Casey Voigt; Julie Voigt v. Coyote Creek Mining Company, LLC, a North Dakota Corporation

United States Court of Appeals
For the Eighth Circuit
___________________________
No. 18-2705
___________________________
Casey Voigt; Julie Voigt
lllllllllllllllllllllPlaintiffs - Appellants
v.
Coyote Creek Mining Company, LLC, a North Dakota Corporation
lllllllllllllllllllllDefendant - Appellee
------------------------------
State of North Dakota; Lignite Energy Council
lllllllllllllllllllllAmici on Behalf of Appellee(s)
____________
Appeal from United States District Court
for the District of North Dakota - Bismarck
____________
Submitted: October 17, 2019
Filed: November 20, 2020
____________
Before LOKEN, SHEPHERD, and STRAS, Circuit Judges.
____________
SHEPHERD, Circuit Judge.
Casey and Julie Voigt, the owners of a large ranch in rural North Dakota, filed
suit against Coyote Creek Mining Company, LLC (CCMC), alleging CCMC failed to

-- 1 of 25 --

obtain the proper construction permit under the Clean Air Act (CAA), 42 U.S.C. §
7401, et seq., and failed to implement the requisite dust control plan for the Coyote
Creek Mine, which is adjacent to the Voigts’ ranch. CCMC moved for summary
judgment on the Voigts’ claims and the Voigts moved for partial summary judgment
on issues of liability. The district court1 granted summary judgment in favor of
CCMC, concluding the federal regulations imposing permitting and dust control
requirements do not apply to CCMC’s operations. The Voigts appeal, arguing the
district court erroneously determined the regulations are ambiguous and improperly
relied on the North Dakota Department of Health (NDDOH) permitting decision to
reach its conclusion. Having jurisdiction under 28 U.S.C. § 1291, we affirm.
I.
Pursuant to the CAA, the Environmental Protection Agency (EPA) established
National Ambient Air Quality Standards (NAAQS), which are designed to improve
air quality by placing limits on six specific air pollutants, including, as relevant here,
particulate matter. 42 U.S.C. §§ 7408-09; see also Util. Air Regulatory Grp. v. EPA,
573 U.S. 302, 308 (2014). Particulate matter is the air pollutant most commonly
associated with mining operations. Areas of the country where the air quality meets
the NAAQS are called attainment areas, while areas that do not meet these standards
are known as non-attainment areas. 42 U.S.C. § 7407(d). North Dakota is an
attainment area. As part of its plan to achieve and maintain the NAAQS, the EPA
created New Source Performance Standards (NSPS), which impose emission
standards on new major sources of air pollution, including newly constructed
facilities, and on modifications to existing facilities that would increase emissions.
See Sierra Club v. Otter Tail Power Co., 615 F.3d 1008, 1011 (8th Cir. 2010).
However, because the NSPS are aimed at helping achieve and maintain the NAAQS,
1The Honorable Charles S. Miller, Jr., United States Magistrate Judge for the
District of North Dakota, now retired, to whom the case was referred for final
disposition by consent of the parties pursuant to 28 U.S.C. § 636(c).
-2-

-- 2 of 25 --

they do not prevent air quality degradation in attainment areas, like North Dakota,
where the air quality is already below NAAQS-imposed limits. See Alaska Dep’t of
Envtl. Conservation v. EPA, 540 U.S. 461, 470-71 (2004). Recognizing that this gap
existed, Congress amended the CAA to include prevention of significant deterioration
of air quality (PSD) provisions, which apply to attainment areas and impose
permitting requirements on the construction of “major emitting facilities.” 42 U.S.C.
§§ 7475, 7479(1). A major emitting facility may not be constructed until a major
source permit is obtained, which requires compliance with various regulations,
including the planned use of best available control technology for each pollutant
emitted by the facility. 42 U.S.C. § 7475(a)(4); see also Chevron, U.S.A., Inc. v.
Natural Res. Def. Council, Inc., 467 U.S. 837, 846 (1984).
There are two ways for a source to be considered a major emitting facility. See
42 U.S.C. § 7479(1). First, a source constitutes a major emitting facility if it is a
stationary source that is included on the list of specified industrial facilities that have
a potential to emit (PTE) 100 tons per year (tpy) of any air pollutant. Id. Second, any
other stationary source that has a PTE of at least 250 tpy of any air pollutant
constitutes a major emitting facility. Id. Surface coal mines are not included on the
list of specified industrial facilities subject to the 100 tpy threshold. See id.
Therefore, the only way for a surface coal mine to be considered a major emitting
facility, and thus to fall within the PSD provisions and require a construction permit,
is if it has a PTE of at least 250 tpy of any air pollutant.
As a general matter, when calculating whether a source’s PTE air pollutants
satisfies the threshold so as to constitute a major emitting facility, the source’s fugitive
emissions are excluded. Fugitive emissions are “those emissions which could not
reasonably pass through a stack, chimney, vent, or other functionally equivalent
opening.” 40 C.F.R. § 51.166(b)(20). For mining operations, fugitive emissions
generally take the form of coal dust. Although fugitive emissions are generally
excluded, the EPA has promulgated a list of categories of sources for which fugitive
emissions must be counted. See id. §§ 51.166(b)(1)(iii), 52.21(b)(1)(iii). Surface coal
-3-

