Hearth, Patio & Barbecue Association v. Environmental Protection Agency

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued January 28, 2021 Decided August 27, 2021
No. 15-1056
HEARTH, PATIO & BARBECUE ASSOCIATION,
PETITIONER
v.
ENVIRONMENTAL PROTECTION AGENCY,
RESPONDENT
AMERICAN LUNG ASSOCIATION , ET AL .,
I NTERVENORS
On Petition for Review of a Final Agency Action
of the Environmental Protection Agency
David Y. Chung argued the cause for petitioner. With him
on the briefs was Amanda Shafer Berman. David E. Menotti
entered an appearance.
Simi Bhat, Senior Trial Attorney, U.S. Department of
Justice, argued the cause for respondent. With her on the brief
were Jonathan Brightbill, Principal Deputy Assistant Attorney
General, Eric Grant, Deputy Assistant Attorney General, and
Scott Jordan, Counsel, U.S. Environmental Protection Agency.
Timothy D. Ballo was on the brief for intervenor American
Lung Association, et al. in support of respondent.

-- 1 of 32 --

2
Letitia James, Attorney General, Office of the Attorney
General for the State of New York, Barbara D. Underwood,
Solicitor General, Anisha S. Dasgupta, Deputy Solicitor
General, Philip J. Levitz, Assistant Solicitor General, Michael
J. Myers, Senior Counsel, Nicholas C. Buttino, Assistant
Attorney General, Clyde
AEd
@ Sniffen, Jr., Acting Attorney
General, Office of the Attorney General for the State of Alaska,
William Tong, Attorney General, Office of the Attorney
General for the State of Connecticut, Kwame Raoul, Attorney
General, Office of the Attorney General for the State of Illinois,
Brian E. Frosh, Attorney General, Office of the Attorney
General for the State of Maryland, Keith Ellison, Attorney
General, Office of the Attorney General for the State of
Minnesota, Gurbir S. Grewal, Attorney General, Office of the
Attorney General for the State of New Jersey, Ellen F.
Rosenblum, Attorney General, Office of the Attorney General
for the State of Oregon, Peter F. Neronha, Attorney General,
Office of the Attorney General for the State of Rhode Island,
Thomas J. Donovan, Jr., Attorney General, Office of the
Attorney General for the State of Vermont, Robert W.
Ferguson, Attorney General, Office of the Attorney General
for the State of Washington, and Jennifer A. Dold, General
Counsel, were on the brief for amici curiae States of New York,
et al. in support of respondent. Paul A. Garrahan, Attorney-
in-Charge, Office of the Attorney General for the State of
Oregon, Carol A. Iancu, Assistant Attorney General, Office of
the Attorney General for the Commonwealth of Massachusetts,
Scott N. Koschwitz, Assistant Attorney General, Office of the
Attorney General for the State of Connecticut, Gregory S.
Schultz, Special Assistant Attorney General, Office of the
Attorney General for the State of Rhode Island, and Nicholas
F. Persampieri and William H. Sorrell, Assistant Attorneys
General, Office of the Attorney General for the State of
Vermont, entered appearances.

-- 2 of 32 --

3
Before: HENDERSON, PILLARD and W ILKINS , Circuit
Judges.
Opinion for the Court filed by Circuit Judge PILLARD.
PILLARD, Circuit Judge: Petitioner Hearth, Patio &
Barbecue Association (HPBA) challenges the portion of the
Environmental Protection Agency’s (EPA) 2015 rule updating
the standards under which EPA audits wood-burning heaters’
compliance with emissions limits required by the Clean Air Act
(the Act). See Standards of Performance for New Residential
Wood Heaters, New Residential Hydronic Heaters and Forced-
Air Furnaces, 80 Fed. Reg. 13,672, 13,708-09, 13,721 (Mar.
16, 2015) (hereinafter 2015 Rule). Based on the danger wood
smoke poses to human health, the Act requires EPA to regulate
emissions from residential wood heaters. Such devices heat
many American homes—hundreds of thousands in the amici
states alone—and produce pollution across the country, which
transcends geographic borders and exposes major portions of
the population to grave health consequences.
Unlike industrial and commercial facilities regulated by
EPA under the Act, residential wood heaters are mass-
produced consumer items typically purchased by individuals
and installed and operated in private homes. To account for
differences between residential and industrial or commercial-
scale sources, and in recognition that residential wood heater
manufacturers are often small businesses, EPA regulates those
heaters through a certification program. Instead of requiring
the testing of every heater, the program allows manufacturers
to obtain certification to sell an entire model line based on
satisfactory emissions testing of a single representative heater
in the model line. As support for a certification application, the
Agency accepts test results from private, EPA-approved
laboratories that the manufacturers hire to test their heaters.

-- 3 of 32 --

4
To guard against errors at the certification stage and assure
continued compliance, the Agency may randomly select
heaters from model lines certified under the program for audit
testing by an EPA-approved laboratory at the manufacturer’s
expense. All audit laboratories are required to use the same test
method. Under its original rule, EPA called on the same
laboratory that had done the certification testing to do any audit
testing. See Standards of Performance for New Stationary
Sources; New Residential Wood Heaters, 53 Fed. Reg. 5,860,
5,871, 5,878 (Feb. 26, 1988) (hereinafter 1988 Rule). In the
first update to the rule in more than 25 years, the 2015 Rule
authorizes audit testing by any accredited and EPA-approved
laboratory.
Petitioner HPBA asserts that the audit provision of the
2015 Rule is invalid because, unlike the 1988 Rule, it
authorizes testing at other labs and neither accounts specifically
for interlaboratory variability in emissions testing nor
adequately acknowledges and explains the change. The
petition highlights that wood heaters’ emissions of particulate
matter fluctuate, posing persistent regulatory challenges.
HPBA and EPA agree that some variability in test results arises
from natural variations inherent in wood that affect how it
burns. And HPBA does not dispute the validity of the 2015
Rule’s amended emissions limits for initial certification of
wood-stove room heaters, 40 C.F.R. § 60.532, or central
heaters, 40 C.F.R. § 60.5474—two subcategories of wood
heaters regulated under the Act. It takes issue here only with
the amended compliance auditing provisions for those heater
categories. See id. §§ 60.533(n), 60.5475(n).
The challenge to the audit provision hinges on the
assertion that the use of different labs raises a distinct problem:
variable results of the same test on the same model heater,
which HPBA asserts derive in significant part from running the

-- 4 of 32 --

5
test at a different laboratory from the one used for initial
certification. HPBA claims the 2015 Rule’s allowance for
audit testing by any EPA-certified laboratory is arbitrary and
capricious insofar as it does not provide more leeway
specifically to account for interlaboratory test variability. In
other words, HPBA argues that, when a laboratory other than
the one that did the certification-stage test performs an audit,
EPA must allow worse test results, and that it has failed to
explain why it does not.
There is little question that, if it were starting from a clean
slate, EPA has provided substantial evidence and rational
explanation that would suffice to sustain the audit provisions
of the 2015 Rule. The crux of HPBA’s claim is what it treats
as EPA’s promise in the original rule it promulgated in 1988 to
continue “restricting where and how audit testing could occur,
at least until EPA studied and better understood interlaboratory
variability.” HPBA Br. 3. HPBA hangs more weight on two
sentences in the preamble to the 1988 Rule than they can bear.
And in fact, when EPA proposed the current Rule, it explained
the evolution of its understanding of test variability. It
described how, based on analyses of testing proficiency data
and improved testing methods developed since 1988, concerns
about interlaboratory audit testing as a distinct source of
variability were shown to have been overstated. And it refined
the audit procedures to address identified causes of variability.
Because we conclude that EPA acknowledged and
adequately explained the changes in the 2015 Rule, and
substantial evidence in the record supports those changes, we
deny the petition for review.

