OPINION

SUHRHEINRICH, Circuit Judge.

This case arises from a final action of the Environmental Protection Agency (EPA) determining that a natural gas sweetening plant and various sour gas production wells commonly owned by Petitioner Summit Petroleum Corporation (Summit) and separately located within an area of approximately forty-three square miles constitute a single stationary source under the EPA's Clean Air Act Title V permitting program. Specifically at issue is the EPA's conclusion that Summit's facilities satisfy the regulatory requirement of being "located on ... adjacent properties" because, although physically independent, they are "truly interrelated." Summit, together with Amici American Petroleum Institute and American Exploration and Production Counsel, argues that the EPA's determination that the physical requirement of adjacency can be established through mere functional relatedness is unreasonable and contrary to the plain meaning of the term "adjacent." We agree. For these and other reasons fully set forth herein, we VACATE the EPA's final determination and REMAND this case to the EPA to determine whether Summit's sweetening plant and sour gas wells are sufficiently physically proximate to be considered "adjacent" within the ordinary, i.e., physical and geographical, meaning of that requirement.

I. Background

A. Summit's Natural Gas Production Facilities

Summit is a producer of natural gas that owns and operates a natural gas sweetening plant in Rosebush, Michigan. Summit's plant "sweetens" the "sour" gas from approximately one hundred sour gas production wells by removing hydrogen sulfide so that the gas can be used. Summit owns all of the production wells and the subsurface pipelines that connect each of the wells to the sweetening plant. The wells themselves are located over an area of approximately forty-three square miles at varying distances from the plant — from five hundred feet to eight miles away — and Summit does not own the property between the individual well sites or the property between the wells and the plant. None of the well sites share a common boundary with each other, nor do any of the well sites share a common boundary with Summit's production plant. Flares work as part of the plant operations by burning off natural gas waste to relieve pressure on the gas collection equipment. The closest flare is located approximately one half-mile from the plant, while the remaining flares are each over one mile away.

The sweetening plant and the majority of Summit's gas production wells and flares are located within the territory of Michigan's Saginaw Chippewa Indian Tribe's Isabella Reservation. The sweetening plant, gas production wells, and flares emit sulfur dioxides and nitrous oxides, air pollutants subject to regulation under the Clean Air Act (CAA). The plant alone emits, or has the potential to emit, just under one hundred tons of these pollutants per year. Each flare and each well site emits, or has the potential to emit, much lower amounts of pollutants. However, if the emissions of sulfur dioxide from the plant and any one production well were to be combined, they would exceed one hundred tons of pollutants per year.

B. The EPA's Title V Permitting Program

In 1970, Congress passed the CAA, 42 U.S.C. §§ 7401-7671q, "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...." 42 U.S.C. § 7401(b)(1) (2006). In 1990, Congress enacted Title V of the CAA (Title V), 42 U.S.C. §§ 7661-7661f, establishing an operating permit program to regulate stationary sources of air pollution. Rather than imposing new substantive air quality control requirements, Title V supplements the CAA by requiring regulated sources of air pollution to obtain operating permits that include emission limitations, standards, monitoring requirements, compliance schedules, and other conditions necessary to assure compliance with the CAA. See 42 U.S.C. §§ 7661a(a), 7661c(a); Ohio Pub. Interest Research Grp., Inc. v. Whitman, 386 F.3d 792, 794 (6th Cir.2004). Although "[s]tates and local governments bear `primary responsibility'" in administering Title V regulations, Whitman, 386 F.3d at 794 (citing 42 U.S.C. § 7401(a)(3)), certain circumstances permit or require the EPA itself to establish and promulgate Title V regulations in lieu of the EPA-approved state plans. In the absence of an EPA-approved plan regulating Indian territory located within a given state, for example, the EPA is authorized to implement a federal air quality control plan within that jurisdiction. Id. § 7601(d)(4). In this case, the EPA has asserted jurisdiction to regulate Indian territory within the State of Michigan, including the Saginaw Chippewa Indian Tribe's Isabella Reservation. See, e.g., Approval and Promulgation of Air Quality Implementation Plans; Michigan; PSD Regulations (PSD Regulations), 73 Fed. Reg. 53,366 (Sept. 16, 2008) (to be codified at 40 C.F.R. pt. 52) ("Michigan is not authorized to carry out its Federally approved air program in `Indian Country,' as defined in 18 U.S.C. 1151.").