-- 3 of 25 --

mines are not included on that list. Therefore, although most surface coal mines have
the PTE more than 250 tpy of dust, see Natural Resources Def. Council, Inc. v. EPA,
937 F.2d 641, 643 (D.C. Cir. 1991), those emissions consist almost entirely of fugitive
emissions and, thus, the surface coal mines do not, by themselves, constitute major
emitting facilities. The EPA has provided, however, that fugitive emissions must be
counted when calculating the PTE air pollutants for a coal processing plant. See 40
C.F.R. §§ 51.166(b)(1)(iii)(aa), 52.21(b)(1)(iii)(aa). Therefore, a coal processing
plant that has a PTE more than 250 tpy of any air pollutant, the calculation of which
includes fugitive emissions, is considered a major emitting facility. Moreover, where
the coal processing plant meets this threshold and is a part of a mining operation that
also consists of a surface coal mine, the entire mining operation is considered a major
emitting facility. Accordingly, the PSD provisions and construction permit
requirement would apply to the entire mining operation, including the surface coal
mine.
Further, in addition to the PSD provisions’ permitting requirements, generally
applicable NSPS have been established for coal processing plants that process more
than 200 tons of coal per day. These regulations are contained in Subpart
Y–Standards of Performance for Coal Preparation and Processing Plants, 40 C.F.R.
pt. 60. Among the Subpart Y requirements, an open storage coal pile in a coal
processing plant must have a fugitive dust control plan. 40 C.F.R. § 60.254(c). The
parties agree that CCMC’s coal processing plant is subject to Subpart Y; however,
they dispute which portions of CCMC’s operations constitute a part of the coal
processing plant. This is of critical importance because what portions of the operation
are part of the coal processing plant dictates which portions are subject to Subpart Y
NSPS and are included in calculating the major source PTE air pollutants threshold.
In short, those parts of the mining operation that are considered within the coal
processing plant are subject to permitting and dust control requirements simply
because the regulations distinguish between coal processing plants and surface coal
mines.
-4-

-- 4 of 25 --

This framework and these regulations are carried out through a cooperative
relationship between the EPA and individual states. The CAA delegates to states the
primary responsibility for carrying out its purposes, which states accomplish by
enacting a State Implementation Plan (SIP), which details how a state plans to comply
with the provisions of the CAA. See 42 U.S.C. § 7410. A state’s SIP is subject to
EPA approval. North Dakota has an EPA-approved SIP, which includes
administration of PSD provisions. The practical effect of this set-up is that North
Dakota, through the NDDOH, is the permitting authority for new facilities that require
a major source construction permit under the CAA. In addition to the CAA
requirements, North Dakota has adopted regulations that impose their own
requirements on new facilities that do not qualify as major sources under the CAA,
including mandating that these facilities obtain a minor source permit prior to
construction. See N.D. Admin. Code § 33.1-15-14-03. Both the major and minor
source permitting decisions are handled by the NDDOH.
CCMC mines lignite at the Coyote Creek Mine. Lignite is a low-grade coal,
which is typically consumed near the mine based on the economics of lignite
transportation. Coyote Creek Mine consists of two major components: the mine face
itself and the coal processing facility. The mine face is connected to the coal
processing facility by a private hauling road, which covers the several mile distance
between the two locations. After coal is mined, trucks transport it across the haul road
to the coal processing facility, where it is unloaded onto an open storage coal pile at
the coal processing facility. The coal pile covers an area of roughly 8 acres and can
store approximately 180,000 tons of raw, unprocessed coal and abuts a retaining wall
that separates the coal pile from the crushing equipment within the coal processing
facility. Near the top of the retaining wall is an apron feeder, which is where the coal
is fed into the crushing equipment. The apron feeder is located a significant distance
off the ground, but is rarely visible because it is typically covered by the top of the
coal pile. Coal is usually drawn into the apron feeder with the assistance of gravity,
but in the circumstances where the apron feeder is visible because the coal pile is not
high enough to cover it, CCMC uses bulldozers to push the coal directly into the
-5-

-- 5 of 25 --

feeder. Once the coal is loaded from the coal pile through the apron feeder, it is fed
through the primary and secondary crushing equipment, which are housed within an
enclosed area within the coal processing facility. Once the coal is processed, it is
again transported by conveyor system to the Coyote Station, a coal-fired electric
generating plant and CCMC’s lone customer for the Coyote Creek Mine.
Although the coal pile has a capacity of approximately 180,000 tons of coal,
CCMC has generally maintained the coal pile at between 130,000 to 145,000 tons of
coal, and the pile has never dropped below 101,000 tons. CCMC recognizes that it
is unlikely to use the reserve raw coal in the pile, unless a long-term emergency
affected CCMC’s ability to mine or deliver coal. In the case of such an emergency,
the coal amassed in the coal pile would allow CCMC to meet its contractual delivery
obligations for a period of three weeks.
In 2014, prior to construction of the Coyote Creek Mine, CCMC applied for a
minor source permit with the NDDOH. The permit application described the entire
mining operation, from the coal extraction at the mine face to the processing of the
coal at the plant for transfer to Coyote Station. The permit application identified the
beginning of the coal processing plant as the apron feeder, where raw coal entered into
the processing equipment from the coal pile, making a distinction between the
beginning of the crushing and conveying equipment and the coal pile. The application
specifically stated that the coal pile is not a part of the coal processing plant because
its physical location is before the processing unit and thus the coal pile is not subject
to the Subpart Y regulations. Before issuing a permit, the NDDOH reviewed
CCMC’s application and prepared an Air Quality Effects Analysis (AQEA). The
AQEA reflected that the coal pile is not a part of the coal processing plant and thus
is not subject to the Subpart Y. Because the coal pile is not part of the coal processing
plant, the coal pile’s fugitive emissions are not counted in the calculation of the coal
processing plant’s PTE particulate matter for purposes of determining whether it
requires a major source permit, instead of a minor source permit. Based on the
emissions from the processing equipment and system alone, the NDDOH determined
-6-