-- 5 of 32 --

6
I. BACKGROUND
A. EPA’s prior regulation of wood heaters
Residential wood heaters are enclosed, wood-burning
devices such as wood stoves. They are most often installed in
homes for ambiance or to heat the space immediately
surrounding the device. Such heaters emit many pollutants
harmful to human health. EPA promulgated its first residential
wood heater rule under Section 111(b) of the Clean Air Act in
1988, see 1988 Rule, 53 Fed. Reg. at 5,860, and first revisited
the standards in the 2015 Rule at issue here, see 2015 Rule, 80
Fed. Reg. at 13,672-73.
“Congress enacted the Clean Air Act . . . ‘to protect and
enhance the quality of the Nation’s air resources so as to
promote the public health and welfare and the productive
capacity of its population.’” Clean Wis. v. EPA, 964 F.3d 1145,
1153 (D.C. Cir. 2020) (per curiam) (citation omitted) (quoting
Miss. Comm’n on Env’t Quality v. EPA, 790 F.3d 138, 144
(D.C. Cir. 2015) (per curiam), judgment entered, 812 F. App’x
4 (D.C. Cir. 2020)). Section 111 of the Act, which Congress
added in 1970, directs EPA to regulate any new stationary
sources of air pollutants that “cause[], or contribute[]
significantly to, air pollution” and that “may reasonably be
anticipated to endanger public health or welfare.” Am. Lung
Ass’n v. EPA, 985 F.3d 914, 930 (D.C. Cir. 2021) (alterations
in original) (citing 42 U.S.C. § 7411(b)(1)(A)). Pursuant to
Section 111(b), once EPA has identified a pollutant that
endangers the public health or welfare and a category of source
that emits that pollutant, the Agency must issue mandatory
standards of performance for new and newly modified sources
in the category. 42 U.S.C. § 7411(b)(1)(B); see also id.
§ 7411(a)(1), (e). EPA is required to set standards that
“reflect[] the degree of emission limitation achievable through

-- 6 of 32 --

7
the application of the best system of emission
reduction.” Id. § 7411(a)(1). In setting those standards, the
Agency must also account for “the cost of achieving”
emissions reductions, as well as health, environmental, and
energy considerations. Id. The Act requires that the Agency
review those standards “at least every 8 years” insofar as
“appropriate in light of readily available information on the
efficacy of [the] standard[s],” and that it revise the standards as
necessary. Id. § 7411(b)(1)(B).
EPA promulgated new-source performance standards to
regulate emissions from residential wood heaters under Section
111(b)(1) of the Clean Air Act for the first time in 1988. See
1988 Rule, 53 Fed. Reg. 5,860. Wood heaters emit what EPA
refers to as residential wood smoke, which includes particulate
matter. Standards of Performance for New Residential Wood
Heaters, New Residential Hydronic Heaters and Forced-Air
Furnaces, and New Residential Masonry Heaters, 79 Fed. Reg.
6,330, 6,332, 6,337 (Feb. 3, 2014) (hereinafter 2014 Proposed
Rule). Particulate matter—essentially small particles of
pollutants—has a variety of negative human health effects,
including shortness of breath, heart problems, and premature
death. Likewise, negative health effects from harmful
chemicals in wood smoke, such as carbon monoxide,
formaldehyde, nitrogen oxide, and polycyclic aromatic
hydrocarbons, range in severity from allergic reactions to lower
birthweights and increased risk of death for newborns of
mothers exposed to wood smoke. Formaldehyde interferes
with proper lung function and can cause cancer in humans and
animals. Nitrogen oxide affects the lungs and impairs the
body’s ability to fight respiratory infections. Each year,
“residential wood combustion accounts for 44 percent of total
stationary and mobile polycyclic organic matter . . . emissions,
which account for nearly 25 percent of all area source air toxics
cancer risks and 15 percent of noncancer respiratory effects.”

-- 7 of 32 --

8
2015 Rule, 80 Fed. Reg. at 13,673 (footnote omitted). EPA has
noted that the most harmful health effects from residential
wood smoke tend to be felt by older adults, children, and
people with preexisting heart or lung disease, and that its harms
are disproportionately borne by low-income populations and
people of color. Because standards controlling particulate
matter also effectively limit accompanying chemical
pollutants, EPA has chosen to establish a particulate matter
limit. 1988 Rule, 53 Fed. Reg. at 5,861. In the 2015 Rule, EPA
concluded that “the public health benefits of this rule outweigh
the costs by more than 100 times.” 80 Fed. Reg. at 13,673.
In the 1988 Rule, the Agency chose a certification-and-
audit approach to regulate residential wood heaters, which
differs substantially from how the Agency regulates industrial
and commercial-scale sources. EPA’s regulation of
commercial and industrial sources typically imposes new
source performance standards on every new industrial source
and requires compliance testing and monitoring to verify each
source’s adherence to those standards. See Standards of
Performance for New Stationary Sources; Standards of
Performance for New Sources; Residential Wood Heaters, 52
Fed. Reg. 4,994, 5,010 (Feb. 18, 1987) (hereinafter 1987
Proposed Rule). In the proposed version of the original wood-
heater rule, EPA noted that, because “[t]he cost of a test series
is several times the cost of a typical wood heater,” and wood
heaters are mass-produced consumer items, “a compliance
scheme requiring that each facility be tested would be very
costly.” Id. at 4,995. To avoid imposing costs that could prove
prohibitive for the small businesses manufacturing wood
stoves for residential use, EPA instead provided for categorical
certification of an entire model line based on testing of a
representative heater, backstopped by a program of random
audit testing to deter and correct for errors at certification and
to ensure continued compliance. That regulatory approach

