowmobile model on the market. In-
deed, it is not clear whether there is any meaningful distinc-
tion between time and cost here; it may be that the pace of
implementation is simply a function of the level of investment.
Naturally, there will be some uncertainty in any estimate of
how much money and time is needed to apply advanced
technologies to each model or engine family and, accordingly,
what scope of implementation is actually feasible in the time
available. But this does not excuse EPA from offering any
estimate whatsoever. We can defer to the Agency’s predic-
tion of the feasible pace of implementation only if it has
adequately explained the basis of that prediction. As this
court stated in NRDC, ‘‘[t]he Clean Air Act requires EPA to
look to the future in setting standards, but the agency must
also provide a reasoned explanation of its basis for believing
that its projection is reliable. This includes a defense of its
methodology for arriving at numerical estimates.’’ 655 F.2d
at 328 (citing Int’l Harvester Co., 478 F.2d at 629).
E. Catalyst Technology and Other Claims
1. Catalyst Technology
EPA declined to adopt emissions standards based on the
application of catalyst technology, which is widely used in
exhaust aftertreatment devices to reduce emissions from au-
tomobiles and other engines. Bluewater challenges this deci-
sion, claiming that it was based on an impermissible interpre-
tation of the statute and is arbitrary and capricious. We
reject this claim.
31
As noted above, § 213(a)(3) requires adoption of standards
based on technology that EPA determines ‘‘will be available.’’
Bluewater asserts that EPA violated this mandate by basing
its decision on a finding that catalyst technology is not
already available for snowmobiles. The premise of this argu-
ment is unsupported. EPA determined that there were
significant technical barriers to development of catalyst tech-
nology for snowmobiles, including difficulties with achieving
full oxidation of ‘‘rich exhaust’’ and overcoming space con-
straints. See SAC at V–25, J.A. 104. EPA concluded that it
could not, at this time, predict that these hurdles could be
overcome or that such technology would be available by 2012.
Id. Therefore, EPA’s decision was based on the lack of
evidence to support a prediction that catalyst technology ‘‘will
be available’’ within the relevant lead time, not on the mere
fact that such technology is not available now.
Bluewater contends, in the alternative, that EPA’s determi-
nation that catalyst technology will not be available by 2012 is
not supported by the record. Bluewater first challenges
EPA’s reliance on the ‘‘rich exhaust’’ barrier to development
of catalyst technology for snowmobiles. Bluewater asserts
that DI two-stroke and four-stroke engines have comparative-
ly low CO and HC emissions and that application of these
engine technologies would therefore eliminate the ‘‘rich ex-
haust’’ problem. In other words, EPA should have consid-
ered standards based on the simultaneous development and
application of advanced engines and catalyst technology. We
cannot agree that EPA’s failure to analyze this scenario was
arbitrary and capricious. Having concluded that advanced
engine technologies alone could not be fully implemented in
the available lead time, the Agency was not required, on its
own motion, to consider whether simultaneous development
and application of catalyst technology would be feasible. Cf.
Nat’l Petrochemical & Refiners Ass’n, 287 F.3d at 1145
(noting that EPA is entitled to some deference in choosing
the technological basis of its standards).
Bluewater also challenges EPA’s position that space con-
straints posed a barrier to adoption of catalyst technology.
This challenge is premised primarily on EPA’s statements
32
elsewhere in the record indicating that space constraints can
be overcome for many large-SI and recreational-vehicle en-
gines. See Regulatory Support Document (‘‘RSD’’) at 3–19,
J.A. 177. But the record indicates that snowmobiles present
some unique problems, see Written Testimony of the Manu-
facturers of Emission Controls Association at 9, J.A. 272, so
these statements do not present the fatal contradiction sug-
gested by Bluewater. Similarly, Bluewater cites EPA’s find-
ing in an earlier proceeding that catalysts can be built into
handheld lawn and garden equipment engines. We fail to see
how this finding undermines EPA’s conclusion with regard to
snowmobiles, given the significant differences between the
two types of engines and applications, see 67 Fed. Reg. at
68,260; Husqvarna, 254 F.3d at 197 n.1.
2. Remaining Claims
Bluewater raised a number of other claims in its briefs.
We have carefully considered each of these claims and find no
merit in any. We offer our conclusions here in summary
form.