-- 6 of 25 --

that the Coyote Creek Mine is a minor source and issued the permit. The NDDOH
issued the permit without providing the public the opportunity for notice and
comment.
Construction of the mining operation began in 2015, and the mine was
operational in 2016. During construction, the Voigts filed suit against CCMC,
alleging violations of the CAA and seeking declaratory and injunctive relief and civil
penalties. The Voigts alleged that construction of the Coyote Creek Mine required a
major source permit, rather than the minor source permit CCMC obtained, and that
CCMC’s coal processing plant violated the CAA because it did not include the
requisite dust control plans for coal processing facilities. If the coal pile is part of the
coal processing plant, as alleged by the Voigts, Subpart Y would apply to the coal pile
and mandate a fugitive dust control plan. Further, a determination that the coal pile
is subject to Subpart Y as part of the coal processing plant would also bring the coal
pile’s fugitive emissions within the PTE air pollutants threshold calculation. Thus,
whether the coal pile is subject to Subpart Y is determinative of both claims.
Both parties moved for summary judgment on the question of whether Subpart
Y applies to CCMC’s coal pile. The district court granted CCMC’s motion and
denied the Voigts’ motion. In a 96-page opinion and order, the district court noted
that both the Voigts and CCMC provided plausible interpretations of Subpart Y that
would render the coal pile a part of or separate from the coal processing plant.
Because the district court concluded that both parties provided plausible
interpretations, it found Subpart Y ambiguous and relied on other sources to resolve
the ambiguity, including EPA guidance and the NDDOH’s permitting decision
regarding the construction of the Coyote Creek Mine. Giving deference to the
NDDOH’s permitting decision, the district court concluded that the coal pile is not
part of the coal processing plant and thus is not subject to Subpart Y. As a result,
CCMC is not required to implement a fugitive dust control plan for the coal pile and
the coal pile’s fugitive emissions are excluded from the PTE air pollutants
-7-

-- 7 of 25 --

determination, which necessitated only a minor source permit for the Coyote Creek
Mine. The Voigts appeal.
II.
The Voigts assert that the district court erred in granting summary judgment to
CCMC and in denying their motion for summary judgment because Subpart Y clearly
and unambiguously includes the coal pile as part of CCMC’s coal processing plant,
and thus CCMC is required to obtain a major source permit and implement a fugitive
dust control plan. Further, the Voigts argue that even if Subpart Y were ambiguous
regarding whether the coal pile is part of the coal processing plant, the district court
erred in relying on the NDDOH permitting decision to resolve the ambiguity in
CCMC’s favor because the NDDOH is a state agency offering an opinion on federal
law that is not entitled to any deference. “We review a district court’s decision on
cross-motions for summary judgment de novo.” Thirty and 141, L.P. v. Lowe’s Home
Ctrs., Inc., 565 F.3d 443, 445-46 (8th Cir. 2009). “Summary judgment is appropriate
if viewing the record in the light most favorable to the nonmoving party, there are no
genuine issues of material fact and the moving party is entitled to judgment as a matter
of law.” Woods v. DaimlerChrysler Corp., 409 F.3d 984, 990 (8th Cir. 2005).
A.
The Voigts first assert that the district court erroneously concluded that the
regulations are ambiguous, arguing that the clear and unambiguous language pulls the
coal pile squarely within the coal processing plant and thus Subpart Y. CCMC
asserts, in response, that the regulations, coupled with unambiguous EPA guidance,
conclusively demonstrate that the coal pile is not part of the coal processing plant. We
reject both arguments and agree with the district court that Subpart Y is ambiguous
regarding whether the coal pile falls within the coal processing plant.
-8-

-- 8 of 25 --

Subpart Y applies to coal processing plants, defined as “any facility (excluding
underground mining operations) which prepares coal by one or more of the following
processes: breaking, crushing, screening, wet or dry cleaning, and thermal drying.”
40 C.F.R. § 60.251(e). But Subpart Y imposes performance standards on only
“affected facilities in coal preparation and processing plants that process more than
. . . (200 tons) of coal per day.” 40 C.F.R. § 60.250(a) (emphasis added). Affected
facilities for the purposes of Subpart Y performance standards is defined to include
“[t]hermal dryers, pneumatic coal-cleaning equipment (air tables), coal processing and
conveying equipment (including breakers and crushers), coal storage systems, transfer
and loading systems, and open storage piles.”2 Id. § 60.250(d) (emphasis added).
Thus, an open storage pile, defined as “any facility, including storage area, that is not
enclosed that is used to store coal, including the equipment used in the loading,
unloading, and conveying operations of the facility,” id. § 60.251(m), is subject to
Subpart Y as an affected facility where it is in the coal processing plant. However, the
regulations do not define what it means for an affected facility to be “in” a coal
processing plant.
The Voigts assert that the definitions of coal processing plant and open storage
pile clearly demonstrate that Subpart Y broadly applies to open storage piles,
regardless of their location before or after the coal crushing equipment. CCMC argues
the EPA guidance unambiguously dictates that the regulations apply only to open
storage piles where the piles occur past the first hopper, which is the component into
which coal is loaded in bulk and is tapered downward in smaller segments toward the
crushing equipment. We disagree with both parties’ contentions. The regulations
2The parties do not dispute the application of Subpart Y to the coal processing
and conveying equipment, defined as “any machinery used to reduce the size of coal
or to separate coal from refuse, and the equipment used to convey coal to or remove
coal and refuse from the machinery. This includes, but is not limited to, breakers,
crushers, screens, and conveyor belts. Equipment located at the mine face is not
considered to be part of the coal preparation and processing plant.” 40 C.F.R.
§ 60.251(f).
-9-