-- 8 of 32 --

9
minimizes the costs of both certification and audit testing and
spreads the audit-testing costs evenly among manufacturers.
See 1988 Rule, 53 Fed. Reg. at 5,861; 1987 Proposed Rule, 52
Fed. Reg. at 4,995, 5,010.
A manufacturer seeking certification of its residential
wood heaters under the 1988 Rule could elect to test a
representative heater for a model line and submit the results in
support of its application to EPA for the certification needed to
sell that model line. 1988 Rule, 53 Fed. Reg. at 5,861. The
1988 Rule required certification-stage emissions testing to be
conducted in an EPA-accredited laboratory using one of four
EPA-approved emissions sampling methods. Id. at 5,860. The
regulations specified procedures for each step of the emissions
testing, including how to set up the heater, load test fuel into
the heater, and operate the heater. Id. at 5,861. Tests were to
be done using a “crib” wood configuration as fuel, with spacers
and lumber of specified size. Id. at 5,901; see also id. at 5,911;
2015 Rule, 80 Fed. Reg. at 13,677-78. EPA phased in the
emissions limits over two years, resulting in emissions limits
by 1990 of 4.1 grams of particulate matter per hour (g/hr) for
catalytic wood heaters, and 7.5 g/hr for noncatalytic wood
heaters. 1988 Rule, 53 Fed. Reg. at 5,860.
The 1988 Rule authorized EPA to conduct random
compliance audits. It provided that EPA would direct the
laboratory that had done the certification-stage testing to retest
a heater from any model line at the manufacturer’s expense,
using the same test method and procedure used during
certification testing for that model line; in response to
manufacturer concerns, EPA conditionally deferred
authorizing audit testing at any laboratory other than the
certifying laboratory. Id. at 5,878. If a wood heater failed to
meet the applicable emissions limit during audit testing, EPA
would propose revocation of the certificate of compliance for

-- 9 of 32 --

10
that model line. Id. A manufacturer could then seek a hearing
to present its rebuttal case against the proposed revocation
before EPA would suspend the certification. Id. at 5,878-79.
In promulgating the 1988 Rule, EPA recognized that
emissions testing of wood-burning heaters could be imprecise.
The Agency acknowledged a commenter’s concerns about
variability in the results of both tests of a single appliance
conducted at the same laboratory (intralaboratory variability)
and tests of a single appliance across different labs
(interlaboratory variability). See id. at 5,870-71. EPA noted:
The Agency is collecting additional data to
determine the expected precision before enforcement
audits are conducted on an interlaboratory basis.
Analyses of these data will be conducted in a
statistically sound manner and the results will be
published when available. As described in the
preamble to the proposal, the interlaboratory
precision value assumed during the negotiations was
1 g/hr. If the results of the interlaboratory analysis
show a value greater than this is appropriate, the
interlaboratory component of precision will be used
in evaluating audit data.
Id. at 5,871.
The “limited amount of data available” to EPA when it
promulgated the 1988 Rule reflected variability in repeated
testing by the same laboratory “within 1 g/hr for a four test run
average.” See id. Citing unresolved concerns about
“intralaboratory precision,” id., among others, EPA set the
permissible emissions rate to be significantly more lenient than
rates many existing heaters had achieved, see id. at 5,870-71.
Given manufacturers’ concerns and the scant data available,
EPA added a 1 g/hr margin to the emissions limits to account

-- 10 of 32 --

11
for the intralaboratory precision value and deferred authorizing
enforcement audits at laboratories other than the certification-
test lab. Id. at 5,871. Referring to negotiations with
manufacturers in the rulemaking process, EPA’s statement that
“the interlaboratory precision value assumed during the
negotiations was 1 g/hr,” id. (emphasis added), suggests the
shared assumption at the time was that variation among results
of tests conducted in the same way on the same model heater
by a different laboratory was the same as variation of results
from the same laboratory. Noting that it was “collecting
additional data to determine the expected precision before
enforcement audits are conducted on an interlaboratory basis,”
EPA anticipated that if further analysis demonstrated
variability greater than 1 g/hr it would consider the
“interlaboratory component of precision . . . in evaluating audit
data,” id., and presumably account for it in any authorization
of testing by different labs.
B. History of the 2015 Rule
In 2014, EPA proposed updating the performance
standards for residential wood heaters. It did so in light of the
proliferation of cleaner and more efficient designs and
technology for such devices, and in response to lawsuits by
several states and environmental groups seeking to force the
Agency to fulfill its statutory duty to update its residential
wood heater standard pursuant to Section 111(b)(1)(B) of the
Clean Air Act. See 2014 Proposed Rule, 79 Fed. Reg. at 6,338;
see also State Amici Br. 6-7. After notice and comment, EPA
finalized revised standards in 2015. 2015 Rule, 80 Fed. Reg.
13,672. The 2015 Rule applies prospectively to heaters
manufactured after its effective date. Id. at 13,676.
The 2015 Rule maintains the certification-and-audit
regulatory framework established in the 1988 Rule. It expands

-- 11 of 32 --

12
the applicability of the standard, which under the 1988 Rule
had only applied to “room heaters” that heat an area
immediately surrounding the device, to also cover “central
heaters” used to warm larger areas with air or liquid heated in
the device and distributed via ducts or pipes throughout the
residence. See id.; 40 C.F.R. § 60.5473.
Commenting on the proposed rule, petitioner HPBA urged
EPA to abandon audit testing altogether, or at least to authorize
audit tests by a laboratory other than the one that did the
certification testing only if it employed less stringent emissions
limits to account for what HPBA asserted was greater inter-
than intralaboratory variability. Comments of Hearth, Patio &
Barbeque Association on EPA’s Proposed Standards, 79 Fed.
Reg. 6,330 (Feb. 3, 2014) at 30 (hereinafter HPBA Comment).
HPBA pointed to a 2010 study by Rick Curkeet and Robert
Ferguson. Id. (citing Rick Curkeet & Robert Ferguson, EPA
Wood Heater Test Method Variability Study (Oct. 6, 2010)
(hereinafter Curkeet-Ferguson Study)). That study generally
concluded that “[v]ariability in wood heater emission testing
results for any given appliance is most likely a function of the
random nature of burning wood, no matter how tightly you try
to control the process.” Curkeet-Ferguson Study 19. The
authors acknowledged that testing is “certainly capable of
reliably distinguishing between good and bad performance,”
but nonetheless concluded that “it cannot reliably distinguish
between ‘good, better and best’ performance.” Id. In urging
EPA to restrict audit testing to the certification-test lab, HPBA
highlighted the study’s conclusion that variability could range
from 4.9 to 9.8 grams per hour higher or lower “if the appliance
were tested again at a different laboratory” from the certifying
lab. Id. at 8; see also HPBA Comment 69 n.179.
EPA chose not to restrict audit testing to the certification
test lab. In its response to comments on the 2014 Proposed