Bluewater claims, based on the December 2000 contribution
finding for land-based recreational vehicles, that EPA was
required to regulate HC and NOx under § 213(a)(3), instead
of (a)(4). Bluewater argues that, once EPA found that the
category including snowmobiles contributes to ozone pollution
in more than one nonattainment area, the Agency was re-
quired to regulate snowmobile emissions of ozone precursors
under § 213(a)(3). We disagree. EPA reasonably deter-
mined that snowmobiles should be considered separately from
the recreational vehicle category for purposes of the ozone
contribution finding, because snowmobiles are operated in
cold weather and ozone is a warm-weather phenomenon. See
SAC at II–24, V–31, J.A. 92, 110.
Bluewater also contends that, once EPA made a contribu-
tion finding for ozone or CO, it was obligated to regulate both
CO and ozone precursors under § 213(a)(3). We seriously
doubt that the statute permits this odd reading, but it certain-
ly does not require it. EPA reasonably interpreted
§ 213(a)(3) to require regulation of ozone precursors under
33
that provision only if it had made a finding that the relevant
individual class or category of engines – in this case snowmo-
biles – contributes to ozone concentrations in more than one
ozone nonattainment area.
In addition, Bluewater argues that EPA improperly failed
to consider the noise and energy impacts of its decision to
adopt standards reflecting continued use of carbureted two-
stroke engines in up to 30% of new snowmobiles after 2012.
Section 213(a)(3) requires EPA to set standards reflecting the
maximum emissions achievable through adoption of technolo-
gy that will be available, ‘‘giving appropriate consideration’’
to, inter alia, ‘‘noise [and] energy TTT factors associated with
the application of such technology.’’ EPA clearly met this
requirement by giving consideration to the (undisputedly
positive) noise and energy impacts of basing standards on the
adoption of DI two-stroke and four-stroke engine technolo-
gies. See RSD at 4–48, J.A. 189. The statute does not
require EPA to set standards maximizing energy conserva-
tion or noise reduction.
Bluewater next claims that EPA violated § 213(a)(3) by
failing to ‘‘first consider standards equivalent in stringency to
standards for comparable [onroad] motor vehicles or engines
(if any) regulated under’’ § 202 of the Act. See 42 U.S.C.
§ 7547(a)(3). In response to comments on this point, EPA
determined that automobiles are not ‘‘comparable’’ to snow-
mobiles and that the threshold for ‘‘first considering’’ similar
standards therefore was not triggered. See SAC at V–22,
J.A. 101. Bluewater argues that EPA should have considered
whether onroad motorcycles are ‘‘comparable,’’ but this spe-
cific claim was not raised in administrative proceedings and
therefore was waived.
Finally, Bluewater argues that EPA acted arbitrarily and
capriciously by failing to adopt standards based on the exclu-
sive application of four-stroke technology, rather than a mix
of DI two-stroke and four-stroke technology. This claim is
baseless. EPA reasonably concluded that DI two-stroke
engines would achieve greater CO emission reductions than
comparably powered four-stroke engines, but would yield
34
somewhat lower HC reductions. 67 Fed. Reg. at 68,272.
EPA accordingly concluded that neither technology was obvi-
ously superior. In reaching its conclusion, EPA relied on
data from personal watercraft (jet skis) to predict expected
emissions from high-powered four-stroke engines. RSD at 4–
42, J.A. 183. This was appropriate, because EPA reasonably
determined that emissions from the two types of engines
would be similar, and because there were no data on high-
powered four-stroke snowmobile engines available at the time
of the rulemaking. See id.; see also, e.g., Am. Iron & Steel
Inst. v. EPA, 115 F.3d 979, 1004–05 (D.C. Cir. 1997) (uphold-
ing the Agency’s reliance on a reasonable model to estimate
results where no field data were available).
III. CONCLUSION
For the reasons set forth above, we grant in part and deny
in part each of the two petitions for review. We hereby
vacate the NOx standard on the ground that EPA lacks
statutory authority to regulate NOx emissions under
§ 213(a)(4). We hold that the Agency acted within its statu-
tory authority in promulgating the CO and HC standards
under § 213(a)(3) and (a)(4), respectively. However, we re-
mand the CO and HC standards for EPA to clarify the
analysis and evidence upon which the standards are based.
Specifically, we direct EPA to clarify (1) the statutory and
evidentiary basis of the Agency’s assumption that the stan-
dards must be sufficiently lenient to permit the continued
production of all existing snowmobile models, and (2) the
analysis and evidence underlying the Agency’s conclusion that
advanced technologies can be applied to no more than 70% of
new snowmobiles by 2012. We uphold the CO and HC
standards in all other respects.