-- 9 of 25 --

simply do not provide an unambiguous answer to the inquiry here: whether a coal pile
that is adjacent to the coal processing equipment, and is used for both storage and
loading coal into the coal processing equipment, is “in” the coal processing plant
itself. Indeed, as the district court noted, both parties provide plausible conflicting
interpretations of the regulations, underscoring the ambiguity that exists. While the
regulations clearly contemplate the inclusion of coal piles that are within coal
processing plants, they do not provide unambiguous direction as to when exactly a
coal pile is “in” a coal processing plant so as to be considered an affected facility
subject to Subpart Y requirements.
Because we conclude the regulations are ambiguous, we turn to subsequent
interpretative guidance to aid us in determining whether the coal pile is part of the coal
processing plant. See Coeur Alaska, Inc. v. Se. Alaska Conservation Council, 557
U.S. 261, 278 (2009); see also Kisor v. Wilkie, 139 S. Ct. 2400 (2019). Kisor
instructs that deference to EPA guidance is appropriate where “(1) the regulation [is]
genuinely ambiguous; (2) the agency’s interpretation of the regulation [is] reasonable;
(3) the interpretation [is] the agency’s authoritative or official position; (4) the
interpretation . . . in some way implicate[s] the agency’s substantive expertise; and (5)
the interpretation . . . reflect[s] fair and considered judgment.” Wells Fargo & Co. v.
United States, 957 F.3d 840, 855 (8th Cir. 2020) (Grasz, J., dissenting in part) (citing
Kisor, 139 S. Ct. at 2415-18).
With respect to the dispositive issues in this case, the EPA has offered some
clarification on when a coal pile is considered to be “in” a coal processing plant:
It should be noted that if the coal is unloaded for the purpose of storage,
then the unloading activity is not an affected facility under NSPS
Subpart Y. The coal must be directly unloaded into receiving
equipment, such as a hopper, to be subject to the provisions of NSPS
Subpart Y.
-10-

-- 10 of 25 --

New Source Performance Standards (NSPS)—Applicability of Standards of
Performance for Coal Preparation Plants to Coal Unloading Operations, 63 Fed. Reg.
53288-01, 53289 (Oct. 5, 1998). And the EPA has reiterated this point, again stating
that “coal must be directly unloaded into receiving equipment” for Subpart Y
applicability. Id. The EPA further stated in its responses to comments on proposed
amendments to Subpart Y that it “interprets coal unloading into the first hopper
‘downstream’ from any form of transportation to be the beginning of the ‘coal
preparation plant.’” Response to Comments Received on Proposed 2009
Amendments, Standards of Performance for Coal Preparation and Processing Plants
(Subpart Y), R. Doc. 38-5, at 89; see also Standards of Performance for Coal
Preparation and Processing Plants, 74 Fed. Reg. 51950-01, 51958 (Oct. 8, 2009) (“A
coal preparation and processing plant begins at the first hopper (i.e. drop point) used
to unload coal . . . .”).
Although this EPA guidance expands upon the gap in the regulations defining
where precisely a coal processing plant begins for the purposes of whether an affected
facility is “in” a coal processing plant, we agree with the district court that it too does
not provide a conclusive answer, particularly where, as here, the coal pile is used for
storage, unloading, and feeding purposes. The record reflects that CCMC’s coal pile
plays a necessary role in the process by which coal is directly unloaded into receiving
equipment, or the apron feeder; however, the record also reflects that the coal pile is
maintained at its size for storage purposes to allow CCMC to fulfill contractual
obligations in the event of a delay or shutdown at the mine face. The coal pile is, in
essence, a hybrid between a storage and unloading pile. There is thus no clear cut
answer as to whether the coal pile is for storage—and unaffected by Subpart Y—or
part of direct unloading into receiving equipment—and subject to Subpart Y. Having
reviewed the relevant regulations and EPA guidance, we conclude that they do not
provide a clear answer as to whether CCMC’s coal pile is “in” the coal processing
plant so as to qualify as an affected facility subject to Subpart Y.
-11-

-- 11 of 25 --

Given the foregoing, the district court did not err in concluding that the
regulations are ambiguous or in ultimately concluding that the regulations, combined
with the guidance, do not resolve the relevant inquiry. We agree with the district court
that the best interpretative aid to determine whether Subpart Y applies to the coal pile
is the NDDOH permitting decision, which concluded that the coal pile is not part of
the coal processing plant and thus is not subject to Subpart Y. And, as discussed
below, that decision is entitled to deference.
B.
The Voigts next argue that the NDDOH permitting decision is not entitled to
any deference because it represents no more than a state agency’s interpretation of
federal law. The Voigts specifically argue that the EPA has expressly reserved the
authority to interpret its own NSPS regulations and the EPA does not and cannot
delegate authority to states to make decisions affecting the uniform applicability and
consistency of NSPS. The Voigts also assert that the issuance of a permit without a
public notice and comment period further demonstrates why deference is unwarranted.
First, the Voigts’ contention regarding the EPA’s authority ignores the system
of cooperative federalism that exists to help achieve the aims of the CAA. “Under the
CAA’s cooperative-federalism scheme, the EPA directs states to submit state
implementation plans to assure reasonable progress toward the CAA’s national
visibility goals.” Nat’l Parks Conservation Ass’n v. EPA, 759 F.3d 969, 971 (8th Cir.
2014) (internal quotation marks omitted).
The Clean Air Act requires states to develop state statutory and
regulatory programs that implement the air quality planning objectives
of the Clean Air Act. These state programs are incorporated into a State
Implementation Plan (“SIP”), which the EPA reviews and approves.
Once the EPA approves a SIP, the state acquires “SIP-approved” status
for the EPA-approved air quality programs. Thereafter, the state has
-12-