-- 12 of 32 --

13
Rule, EPA explained that inherent methodological and
computational flaws in the study on which HPBA relied
seriously undermined that study’s conclusions regarding the
poor reproducibility of proficiency testing. EPA pointed to
other data showing that testing variability could be sufficiently
addressed by using only the test method shown to be the most
reliable and adding regulatory incentives for laboratories to
report accurate results. EPA Response to Comment on
Proposed Rule (Feb. 2015) 236 (hereinafter Response to
Comments). Remaining open to the prospect there might
nonetheless be some degree of interlaboratory variability not
addressed by those measures, EPA included in the audit
provisions an opportunity for manufacturers to submit
evidence to rebut a proposed certification-revocation decision,
thereby allowing a manufacturer to present evidence that, for
example, a finding of excess emissions should be discredited
due to test variability. The 2015 Rule also accounted for
potential residual testing variability by adding a compliance
margin to the emissions limit announced in the proposed rule.
In HPBA’s view, EPA failed adequately to respond to its
comments, adopting audit provisions in the 2015 Rule that are
unsupported by the record. Three changes effected by the 2015
Rule provide relevant context for HPBA’s challenge.
First, the 2015 Rule refines how emissions testing of wood
heaters is to be conducted. Whereas the 1988 Rule approved
four different test methods, EPA recognized by 2015 that
better-controlled methods decreased the variability of results.
EPA thus limited its approval to a single method, Method 28,
an improved version of the most repeatable (by the same lab)
and reproducible (by different labs) of the four testing methods
allowed under the 1988 Rule. See 2015 Rule, 80 Fed. Reg. at
13,682, 13,709. Even the study on which HPBA principally
bases its challenge concluded that the variability of the best of

-- 13 of 32 --

14
the previously approved methods (Method 5G-3) is ten times
lower than that of other previously approved methods.
Curkeet-Ferguson Study 15 n.8, 19. The 2015 Rule also
authorizes an alternative configuration of the wood that may be
used in emissions testing, expanding fuel choice beyond the
“crib wood” previously required to also allow testing with
“cord wood,” which more closely resembles what a typical
homeowner might use. EPA authorized use of either crib wood
or cord wood and set different emissions limits for tests using
each wood type. See 2015 Rule, 80 Fed. Reg. at 13,684; 2014
Proposed Rule, 79 Fed. Reg at 6,340.
Second, the final audit provisions in the 2015 Rule
eliminate the requirement that the same laboratory that
conducted the original certification test for a model line also
conduct any audit testing, instead allowing audit testing under
the EPA-approved method by any accredited, EPA-approved
laboratory. EPA has also imposed a conflict-of-interest
requirement for laboratories, stating that test labs may not
receive any financial benefit from the outcome of emissions
tests. 2015 Rule, 80 Fed. Reg. at 13,710.
Third, the 2015 Rule adopts more stringent emissions
limits, to be phased in over five years. But, in response to
comments calling for the standards to incorporate a compliance
margin to accommodate testing variability, the final rule
includes less stringent emissions limits than the 2014 proposal.
Id. at 13,686-87. Whereas the 1988 Rule had set limits of 4.1
to 7.5 g/hr for room heaters, EPA set the final limits for room
heaters tested with crib wood at 2 g/hr, and for those tested with
cord wood at 2.5 g/hr. For all central heaters, which the prior
rule had not reached, EPA set the emissions limit at 0.15
pounds per million British thermal units (lb/mmBtu). Id.

-- 14 of 32 --

15
In view of the evidence before it and the other measures it
promulgated, EPA deemed it unnecessary to adopt the
manufacturers’ proposal to add more leeway to the audit
provision for any imprecision specifically associated with
interlaboratory testing. See id. In both the 2014 Proposed Rule
and its Response to Comments on that proposal, EPA explained
that imprecision attributable to testing being done by a different
lab had not been demonstrated. The Agency stated that
requiring better testing methodology and controlling for
conflicts of interest on the part of certification-test labs sufficed
to ensure emissions would be measured accurately, whether by
the same or a different EPA-approved laboratory. See 2014
Proposed Rule, 79 Fed. Reg at 6,356; Response to Comments
236.
C. Procedural history
On March 16, 2015, petitioner HPBA and four other trade
association and manufacturer petitioners sought review of the
2015 Rule in this court. This case was held in abeyance after
EPA published an advance notice of proposed rulemaking in
2018, which did not suggest any changes to the 2015 Rule but
solicited comments on specified issues to inform future
rulemaking, including the compliance date for the final phased-
in emissions standard and the question whether the audit
provisions should more specifically address variability.
Standards of Performance for New Residential Wood Heaters,
New Residential Hydronic Heaters and Forced-Air Furnaces
(Advance Notice of Proposed Rulemaking), 83 Fed. Reg.
61,585, 61,592 (Nov. 30, 2018). Once EPA responded to the
comments it had solicited, see Standards of Performance for
New Residential Wood Heaters (Final Rule), 85 Fed. Reg.
18,448, 18,453 (Apr. 2, 2020), all petitioners other than HPBA
voluntarily dismissed their petitions.

-- 15 of 32 --

16
The current petition challenges only EPA’s decision to
allow audit testing—included in the rule since it was
promulgated in 1988—to be conducted by laboratories other
than the same one that did the certification testing on the heater
in question. HPBA does not challenge the more demanding
emissions limits set by EPA in the 2015 Rule. Rather, it claims
only that authorization of testing by any accredited, EPA-
approved lab is invalid for failure to account for what HPBA
claims is testing variability that occurs when audit tests are
done by a lab other than the one the manufacturer hired at the
certification stage. HPBA argues that EPA arbitrarily reversed
course from the 1988 Rule without acknowledging its change
of course or supporting and explaining the 2015 Rule’s
approach to interlaboratory variability, all of which HPBA
contends renders the rule arbitrary and capricious.
II. JURISDICTION
Before reaching the merits, we must determine whether we
have jurisdiction. Neither EPA nor the intervenors or amici in
its support questioned HPBA’s standing, but the court has an
independent obligation to satisfy itself of its own jurisdiction.
EPA does, however, raise a ripeness challenge to the petition.
So we first confirm that HPBA has standing and that its petition
is ripe. See Ctr. for Sustainable Econ. v. Jewell, 779 F.3d 588,
596 (D.C. Cir. 2015).
A. Petitioner HPBA has demonstrated its standing
Petitioner HPBA claims “associational” or
“representational” standing to challenge the audit-testing
provision of the 2015 Rule on behalf of its member
businesses—as distinct from “organizational” standing on its
own behalf. See Hunt v. Wash. State Apple Advert. Comm’n,
432 U.S. 333, 343 (1977). “An organization has associational
standing to bring suit on its members’ behalf when: (1) at least