Title V requires every "major source" of air pollution to obtain a Title V operating permit. 42 U.S.C. § 7661a(a). The EPA's Title V plan defines a major source to include "any stationary facility or source of air pollutants which directly emits, or has the potential to emit, one hundred tons per year of any pollutant," including nitrous oxides and sulfur dioxides (the pollutants emitted by Summit's operations). See 42 U.S.C. § 7602(j); see also 40 C.F.R. § 71.2. The EPA defines a "stationary source" as "any building, structure, facility, or installation which emits or may emit a regulated [air] pollutant." 40 C.F.R. § 52.21(b)(5). Multiple pollutant-emitting activities, such as Summit's sweetening plant and sour gas production wells, can be aggregated together and considered a "building, structure, facility, or installation," i.e., a single stationary source, under Title V only if they: (1) are under common control; (2) "are located on one or more contiguous or adjacent properties"; and (3) belong to the same major industrial grouping. 40 C.F.R. § 71.2.1 If the pollutant-emitting activities fail to satisfy any one of these three criteria, they are considered separate stationary sources and their emissions cannot be aggregated to meet the major source threshold for which a Title V permit is required.

C. Major Source Determination Correspondence Between Summit and the EPA

In January 2005, Summit, together with the Michigan Department of Environmental Quality (MDEQ), submitted a request to the EPA to determine whether Summit's facilities met the definition of a Title V major source of air pollution.2 Summit's request correctly noted that emissions from its sweetening plant were insufficient to constitute it a major source, but "if potential emissions from the sour gas production wells are included with the gas sweetening plant emissions, the aggregated sources combined may be considered a single major source for criteria pollutants, namely nitrogen oxides and sulfur dioxides." Summit's request suggested that aggregation of its facilities would be improper for two principal reasons. First, Summit's wells were "located at great distances from its production facility on entirely different tracts, leases and surface sites," making them neither contiguous nor adjacent to one another. Second, the EPA had not distinguished between the definition of a "facility" found in its hazardous air pollutant (HAP) regulations from the definition of a "facility" for criteria pollutant (Title V) purposes. The EPA's HAP regulations specifically considered aggregation of multiple facilities in the oil and gas industry and concluded that aggregation was improper: "Emissions from any oil or gas exploration or production well, with its associated equipment, and emissions from any pipeline compressor or pump station shall not be aggregated with emissions from other similar units, whether or not these units are in a contiguous area under common control...." See Article II, Part 55, Air Pollution Control, of the Michigan Natural Resources and Environmental Protection Act, 1994 PA 451, as amended (Act 451), and 1980 AACS R 336.1211 (Rule 211) (emphasis added).

The EPA responded to Summit's source determination request on April 26, 2007, confirming that it was asserting jurisdiction to regulate Summit's facilities within the Isabella Reservation but stating that it was "unable to conclude if the wells and plant constituted a single source for Title V purposes." The EPA distinguished its Title V and HAP regulations, maintaining that pollutant-emitting facilities could be aggregated to constitute a single stationary source under Title V if they operated under common control, were located on contiguous or adjacent properties, and belonged to the same industrial grouping. It then identified the proper procedure for determining a single stationary source in the oil and gas industry, outlined in an EPA guidance memorandum issued by Acting Assistant Administrator William Wehrum:

Memorandum from William L. Wehrum, Acting Assistant Adm'r, to Reg'l Adm'rs I-X (Jan. 12, 2007) available at http://www. eenews.net/public/25/12769/features/ documents/2009/10/13/document_pm_02. pdf (last visited May 16, 2012) (Wehrum Memorandum).

The EPA found that based on the information provided by Summit's request, its wells and production plant were undisputedly commonly owned, part of the same industrial grouping, and not located on contiguous surface sites. Accordingly, the agency concluded that it "must evaluate the adjacency, (i.e., the nearness or closeness) of the sour gas wells to the sweetening plant" to determine whether Summit's facilities constituted a single stationary source subject to regulation. Although it noted that "because the production wells supply the gas to the gas sweetening plant, we believe that the sites meet the common sense notion of a plant," the EPA found that the Wehrum Memorandum "makes clear that proximity is the most informative factor in making a source determination involving oil and gas activities." Finally, the EPA requested additional information — a map of Summit's facilities and a schematic of its processes that showed the connection between Summit's "exploration and production activities" — to assist it in reaching a decision.

Summit provided the requested information in a letter dated April 24, 2008. In the letter, Summit agreed that under the Wehrum Memorandum, proximity was the most crucial factor in making source determinations for the oil and gas industry. Summit reminded the EPA of the Wehrum Memorandum's additional comments regarding the proximity factor, e.g., the EPA's position that it was unreasonable to aggregate wells and production field activities that "occur over large geographic distances with the downstream processing plant into a single stationary source." See id. Summit submitted that the concepts of contiguousness and adjacency should assess "whether the land associated with the pollutant-emitting activity is connected to, or is nearby, land associated with another pollutant-emitting activity." In support, Summit pointed again to the Wehrum Memorandum, which advised that "[a] reviewing authority can consider two surface sites to be in close proximity if they are physically adjacent, or if they are separated by no more than a short distance (e.g., across a highway, separated by a city block or some similar distance)." Id. Summit noted that its wells and production plant were separated by much larger distances than those contemplated by the Wehrum Memorandum and reiterated its request that the EPA not consider them a single stationary source under Title V.