-- 12 of 25 --

primary responsibility for implementing federal air quality planning
goals.
United States v. Minnkota Power Co-op Inc., 831 F. Supp. 2d 1109, 1113 (D.N.D.
2011). And the EPA has expressly delegated authority to the State of North Dakota
to implement NSPS rules. See Automatic Delegation of Authority to the States of
Colorado, Montana, North Dakota, South Dakota, Utah, and Wyoming to Implement
and Enforce New Source Performance Standards (Automatic Delegation), 79 Fed.
Reg. 60993-01, 60994 (Oct. 9, 2014) (“The CAA section 111(c) authorizes the EPA
to delegate authority to any state agency which submits adequate regulatory
procedures for implementation and enforcement of the NSPS program.”).
The EPA-delegated primary responsibility for carrying out the CAA air quality
goals takes the form of issuance of preconstruction permits: “states issue the
preconstruction permits in accordance with their SIPs and federal minimum
standards.” Nucor Steel-Ark. v. Big River Steel, LLC, 825 F.3d 444, 447 (8th Cir.
2016). Because implementation of the CAA hinges on a system of cooperative
federalism and North Dakota has an EPA-approved SIP, North Dakota is the primary
party enforcing the CAA for the State.
Second, the district court’s exhaustive and well-reasoned opinion demonstrates
that the NDDOH permitting decision does not run afoul of the relevant regulations
and EPA guidance. The district court explicitly noted that the NDDOH decision was
entitled to deference because (1) the determination that the coal pile is not part of the
coal processing plant is consistent with EPA guidance describing the beginning of a
coal processing plant as the “first hopper”; (2) the coal pile contains only unprocessed
raw coal; (3) most of the coal pile consists of coal stored long-term in the event of a
delay or shutdown at the mine face; and (4) the exclusion of the coal pile from the coal
processing plant does not eviscerate the regulations as they would still apply to coal
piles in a coal processing plant, particularly those that contained processed coal. As
the primary body responsible for issuing permits based upon the CAA standards,
North Dakota is in the best position to decide whether a given facility falls within or
-13-

-- 13 of 25 --

satisfies the CAA standards, and that decision is entitled to deference. We agree with
the district court that the above factors demonstrate that the NDDOH’s conclusion that
the coal pile is not part of the coal processing plant is not an arbitrary or unreasonable
position inconsistent with the EPA and CAA’s aims. We decline to second-guess the
NDDOH’s exercise of its authority. See Minnkota, 831 F. Supp. 2d at 1121 (“North
Dakota’s conclusions regarding such highly technical matters are entitled to deference
unless the EPA proves them to be unreasonable, arbitrary, or capricious.”).
Regarding the Voigts’ assertion that giving deference to the NDDOH permitting
decision undercuts the EPA’s non-delegable authority to make legal determinations
in order to preserve the uniformity and consistency of NSPS on a national level, the
Voigts again ignore the cooperative framework where states are tasked with carrying
out the CAA’s aims, which include making determinations regarding NSPS
applicability. Although factual determinations are often necessarily intertwined with
legal issues, there is no suggestion that the NDDOH exceeded its authority, delegated
by the EPA, in carrying out its SIP to make an NSPS applicability determination and
in issuing the preconstruction minor source permit to CCMC. See Automatic
Delegation, 79 Fed. Reg. at 60994 (stating that delegated authority under SIP includes
“implementation and enforcement of NSPS program”). The process for NSPS
enforcement would be significantly impaired if the state authority did not have the
ability to make determinations based on application of given facts to the SIP and EPA
framework. Further, state permitting decisions are not immune from review;
“unreasonable, arbitrary, or capricious” exercises of the NDDOH’s authority that
could result in national inconsistency may be challenged. See Minnkota, 831 F. Supp.
2d at 1121. The existence of this avenue to challenge aberrant decisions guards
against the risk of national inconsistency. We are thus unpersuaded by the Voigts’
contention that affording the NDDOH permitting decision deference will result in
national inconsistency.
We finally note that our conclusions are not affected by the Voigts’ purported
exclusion from the permitting process due to a lack of a public notice and comment
-14-

-- 14 of 25 --

period. As the district court noted, the lack of a public notice and comment period
resulted in this case coming to us “in a much different posture” than it would have
“had defendant requested the NDDOH follow a more formal process in the handling
of its application for the construction permit[.]” R. Doc. 127, at 7. The district court
acknowledged that this process would have allowed the Voigts “the opportunity to
make the arguments they are now making to the NDDOH; the NDDOH could have
addressed them; and then, if either plaintiffs or defendant were disappointed in its
decisions, there would have been the right of an appeal . . . that . . . could have been
resolved based upon a formal administrative record.” R. Doc. 127, at 7-8. Although
that may have been the preferred course, the absence of public notice and comment
does not require us to unwind the NDDOH permitting decision on that basis alone,
particularly where the Voigts have had judicial recourse to raise the arguments they
would have raised during a notice and comment period.
On the record before us and given the overarching framework of the CAA,
including the cooperative relationship between the EPA and the states, we conclude
the district court appropriately gave deference to the NDDOH permitting decision to
resolve the regulations’ ambiguity in favor of CCMC. The district court thus did not
err in granting summary judgment to CCMC on the basis that the coal pile is not
subject to Subpart Y regulations, which would have required a major source permit
and a fugitive dust control plan.
III.
For the foregoing reasons, we affirm the judgment of the district court.
STRAS, Circuit Judge, dissenting.
Most Americans would be surprised to learn that state bureaucrats can play an
even larger role than federal judges do in interpreting federal law. Yet by deferring
to the North Dakota Department of Health’s interpretation of a Clean Air Act
-15-