-- 16 of 32 --

17
one of its members would have standing to sue in his or her
own right; (2) ‘the interests it seeks to protect are germane to
the organization’s purpose’; and (3) ‘neither the claim asserted
nor the relief requested requires the participation of individual
members in the lawsuit.’” Sierra Club v. FERC, 827 F.3d 59,
65 (D.C. Cir. 2016) (quoting WildEarth Guardians v. Jewell,
738 F.3d 298, 305 (D.C. Cir. 2013)).
It is HPBA’s burden as the party invoking federal
jurisdiction to demonstrate standing, “which must be supported
in the same way as any other matter on which the party bears
the burden of proof.” Twin Rivers Paper Co. v. SEC, 934 F.3d
607, 613 (D.C. Cir. 2019) (formatting modified). We review
agency action on the administrative record—akin to our review
of a district court’s decision on cross motions for summary
judgment in a case in which the factual record is uncontested
and the parties’ dispute is over legal implications. See CTS
Corp. v. EPA, 759 F.3d 52, 64 (D.C. Cir. 2014). “It is ‘black-
letter administrative law that in an [Administrative Procedure
Act] case, a reviewing court should have before it neither more
nor less information than did the agency when it made its
decision.’” Id. (quoting Hill Dermaceuticals, Inc. v. FDA, 709
F.3d 44, 47 (D.C. Cir. 2013)).
Because Article III’s constraints apply to courts and not
executive agencies, the obligation to establish standing in a
case that originates in an agency arises with the first court
filing. Regarding challenges to agency action that may be filed
directly in this court, we have noted that
[i]n many if not most cases the petitioner’s standing
to seek review of administrative action is self-
evident; no evidence outside the administrative
record is necessary for the court to be sure of it. In
particular, if the complainant is “an object of the

-- 17 of 32 --

18
action (or forgone action) at issue”—as is the case
usually in review of a rulemaking and nearly always
in review of an adjudication—there should be “little
question that the action or inaction has caused him
injury, and that a judgment preventing or requiring
the action will redress it.”
Sierra Club v. EPA, 292 F.3d 895, 899-900 (D.C. Cir. 2002)
(quoting Lujan v. Defs. of Wildlife, 504 U.S. 555, 561-62
(1992)).
But sometimes the evidence supporting standing is absent
from or not readily identified in the administrative record. Our
circuit rules provide that “[w]hen the appellant’s or petitioner’s
standing is not apparent from the administrative record, the
brief must include arguments and evidence establishing the
claim of standing.” D.C. Cir. R. 28(a)(7) (concerning contents
of briefs and requirements related to standing). We have held
that, “[w]hen the petitioner’s standing is not self-evident . . .
the petitioner must supplement the record to the extent
necessary to explain and substantiate its entitlement to judicial
review.” Sierra Club, 292 F.3d at 900.
Petitioner HPBA’s standing statement in its brief was
conclusory. The briefing did not point to anything in the
administrative record establishing that it has one or more
members that manufacture heaters regulated by the challenged
provisions of the 2015 Rule, nor did HPBA submit to us
declarations or other competent evidence to that effect.
HPBA’s counsel was unable at oral argument to name any of
HPBA’s members that are manufacturers of wood heaters
subject to the Rule. See Transcript of Oral Argument at 35-36.
The question, then, is whether it was reasonable for HPBA,
whose members appeared to include some businesses
manufacturing wood stoves, perhaps including residential

-- 18 of 32 --

19
wood heaters subject to the challenged rule, to have assumed
that its standing was “self-evident” based on the administrative
record.
We have acknowledged that the language of the circuit
rule explaining when a petitioner must make factual
submissions supplemental to the record “is hardly free from
ambiguity because what may be apparent from the
administrative record to one reasonable person may seem less
clear to another.” Ams. for Safe Access v. DEA, 706 F.3d 438,
445 (D.C. Cir. 2013) (internal quotations omitted). This court
takes care to eschew any “‘gotcha’ construction” that “would
have the undesirable effect of causing parties to include long
jurisdictional statements in practically all opening briefs for
fear that the court might find their standing less than self-
evident.” Am. Library Ass’n v. FCC, 401 F.3d 489, 494 (D.C.
Cir. 2005). After all, “[t]his would waste, rather than conserve,
judicial resources and place an unnecessary burden on
litigants.” Id.
That said, because “[t]here may be a disjunction between
what a petitioner assumes that the court knows about its
organization and operation and what the court actually knows,”
id. at 494-95, and because we cannot simply “accept[] [an]
organization[’s] self-description[] of [its] membership,”
Summers v. Earth Island Inst., 555 U.S. 488, 499 (2009),
“[w]hen a petitioner claims associational standing, it is not
enough to aver that unidentified members have been injured.”
Chamber of Com. of U.S. v. EPA, 642 F.3d 192, 199 (D.C. Cir.
2011). Instead, a petitioner relying upon associational standing
should explain in its opening brief how each of the challenged
agency action affects or injures one or more of its members,
identify those members, and point to evidence in the
administrative record or in a declaration that shows those
persons or entities are actually members of petitioner’s

-- 19 of 32 --

20
association and how the challenged agency action affects them.
Id.; see also Sierra Club v. EPA, 754 F.3d 995, 1001 (D.C. Cir.
2014) (standing not shown where petitioner failed to present
evidence one of its members actually suffered an injury in fact).
Nonetheless, “under the law of this circuit, the members of
a panel retain discretion to seek supplemental submissions on
standing to fulfill the obligation of the court to determine
whether the requirements of Article III have been met.” Ams.
for Safe Access, 706 F.3d at 444. We have invited
supplemental submissions in cases where, as here, the nature
of the business of a petitioner’s individual members or the
specific products they might manufacture, or some other
pertinent characteristic, was not sufficiently clear from the
administrative record to allow us to discern a claim of Article
III injury. See Am. Library Ass’n, 401 F.3d at 496. See also
generally McNary v. Fed. Mine Safety & Health Rev. Comm’n,
989 F.3d 21, 23 (D.C. Cir. 2021) (supplemental submission
invited on redressability); Sissel v. U.S. Dep’t of Health &
Hum. Servs., 760 F.3d 1, 5 (D.C. Cir. 2014) (same on plaintiff’s
employment and income); Pub. Citizen, Inc. v. Nat’l Highway
Traffic Safety Admin., 489 F.3d 1279, 1296-97 (D.C. Cir. 2007)
(same on risk of harm to organization’s members); Ctr. for
Sustainable Econ., 779 F.3d at 598-99 (same on organization’s
mission and bylaws); Am. Chemistry Council v. Dep’t of
Transp., 468 F.3d 810, 815, 819 (D.C. Cir. 2006) (same on
nature and imminence of harm to organization’s members).
Following oral argument, we invited HPBA to submit
information not included in the Joint Appendix to point out to
the court the support they referenced at argument for their very
general assertions that they have members directly regulated by
the 2015 Rule. HPBA by letter submitted the full text of its
comment to EPA, which identified Jotul and Pacific Energy as
member manufacturers of room wood heaters and central wood