In July 2009, the EPA responded that even with Summit's supplemental materials, it was again unable to conclusively determine whether Summit's facilities constituted a single stationary source. The EPA addressed Summit's emphasis on the Wehrum Memorandum and stressed that physical proximity was not the only factor in adjacency determinations, stating that "generally, the EPA determines whether two facilities are `adjacent' based on a `common sense' notion of a source and the functional interrelationship of the facilities, rather than simply on the physical distance between the facilities." The EPA concluded by requesting that Summit contact it as soon as possible to further "discuss the requirements of the ... Title V permit applications for sources located on tribal lands."

The parties engaged in a conference call in July 2009. According to Summit, the EPA indicated that it might no longer be using the Wehrum Memorandum as guidance in Summit's source determination. The EPA said it would investigate the status of the Wehrum Memorandum and gather information on permit conditions for other sources before making a final source determination.

1. The EPA determines Summit's facilities to be a single and major source under Title V.

On September 8, 2009, EPA issued its determination that Summit's sweetening plant and sour gas wells constituted a single stationary source, and were therefore a major source, under Title V. The EPA did not explain the basis for its decision, save for citing to its prior April 26, 2007 and April 16, 2009 letters to Summit. Shortly after the EPA issued its decision, it forwarded Summit a September 22, 2009 guidance memorandum prepared by the EPA's Office of Air and Radiation Assistant Administrator Gina McCarthy entitled "Withdrawal for Source Determinations for Oil and Gas Industries" (McCarthy Memorandum). See Memorandum from Gina McCarthy, Assistant Adm'r, to Reg'l Admr's Regions I-X (Sept. 22, 2009), available at http://www.eenews.net/public/ 25/12769/features/documents/2009/10/13/ document_pm_01.pdf (last visited May 16, 2012). In pertinent part, the McCarthy Memorandum reasoned that the simplified approach of the Wehrum Memorandum, with its emphasis on physical proximity in the determination of what constituted a stationary source, was no longer appropriate. The McCarthy Memorandum called permitting authorities to instead "rely foremost on the three regulatory criteria for aggregating emissions from multiple sources, ... (1) whether the activities are under the control of the same person (or person under common control); (2) whether the activities are located on one or more contiguous or adjacent properties; and (3) whether the activities belong to the same industrial grouping." Id. (citing 40 C.F.R. § 52.21(b)(6)). The guidance memorandum nevertheless conceded that stationary source determinations should be made on a case-by-case basis and that "in some cases, `proximity' may serve as the overwhelming factor in a permitting authority's source determination decision." Id.

In response to the EPA's September 8, 2009 determination, Summit filed its initial Petition for Review with this Court on November 4, 2009. On January 5, 2012, Summit sent a letter to the EPA requesting an administrative stay of the deadline the EPA had imposed on Summit to obtain a Title V operating permit so that Summit could submit additional information for the EPA's consideration. The EPA granted Summit's request and the parties submitted a joint motion for abeyance to this Court on March 4, 2010. This Court granted the motion on March 18, 2010, holding the case in abeyance and vacating the briefing schedule to allow the parties time to resolve their pending administrative issues.

On October 18, 2010, the EPA sent Summit a final letter confirming that even with Summit's supplemental information, the EPA considered Summit's facilities a single stationary source, and thus a major source, under Title V. The EPA explained that the McCarthy Memorandum directed it to consider all three regulatory factors. On the issue of the facilities' adjacency, the EPA indicated that it had never established a specific distance between pollutant-emitting entities at which they were no longer adjacent. Id. The EPA reiterated that, historically, factors such as the "nature of the relationship between the facilities" and the "degree of interdependence between them" had been important to the question of whether two facilities were adjacent. The agency discussed its view that Summit's plant, wells, and flares worked together as a single unit that "together produced a single product" and found that Summit had failed to provide any evidence to demonstrate that the emissions sources were not "truly interdependent." The EPA thus concluded that given this functional interrelationship, Summit's facilities "should not be considered separate emission sources."

Summit filed a second Petition for Review in this Court on October 18, 2010. The EPA opposed, and this Court denied, a motion to intervene filed by American Petroleum Institute (API). This Court granted the unopposed motions of API and American Exploration and Production Council (AXPC) (collectively Amici) to submit amicus briefs in the case.