-- 15 of 25 --

regulation, the court’s decision has just that effect. In my view, even if we must defer
to a federal agency’s interpretation of a federal statute, see Chevron U.S.A. Inc. v.
Natural Res. Def. Council, Inc., 467 U.S. 837 (1984), and a federal agency’s
interpretation of a federal regulation, see Auer v. Robbins, 519 U.S. 452 (1997), it
defies basic constitutional principles to defer to a state agency’s interpretation of
federal law. For that reason, I respectfully dissent from today’s unveiling of Voigt
deference.
I.
Our interpretive task in this case is technical. We must determine whether
Coyote Creek’s coal pile is an “affected facilit[y] in [a] coal preparation and
processing plant[].” 40 C.F.R. § 60.250(a) (emphasis added). If it is, then the
Environmental Protection Agency’s (“EPA”) Standards of Performance for Coal
Preparation and Processing Plants (“Subpart Y”) apply. Relying on an interpretation
by the North Dakota Department of Health, the court holds that it is not.
At first glance, the court’s holding may seem mundane, given that this case is
about whether a particular pile of coal happens to be “in” a coal plant. 40 C.F.R.
§ 60.250(a). But there is a more fundamental principle at stake about who gets to say
what federal law means. Although the court allows a state agency to do it, I believe
the Constitution places the responsibility with us.
II.
One might be under the mistaken impression that looking to states for assistance
is unique to the Clean Air Act. It is not. In a sprawling administrative state, the
federal government often asks states to implement federal policy. There are numerous
examples of cooperative federalism at work: the Medicaid Act, the
Telecommunications Act of 1996, and the Clean Water Act, to name just a few.
-16-

-- 16 of 25 --

In the abstract, cooperative federalism makes sense. It allows states to account
for their own unique needs in implementing federal law. But when implementation
spills over into interpretation, the calculus changes. At that point, state agencies
should not have a special role—much less a decisive one—in telling us what a federal
statute or regulation means.
For its part, the court cites no controlling authority that requires us to defer to
a state agency. No Supreme Court or Eighth Circuit case even suggests that it is
appropriate, and for good reason. See, e.g., Kisor v. Wilkie, 139 S. Ct. 2400, 2414–18
(2019) (discussing the requirements for courts to defer to a federal agency’s
interpretation of its own ambiguous regulations); United States v. Mead Corp., 533
U.S. 218, 227–31 (2001) (explaining when a federal agency’s interpretation of a
federal statute deserves deference). It runs afoul of both horizonal and vertical
separation-of-powers principles.
A.
Let’s start with the horizontal component—what typically comes to mind when
hearing the words, “separation of powers.” The Constitution divides the powers of
the federal government into three distinct branches: the executive, the legislative, and
the judicial. See INS v. Chadha, 462 U.S. 919, 951 (1983). As Joseph Story
straightforwardly put it, “the powers of one [branch] ought not to be exercised by
either of the others.” 2 Joseph Story, Commentaries on the Constitution of the United
States § 1416, at 260 (Boston, Charles C. Little & James Brown eds., 1851).
The judicial branch, which includes “one supreme Court, and [] such inferior
Courts as the Congress may from time to time ordain and establish,” alone exercises
“[t]he judicial Power of the United States.” U.S. Const. art. III, § 1. This power, as
Alexander Hamilton explained in Federalist 78, includes “[t]he interpretation of the
laws,” which he called “the proper and peculiar province of the courts.” The
Federalist No. 78, at 467 (Alexander Hamilton) (Clinton Rossiter ed., 1961). The duty
-17-

-- 17 of 25 --

was a familiar one, for judges had already been doing it for centuries. See Perez v.
Mortgage Bankers Ass’n, 575 U.S. 92, 119–20 (2015) (Thomas, J., concurring in the
judgment) (collecting historical sources).
The Framers had reason to place the interpretive power in the judicial branch.
See Thomas Sergeant, Constitutional Law: Being a View of the Practice and
Jurisdiction of the Courts of the United States, and of Constitutional Points Decided
18 (Philadelphia, P.H. Nicklin & T. Johnson eds., 1830). Experiences in England and
under the Articles of Confederation had shown a tendency for the other branches of
government to borrow the judicial power piece by piece. See Perez, 575 U.S. at
124–25 (Thomas, J., concurring in the judgment) (discussing King James I’s pressure
on judges to approve his attempts to raise revenue without Parliament’s participation);
see also Prohibitions Del Roy (1607), 77 Eng. Rep. 1342, 12 Co. Rep. 64 (KB);
Gordon S. Wood, The Creation of the American Republic: 1776–1787, at 408 (1998)
(describing “[t]he assumption of the judicial [power] . . . into the hands of the [state]
legislative branch[es]” (quotation marks omitted)).
The natural tendency of the other branches to try to accumulate power is why
the Framers chose to create three distinct branches of government, with “fences”
between them. Plaut v. Spendthrift Farm, Inc., 514 U.S. 211, 240 (1995). As James
Madison warned, “[t]he accumulation of all powers, legislative, executive, and
judiciary, in the same hands . . . may justly be pronounced the very definition of
tyranny.” The Federalist No. 47, supra, at 301 (James Madison). Alexander
Hamilton foresaw that this danger could extend to the courts, writing that “liberty can
have nothing to fear from the judiciary alone, but would have everything to fear from
its union with either of the other departments.” The Federalist No. 78, supra, at 466
(Alexander Hamilton).
The threat to the judiciary’s interpretive power is once again right out in the
open. Over the last century, the growth of the administrative state has chipped away
at it, while judicial-deference doctrines have acquiesced in—perhaps even
-18-