-- 20 of 32 --

21
heaters. See Pet’r Letter Providing Information Requested by
the Court During January 28, 2021 Oral Argument, ECF No.
1882541, at 168-78 (Jotul); id. at 179-81 (Pacific Energy).
Those are the two types of heaters covered by the other-lab
audit authorization HPBA challenges. HPBA should have
included those parts of the administrative record in the Joint
Appendix and should have specifically explained and
referenced them in its opening brief. See D.C. Cir. R. 28(a)(7).
Nonetheless, we are satisfied that HPBA’s post-argument
submission suffices to establish standing to challenge the 2015
Rule’s audit provisions as they apply to both room heaters and
central heaters.
Having established that at least one HPBA member has
standing to challenge the audit provisions for each heater type,
we turn to a second threshold question: whether that challenge
is ripe.
B. Petitioner HPBA’s claims are ripe for review
EPA argues that HPBA’s challenge to the 2015 audit
provisions is unripe because “EPA has not yet audited a single
device under the 2015 Rule,” and “HPBA’s hypothetical
concerns of revocation following a failed audit test may never
come to pass.” EPA Br. 37. In the event the Agency issues a
notice of revocation of a device’s certification as the result of a
failed audit, EPA points out that it may consider “any relevant
information” at a rebuttal hearing, including evidence of the
variability HPBA claims must be taken into account at the audit
stage. 40 C.F.R. § 60.533(n)(3)(v).
Those arguments characterize the petition too narrowly.
“Determining whether administrative action is ripe for judicial
review requires us to evaluate (1) the fitness of the issues for
judicial decision and (2) the hardship to the parties of
withholding court consideration.” Cement Kiln Recycling

-- 21 of 32 --

22
Coal. v. EPA, 493 F.3d 207, 214 (D.C. Cir. 2007) (quoting
Nat’l Park Hosp. Ass’n v. Dep’t of Interior, 538 U.S. 803, 808
(2003)). It is well established that “a purely legal claim in the
context of a facial challenge . . . is presumptively reviewable.”
Id. at 215 (quoting Nat’l Ass’n of Home Builders v. U.S. Army
Corps of Eng’rs, 417 F.3d 1272, 1282 (D.C. Cir. 2005)); see
Nat’l Ass’n of Home Builders, 417 F.3d at 1282 (“We have
repeatedly held that claims that an agency’s action is arbitrary
and capricious or contrary to law present purely legal issues.”
(formatting modified)). HPBA facially challenges “EPA’s
authority to require audits to be performed at a laboratory of
EPA’s choosing, without accounting for interlaboratory
variability . . . not whether any individual manufacturer might
somehow avoid revocation despite failing an audit.” Reply
Br. 3. That claim “presents purely legal issues that must be
decided based on the administrative record.” Id. “And because
‘Congress has emphatically declared a preference
for immediate review’ with respect to Clean Air Act
rulemaking, we have no need to consider the ripeness test’s
second element, namely, the hardship to the parties of
withholding review.” Nat. Res. Def. Council v. EPA, 643 F.3d
311, 320 (D.C. Cir. 2011) (quoting Cement Kiln Recycling
Coal., 493 F.3d at 215).
Petitioner HPBA has standing and its challenge to the 2015
Rule’s audit provisions is ripe, so we turn to the merits of its
claim that those provisions are arbitrary and capricious.

-- 22 of 32 --

23
III. MERITS
We review the 2015 Rule pursuant to 42 U.S.C.
§ 7607(b)(1). Under that section, agency action may be
reversed if it is arbitrary, capricious, an abuse of discretion, not
in accordance with law, or in excess of statutory authority.
Id. § 7607(d)(9)(A), (C); see Am. Petrol. Inst. v. EPA, 684 F.3d
1342, 1347 (D.C. Cir. 2012). To withstand review, an agency
must have examined all relevant facts and data and articulated
a rational explanation for its decision, including a reasonable
connection between the facts and ultimate outcome. Am.
Petrol. Inst., 684 F.3d at 1347-48. “We cannot look at EPA’s
decision as would a scientist, but instead must exercise our
‘narrowly defined duty of holding agencies to certain minimal
standards of rationality.’” Murray Energy Corp. v. EPA, 936
F.3d 597, 608 (D.C. Cir. 2019) (quoting Mississippi v. EPA,
744 F.3d 1334, 1342 (D.C. Cir. 2013)).
HPBA argues that EPA arbitrarily reversed course from
the 1988 Rule in the 2015 Rule’s audit provisions. In HPBA’s
view, the 1988 Rule appropriately addressed both
interlaboratory and intralaboratory variability in test results by
(1) setting emissions limits at levels that assume variability in
test results if a device is tested in the same laboratory and (2)
accounting for imprecision of tests done by different
laboratories by prohibiting audit testing at a laboratory other
than the one that did the testing on which the model-line
certification was based. HPBA Br. 8-11. HPBA claims EPA
acted arbitrarily and without record support in the 2015 Rule
by no longer restricting audit testing to the same lab that did
the certification-stage testing and, in particular, that it could not
lawfully do so without first making specific findings that
interlaboratory variability is no greater than intralaboratory
variability, or at least explaining why it would not. See id. 15-
16, 24-26; Reply Br. 16.

-- 23 of 32 --

24
We conclude that the Agency adequately explained the
changes in the 2015 Rule’s audit provisions and that the
changes are supported by the record. In proposing those
changes in 2014, EPA first explained the refinement of its
understanding. The Agency described how it was amending its
approach to concerns about variability of test results based on
more recent testing proficiency data and analyses, and
improved testing methods. The 2015 Rule does account for
potential variability in several ways. Although EPA did not
adopt the approach HPBA favored, the challenged regulatory
choices are supported by substantial evidence and neither
arbitrary nor capricious.
A. EPA considered the data that HPBA submitted
HPBA’s disagreement with EPA’s interpretation of the
available testing proficiency data is the crux of its challenge to
the 2015 Rule’s audit provisions, so we begin by reviewing
how EPA considered and explained the data. HPBA had
commissioned studies in 2010 that reported a high level of
variability in test results. See generally Curkeet-Ferguson
Study. HPBA contends that EPA failed to account for its data,
instead simply “ignoring the Curkeet Ferguson Study in the
record.” HPBA Br. 31. That argument is squarely contradicted
by the Agency’s direct engagement with the Curkeet-Ferguson
Study in its 2014 Proposed Rule, Response to Comments, and
the measures adopted in the final 2015 Rule.
First, in the 2014 Proposed Rule, EPA anticipated HPBA’s
concern about reproducibility of test results. EPA’s proposal
expressly disagreed that “repeatability [by the same lab] and
reproducibility [by a different lab] of the current test method
for wood heater emissions . . . may be poor.” 2014 Proposed
Rule, 79 Fed. Reg. at 6,356; see id. at n.46 (citing Robert
Ferguson, Final Report: EPA Wood Heater Emission Test

-- 24 of 32 --

25
Method Comparison Study (Dec. 1, 2010)). The Agency
explained that the reported variability likely derived from “the
lack of regulatory requirements or incentives for the test
laboratories to achieve highly reproducible results in
proficiency testing.” Id. at 6,356. EPA addressed that lacuna
not only with a margin for variability in the emissions limits
themselves, but also with more rigorous testing requirements
and a conflict-of-interest provision. And it observed that the
results of the tests HPBA identified, which were conducted
under the 1988 Rule, “do not reflect the proposed changes to
improve the repeatability and reproducibility of the test
method.” Id.
Second, EPA published a study of testing variability it
commissioned from Brookhaven National Laboratory
(Brookhaven), which, the Agency explained, showed that the
“[r]epeatability of cord wood test method” for room heaters can
be “very good.” Notice of Data Availability in Support of the
2014 Proposed Rule, 79 Fed. Reg. 37,259, 37,261 (July 1,
2014) (hereinafter Notice of Data Availability). Brookhaven
conducted emissions testing of a “popular non-catalytic EPA-
certified wood stove using cord wood,” id., at three of the four
burn rates required by EPA Method 28, see Butcher, T. et al.,
Brookhaven Nat’l Lab., Cord Wood Testing in a Non-Catalytic
Wood Stove 6 (June 30, 2014) (hereinafter Brookhaven).
Results for three tests that Brookhaven ran at the highest burn
rate were within fifteen percent of each other. Id.; see also
Notice of Data Availability, 79 Fed. Reg. at 37,261. At the two
other, lower burn rates tested, the results were within three and
ten percent of each other. Brookhaven 7, 17. EPA thus
gathered and described record evidence directly relevant to the
issue on which HPBA claims the Agency lacked support.
Following publication of the 2014 Proposed Rule, HPBA
submitted a comment pressing its interlaboratory-variability

-- 25 of 32 --

26
concern. The comment asserted that test results under the 1988
Rule were highly variable and argued that EPA in the 1988
Rule “expressly obligated itself to evaluate interlab precision
and account for it by adjusting the stringency of the standards,
if necessary, by amending them through a rulemaking
proceeding.” HPBA Comment 10. Asserting that data showed
interlaboratory variability preventing EPA from allowing
audits by different laboratories, HPBA characterized the
Curkeet-Ferguson Study as providing confirmation of “the
variable nature of wood heater emissions testing” and
establishing that “the major contributor to variability was the
random nature of burning wood.” Id. at 30; see Curkeet-
Ferguson Study 7-8. Of course, as EPA has consistently
acknowledged, variability caused by “the random nature of
burning wood” affects all testing, whether by the same or a
different lab; but the Curkeet-Ferguson Study drew
conclusions it ascribed to the use of different labs. Curkeet and
Ferguson tested three wood heaters in different laboratories
under the 1988 standards and calculated that the average
emissions results for the tested heaters were 3.91, 14.01, and
6.35 g/hr. They concluded that “for any emissions rate
measured using the EPA test methods, the result could be 4.9
to 9.8 grams per hour higher or lower if the appliance were
tested again at a different laboratory.” Curkeet-Ferguson Study
8.
Third, EPA further considered HPBA’s position when, in
its Response to Comments on the 2014 Proposed Rule, EPA
explained that a statistical analysis by Washington State’s
Puget Sound Clean Air Agency (Puget Sound) challenged the
Curkeet-Ferguson Study’s conclusions regarding the
magnitude of testing variability. Response to Comments 236.
Puget Sound evaluated that study and its underlying data and
identified several deficiencies that together showed the study’s
data set was neither representative nor reliable, and that any

-- 26 of 32 --

27
real testing variability was far less than Curkeet and Ferguson
calculated. See Craig Kenworthy, Puget Sound Clean Air
Agency, Preliminary Review and Critique of Analyses of
NSPS Test Method Variability (Curkeet, 2010) and the
Relationship of EPA Certified Values to “In-Home Use”
(Houck, 2012) (Dec. 5, 2012) 1, 4-5. Puget Sound identified
three main deficiencies in the Curkeet-Ferguson Study. First,
the study incorrectly applied statistics appropriate for a normal
distribution to “non-normal data.” Id. at 5. Second, it
improperly divided up its data set to draw conclusions from a
subset of the data that was not representative of the data set as
a whole, thereby allowing particularly extreme values to be
highlighted and taken out of context. Id. And third, it conflated
absolute difference in values with a confidence interval,
leading to a dramatic overestimation of uncertainty. Id. EPA’s
consideration of Puget Sound’s detailed critique of the
Curkeet-Ferguson Study supports the Agency’s decision to
discount the latter and the conclusions HPBA draws from it.
A fourth way that EPA engaged with HPBA’s comment
and the study it proffered was by describing and crediting yet
another critical assessment of the Curkeet-Ferguson Study,
conducted by a wood heater manufacturer, the Woodstock
Soapstone Company (Woodstock Soapstone). The Woodstock
Soapstone study further supported EPA’s conclusion that the
Curkeet-Ferguson Study misattributed variable testing results
to interlaboratory variability rather than the inefficiency of
outmoded test methods the 2015 Rule abandoned, and the
conflicted interests of the laboratories conducting the tests.
Woodstock Soapstone submitted comments on the 2014
Proposed Rule that included a comparison of the data used in
the Curkeet-Ferguson Study to Woodstock Soapstone’s more
recent data for three room heaters. See Morrissey, Woodstock
Soapstone Company, Comment (Aug. 15, 2014).