-- 18 of 25 --

encouraged—this development, in a marked departure from both historical practice
and the Framers’ constitutional design. See, e.g., Decatur v. Paulding, 39 U.S. 497,
515 (1840) (stating that courts were not obligated to give weight to an executive
official’s interpretation of law if the court thought differently); see also Perez, 575
U.S. at 114 (Thomas, J., concurring in the judgment) (noting that the “doctrine of
deference . . . has taken on a life of its own”).
This case is a prime example. Executive-branch officials in North Dakota
determined that Coyote Creek’s coal pile is not “in” its plant. 40 C.F.R. § 60.250(a).
Rather than using “recognized tools of interpretation” and exercising independent
judgment, Perez, 575 U.S. at 123 (Thomas, J., concurring in the judgment), the court
starts with a thumb on the scale in favor of North Dakota’s interpretation. The
question becomes what the executive branch believes the regulation means, not what
we, as judges, think. This “transfer of interpretive judgment raises serious separation-
of-powers concerns,” id. at 124, because it involves the exchange of an “independent
decisionmaker” for an “avowedly politicized administrative agent seeking to pursue
whatever policy whim may rule the day,” Gutierrez-Brizuela v. Lynch, 834 F.3d 1142,
1153 (10th Cir. 2016) (Gorsuch, J., concurring).
B.
The case for deference becomes even weaker when the interpreter is a state or
local government official. The reason is the vertical separation of powers, more
commonly known as federalism. One of “[t]he great innovation[s]” of the
Constitution, Printz v. United States, 521 U.S. 898, 920 (1997), was the creation of
a “system of dual sovereignty,” Gregory v. Ashcroft, 501 U.S. 452, 457 (1991)—the
idea of “an indestructible Union, composed of indestructible States,” Texas v. White,
74 U.S. 700, 725 (1869).
The “characteristic difference” between the Constitution and the Articles of
Confederation was the ability of the federal government to regulate citizens directly,
-19-

-- 19 of 25 --

Cohens v. Virginia, 19 U.S. 264, 309 (1821), without “the intervention of the State
legislatures,” The Federalist No. 16, supra, at 117 (Alexander Hamilton); see The
Federalist No. 15, supra, at 108 (Alexander Hamilton). Any legislation passed by the
new federal government became binding, even on the states, under the Supremacy
Clause:
This Constitution, and the Laws of the United States which shall be
made in Pursuance thereof; and all Treaties made, or which shall be
made, under the Authority of the United States, shall be the supreme
Law of the Land; and the Judges in every State shall be bound thereby,
any Thing in the Constitution or Laws of any State to the Contrary
notwithstanding.
U.S. Const. art. VI, cl. 2; see Cohens, 19 U.S. at 382 (describing “the subordination
of the State governments to th[e] Constitution”).
Judges would play a role in ensuring federal supremacy, at least when the
federal government operated within its enumerated powers. See The Federalist No.
82, supra, at 491 (Alexander Hamilton) (stating that dual sovereignty would
“originate questions of intricacy and nicety”). As the Supreme Court soon made clear,
judges are responsible for saying “what the law is,” Marbury v. Madison, 5 U.S. 137,
177 (1803), which sometimes entails checking local interests, see Martin v. Hunter’s
Lessee, 14 U.S. 304, 347 (1816).
The “structural protections” afforded to judges, like life tenure and non-
diminishment of salary, Perez, 575 U.S. at 121–22 (Thomas, J., concurring in the
judgment), would, when necessary, guard against the dominance of those interests,
Martin, 14 U.S. at 347; see also Cooper v. Aaron, 358 U.S. 1, 23–24 (1958)
(Frankfurter, J., concurring); Story, supra, § 1599, at 388 (explaining the
constitutional choice to insulate judicial appointments from popular elections).
Federal judges, not states or officials within those states, would have the final say over
what federal law means, see Martin, 14 U.S. at 355, even if the question was
-20-

-- 20 of 25 --

“difficult[],” “doubtful,” or “the strongest State feelings were engaged,” Cohens, 19
U.S. at 404, 421.
C.
Under these basic structural principles, it should be clear that the Constitution
does not leave room for state executive-branch officials to tell federal judges what
federal law means. Rather than bending to “local habits [or] feelings,” Cooper, 358
U.S. at 25 (Frankfurter, J., concurring), our job is to exercise independent judgment
“to say what the law is,” Marbury, 5 U.S. at 177, even when—perhaps especially
when—someone else interprets the law differently, see The Federalist No. 78, supra,
at 467 (Alexander Hamilton).
By extending deference to state executive-branch officials, the court has taken
what so far has been only a horizontal separation-of-powers problem and expanded
it vertically. A state bureaucrat, who unquestionably has “state attachments . . . and
[] interests,” Martin, 14 U.S. at 347, now has more say over Clean Air Act regulations
than we do. No one would seriously argue, to use a slightly different example, that
we should defer to what state judges say federal law means, see Cohens, 19 U.S. at
420, even though they are “bound” to interpret and apply it just as we are, U.S. Const.
art. VI, cl. 2. But see 28 U.S.C. § 2254(d) (requiring deference to state-court
judgments in federal-habeas-corpus proceedings). Yet in the court’s view, somehow
state executive-branch officials are different.
It also apparently does not matter that this case involves a double delegation:
one from Congress to a federal agency and another from a federal to a state agency.
Even if we have to accept the validity of the first delegation, see Chevron, 467 U.S.
at 843–44, state agencies “are neither regional offices nor administrative agencies of
the Federal Government,” New York v. United States, 505 U.S. 144, 188 (1992). Nor,
unlike state judges, are they entrusted by the Constitution with the duty to interpret
federal law. U.S. Const. art. VI, cl. 2.
-21-