-- 27 of 32 --

28
Woodstock Soapstone found variability in test results for
its own heaters ranging from 0.17 to 0.61 g/hr—far less
variability than reported by Curkeet and Ferguson. Woodstock
Soapstone attributed the large difference between the two
studies’ reported variability to “deep[] flaw[s]” in the Curkeet-
Ferguson analysis. Id. 1. The Curkeet-Ferguson data set was
“very small” and some of the “data [wa]s up to 27 years old.”
Id. 3. And, strikingly, Woodstock Soapstone observed in
Curkeet and Ferguson’s data that initial certification tests of
heaters almost always yielded results showing lower levels of
relevant pollutants than subsequent tests. That skew did not
seem to reflect random chance. Woodstock Soapstone
calculated that, all else being equal, there was only a 0.52%
chance that the pattern of audit test results so much higher than
certification test results could be attributed to random chance.
Id. 1. The most reasonable interpretation of the variability in
the Curkeet-Ferguson data set, Woodstock Soapstone
concluded, was that the certification tests were conducted in a
different manner than the subsequent tests. Id.
In all these ways, the Agency explained how it based its
current thinking on additional data and critical analysis of the
study HPBA had presented as substantiating its original
concerns. We conclude that EPA has “adequately explained its
reasons for rejecting the data” HPBA highlighted. City of
Waukesha v. EPA, 320 F.3d 228, 248 (D.C. Cir. 2003). The
Agency “specifically analyzed and responded to” the Curkeet-
Ferguson Study, permissibly concluding that additional
analyses showed testing variability to be a less fundamental
problem than HPBA asserted. Int’l Fabricare Inst. v. EPA, 972
F.2d 384, 394 (D.C. Cir. 1992). The Agency acted responsibly
by taking account of the more recent analysis of the older data
used in the Curkeet-Ferguson Study, and by considering the
newer data used by Brookhaven and in the comment from
Woodstock Soapstone. See Dist. Hosp. Partners, L.P. v.

-- 28 of 32 --

29
Burwell, 786 F.3d 46, 57 (D.C. Cir. 2015) (“[A]n agency
cannot ignore new and better data.”); Catawba County v.
EPA, 571 F.3d 20, 45 (D.C. Cir. 2009) (noting that agencies
“have an obligation to deal with newly acquired evidence in
some reasonable fashion”). While “[r]easonable people might
disagree with EPA’s interpretations of the scientific evidence,”
relative institutional competencies generally mean that “any
such disagreements must come from those who are qualified to
evaluate the science, not [this court].” Mississippi v. EPA, 744
F.3d at 1345. We see no grounds to hold that EPA acted
arbitrarily or without substantial evidentiary support in relying
on new data and expert analyses of the Curkeet-Ferguson Study
to reassess the significance of testing variability.
B. EPA explained its changed position
There is also no reason to conclude that EPA made an
unexplained about-face. The Agency spelled out in the 2014
Proposed Rule and the Response to Comments how and why
its understanding had developed. In the 2014 Proposed Rule,
EPA made clear that it had further considered any potential
variability in emissions testing results and did not interpret the
available data to show that emissions from room heaters could
not be accurately measured. See 79 Fed. Reg. at 6,356. At the
same time, contrary to HPBA’s assertion, the 2014 Proposed
Rule did not deny all variability in testing. Rather, based in
part on evidence of successful implementation of more
stringent emissions limits in some states, EPA explained that it
“believe[d] that previous [testing] results” appearing to
demonstrate substantial variability, while meriting
consideration, “d[id] not mean that lower emissions standards
cannot be measured accurately.” Id. The Agency reiterated
that position in the Response to Comments, where it further
explained that methodological and computational flaws in the
Curkeet-Ferguson Study undermined that study’s conclusions

-- 29 of 32 --

30
regarding the poor reproducibility of proficiency testing.
Response to Comments 236. In view of the intervening
analyses and new data, the Agency concluded, improved test
methods and the addition of regulatory incentives for
laboratories to report accurate results would more
appropriately address testing variability than would HPBA’s
preferred response of abandonment or further relaxation of
standards for audit testing at other labs.
In the 1988 Rule, EPA had set emissions limits based on
“the achievable level of wood heater performance, taking into
account several sources of variability,” and declined at that
time to authorize audit testing at different labs. 53 Fed. Reg. at
5,870. The Agency referred to the “limited . . . data available”
indicating that intralaboratory precision “was within 1 g/hr for
a four test run average.” Id. at 5,871. The key passage in the
1988 Rule that HPBA asserts created for the Agency a
“longstanding obligation to analyze imprecision/variability in
detail,” HPBA Br. 27, in fact presumptively did not distinguish
between the degrees of variability within and between testing
laboratories. EPA expressed readiness to use a different
precision value if data so indicated, and conditionally
undertook to further investigate the interlaboratory variability
that, without the benefit of evidence, it assumed might exist.
1988 Rule, 53 Fed. Reg. at 5,871. But the referenced need to
gather data on interlaboratory variability and perhaps to adjust
the precision value for audits at other labs rested on the
Agency’s understanding that there were no other adequate
explanations of the testing imprecision EPA provisionally
assigned to intra- and interlaboratory variability.
What HPBA frames as a broken promise on the part of
EPA is nothing of the sort. HPBA had long asserted that results
of the same test of the same stove model would somehow vary
wildly if a different EPA-approved lab conducted the test. But

-- 30 of 32 --

31
once EPA had data and analyses debunking the very basis for
that suggestion, any reason for EPA to gather and publish
further data on the point before updating the rule had
evaporated. In other words, the assumptions the Agency made
in 1988 when it referred to additional investigation of
interlaboratory precision no longer made sense by the time it
promulgated the 2015 Rule. Contrary to HPBA’s position,
nothing in the 1988 Rule precluded EPA from amending its
view of the interlaboratory variability issue upon further
consideration of potential causes of testing imprecision and
revising the audit provisions as it did in the 2015 Rule.
C. EPA accounted for testing variability
HPBA contends that the Agency “effectively assumes
there is no variability by failing to build imprecision into the
emissions limits themselves and by placing no limitations on
audit testing.” HPBA Br. 34. To the contrary, EPA accounted
for potential variability of test results in the 2015 Rule in at
least four ways. First, in the final version of the Rule, EPA
changed the substantive standard from the proposed rule to add
a 100% compliance margin to the emissions limit for room
heaters, and a margin of approximately 140% to the limit for
central heaters. 2015 Rule, 80 Fed. Reg. at 13,683. Second,
EPA improved the repeatability and reproducibility of
emissions testing by confining approved test methods to those
shown to be most accurate, and requiring improved methods
for cord wood testing as supported by the results of the
Brookhaven study. Id. at 13,677, 13,685; see also 2014
Proposed Rule, 79 Fed. Reg. at 6,356. Third, the Agency
included a conflict-of-interest provision to blunt apparent bias
in certification testing by prohibiting laboratories from
receiving any financial benefit keyed to the outcome of tests.
2015 Rule, 80 Fed. Reg. at 13,710. Fourth, EPA included in
the audit provisions an opportunity for manufacturers to rebut

-- 31 of 32 --

32
a decision revoking certification with any relevant
information—including proof that what the audit lab measured
as excess emissions in fact resulted from test variability. Id. at
13,709, 13,721.
EPA reasonably determined and explained that the
measures it adopted to address potential testing variability for
wood heaters were sufficient, and its decisions are supported
by substantial evidence. The Agency permissibly decided not
to adopt a more lenient emissions level for audit testing by labs
other than the certifying laboratory, and to instead rely on what
it concluded were more appropriate measures to ensure
effective and fair audit testing.
* * *
For the foregoing reasons, we deny the petition for review.
So ordered.

-- 32 of 32 --