-- 21 of 25 --

In the end, the Supremacy Clause is one of the main losers. See Mayo v. United
States, 319 U.S. 441, 445 (1943) (describing how the Supremacy Clause furthers
“uniformity”). North Dakota, one of many states that has the power to implement the
Clean Air Act, see Automatic Delegation of Authority to the States of Colorado,
Montana, North Dakota, South Dakota, Utah, and Wyoming to Implement and
Enforce New Source Performance Standards, 79 Fed. Reg. 60993, 60993 (Oct. 9,
2014), has already decided that Coyote Creek’s coal pile is not “in” the plant.
Suppose that its neighbor, South Dakota, goes the other way when confronted with a
similar coal pile and that the Missouri Department of Natural Resources adopts yet a
third interpretation. See Standards of Performance for New Stationary Sources
(NSPS) and National Emission Standards for Hazardous Air Pollutants (NESHAP);
Delegation of Authority to the States of Iowa, Kansas, Missouri, Nebraska, Lincoln-
Lancaster County, Nebraska, and the City of Omaha, Nebraska, 63 Fed. Reg. 1746,
1746 (Jan. 12, 1998).
In this scenario, a federal regulation would be subject to three different
interpretations—all within a single circuit—despite the stated preference in the Clean
Air Act for federal coordination over air pollution. See 42 U.S.C. § 7401(a)(4)
(“Federal . . . leadership is essential for the development of cooperative Federal, State,
regional, and local programs to prevent and control air pollution.”). No longer would
there be “uniformity” of federal law in this circuit, much less “throughout the whole
United States.” Martin, 14 U.S. at 348. And in a complete turnabout from Marbury,
no one would have the final say over “what the law [actually] is.” 5 U.S. at 177; see
also Martin, 14 U.S. at 348 (discussing the dangers of having “no revising authority
to control [] jarring and discordant judgments[] and harmonize them”).
III.
Even aside from the constitutional problems, there are other good reasons not
to defer here. To start, the Supreme Court has never, to my knowledge, deferred to
an agency’s interpretation of a regulation that it did not write itself. Kisor identifies
-22-

-- 22 of 25 --

the reason: “the agency that promulgated a rule is in [a] better position to reconstruct
its original meaning.” 139 S. Ct. at 2412 (plurality opinion) (internal quotation marks
and brackets omitted). However persuasive this justification may otherwise be,3 it
does not apply to the North Dakota Department of Health, which played no part in
writing Subpart Y.
Nor does the agency’s interpretation represent its “fair and considered
judgment.” Id. at 2417 (majority opinion) (quotation marks omitted). Even if we
were to assume that the North Dakota Department of Health has technical expertise
under the Clean Air Act, its decision was made by an engineer who took Coyote
Creek at its word that the coal pile was not covered by Subpart Y. Then the agency,
apparently without studying the matter further, simply stood by the engineer’s
determination. This sequence of events raises the possibility that its interpretation
may just be a “post hoc rationalization.” Id. (italics, quotation marks, and brackets
omitted). At the very least, it fosters doubt about whether it is the “official position”
of the agency. Id. at 2416 (quotation marks omitted).
IV.
Deference makes a difference here. Deference leads to one conclusion: Coyote
Creek’s coal pile, an “affected facilit[y],” is not “in” the “coal preparation and
processing plant[].” 40 C.F.R. § 60.250(a). The plain language leads to another: the
pile is in the plant.
3Another justification for deference is that a single agency can resolve
interpretive issues by uniform administrative decision rather than
“piecemeal . . . litigation.” Kisor, 139 S. Ct. at 2413 (plurality opinion). This
rationale goes out the window, however, when the administrative decision gets spread
among multiple separate state agencies. Indeed, even when only federal agencies are
involved, deference is not automatic when decisionmaking authority is shared. See
Martin v. Occupational Safety & Health Rev. Comm’n, 499 U.S. 144, 151–57 (1991).
-23-

-- 23 of 25 --

The word “in,” used as a preposition to describe location, has an unambiguous
meaning: “[w]ithin the limits, bounds, or area of.” The American Heritage Dictionary
of the English Language 885 (5th ed. 2016). Applying this definition and leaving the
technical jargon aside, the regulation requires us to decide whether the “affected
facilit[y]”—here, the coal pile—is physically “[w]ithin the limits, bounds, or area of”
the Coyote Creek plant. 40 C.F.R. § 60.250(a); American Heritage Dictionary, supra,
at 885. Everyone, including the EPA, agrees that the apron feeder, which receives the
coal for conveyance into the crushing equipment, is “in” the plant. See Response to
Comments Received on Proposed 2009 Amendments, Standards of Performance for
Coal Preparation and Processing Plants (Subpart Y), R. Doc. 38-5, at 83–84
(hereinafter, “EPA Guidance”); Appellant Br. at 14, 25; Appellee Br. at 40. So it
follows that the coal pile, which fully envelops and surrounds the apron feeder, is in
the plant, too.
No other conclusion is possible in light of the remainder of the regulation.
First, included among “affected facilities” are “open storage piles.” 40 C.F.R.
§ 60.250(d). What this means, again leaving the technical jargon aside, is that piles
of coal are capable of being “in” a coal plant. Id. § 60.250(a). Second, as the words
“open” and “storage” suggest, an “open storage pile[]” is “not enclosed [and] is used
to store coal.” Id. § 60.251(m) (emphasis added). Combined with the first point, it
means that a non-enclosed coal pile can still be “in” the plant. If Coyote Creek’s coal
pile, which fully envelops the apron feeder, is not one of those piles, it is hard to
imagine what “open storage pile[]” would be. Id. § 60.250(d).
Seeking to defend the agency’s decision, Coyote Creek can barely muster a
textual argument to the contrary. See Appellee Br. 35–36. Its position, based loosely
on an EPA guidance, is that the plant can only begin at the apron feeder, so its pile of
unprocessed coal cannot be in the plant. See EPA Guidance 83–84 (stating that the
beginning of the coal plant is at “the first hopper”). The regulation, however, neither
draws the line where Coyote Creek says nor distinguishes between processed and
unprocessed coal. Instead, by including “coal processing and conveying equipment”
-24-

-- 24 of 25 --

as “affected facilities,” it anticipates that some of the coal “in” the plant will be
unprocessed, at least before it reaches the “coal[-]processing” equipment. 40 C.F.R.
§ 60.250(a), (d). Coyote Creek’s interpretation, in other words, is unreasonable.
* * *
To sum up, I would not defer to the North Dakota Department of Health’s
interpretation of a federal regulation. Rather, using “recognized tools of
interpretation,” Perez, 575 U.S. at 123 (Thomas, J., concurring in the judgment), I
would conclude that Coyote Creek’s coal pile is “in” the plant. 40 C.F.R. § 60.250(a).
______________________________
-25-

-- 25 of 25 --