II. Analysis

A. Standard of Review

Section 307(b) of the CAA, 42 U.S.C. § 7607(b), does not specify a standard for judicial review of an EPA final action. Accordingly, this Court reviews the EPA's source determination under the judicial review provisions of the Administrative Procedure Act (APA). 5 U.S.C. § 706(2)(A). See Alaska Dep't of Envtl. Conservation v. E.P.A., 540 U.S. 461, 496-97, 124 S.Ct. 983, 157 L.Ed.2d 967 (2004). The APA requires us to credit an agency action unless it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A). Where, as here, we review an agency's interpretation of its own regulation and not its application of a statute passed by Congress, we must defer to the agency's interpretation unless it is "plainly erroneous or inconsistent with the regulation." Auer v. Robbins, 519 U.S. 452, 461, 117 S.Ct. 905, 137 L.Ed.2d 79 (1997) (internal quotation marks and citations omitted). We afford an agency's interpretation no deference, however, if the language of the regulation is unambiguous, for doing so would "permit the agency, under the guise of interpreting a regulation, to create de facto a new regulation." Christensen v. Harris Cnty., 529 U.S. 576, 588, 120 S.Ct. 1655, 146 L.Ed.2d 621 (2000). If the regulation is ambiguous and deference is due, we have still noted that "deferential" review is not inconsequential, Ky. Waterways Alliance v. Johnson, 540 F.3d 466, 474 (6th Cir.2008), and we must be satisfied that the agency's action minimally involved a "rational connection between the facts found and the choice made." Motor Vehicle Mfrs. Ass'n of the U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983).

The issue in the instant case is relatively straightforward. Summit, the EPA, and Amici agree that the EPA has properly exercised jurisdiction to regulate the tribal lands on which Summit's facilities are located. They agree that Summit's production plant and sour gas wells are commonly owned and share a major industrial grouping, satisfying two parts of the three-part test for a stationary source under the EPA's Title V permitting program. All parties further agree that if Summit's plant and wells are aggregated into a single stationary source, the total emissions from this source would exceed the threshold required for a stationary source to be considered a major source, and Summit would be required to obtain a Title V permit to regulate its emissions of criteria pollutants. Finally, the parties agree that Summit's production plant and sour gas wells are not located on "contiguous," i.e., bordering, properties. Undisputedly, the sole disputed issue before us is whether Summit's facilities are "adjacent" to one another, thus converting them into a stationary source, and a major source, under Title V.

Together with Amici, Summit argues that the EPA's conclusion that Summit's activities are "adjacent" because they are "truly interrelated" is unreasonable, and urges this Court to vacate the EPA's determination that Summit's plant and wells constitute a single stationary source. Specifically, Summit argues that the term "adjacent" is unambiguous, and that EPA's interpretation of it defies its plain and ordinary meaning. We agree.

B. As used in the EPA's Title V plan, the term "adjacent" is unambiguous.

Summit argues that the term "adjacent," both generally and as it appears in the EPA's regulatory requirement that aggregated activities be "located on contiguous or adjacent properties," is unambiguous. The EPA disagrees, submitting, inter alia, that the term "adjacent" is unquestionably ambiguous because the EPA has never defined a specific physical distance by which it is defined or with which it is simultaneous. See Resp't Br. at 30 (connecting the premise that "adjacent" is ambiguous because it lacks a definite distance with the conclusion that "[a]s applied to plant operations that vary enormously across the different industries regulated by the [CAA], [the term "adjacent"] is at minimum ambiguous as to whether emissions units that are part of an integrated industrial operation and located 500 feet to roughly eight miles apart may be considered to be located on `adjacent' properties....").

We conclude that both the dictionary definition and etymological history of the term "adjacent," as well as applicable case law, support Summit's position.

1. The dictionary definition of "adjacent" implies physical proximity.

This Court, and others as well, have often consulted dictionaries to ascertain the meaning of words. See, e.g., Terrell v. United States, 564 F.3d 442, 451 (6th Cir. 2009) (citing MCI Telecomms. Corp. v. AT & T Co., 512 U.S. 218, 225, 114 S.Ct. 2223, 129 L.Ed.2d 182 (1994)). Our research satisfies us that dictionaries agree that two entities are adjacent when they are "[c]lose to; lying near ... [n]ext to, adjoining." American Heritage Dictionary of the English Language, available at www. ahdictionary.com (search "adjacent") (last visited May 16, 2012). The EPA does not cite, nor could we locate, any authority suggesting that the term "adjacent" invokes an assessment of the functional relationship between two activities. See, e.g., Meriam-Webster Dictionary, available at www.meriam-webster.com (search adjacent") (last visited May 16, 2012) ("not distant: nearby <the city and adjacent suburbs >; having a common endpoint or border <adjacent lots> ...; immediately preceding or following); Oxford Dictionaries, available at www.oxforddictionaries. com (search "adjacent") (last visited May 16, 2012) ("next to or adjoining something else; adjacent rooms; the area adjacent to the fire station ").

Even the etymology of the term belies the EPA's position that the term's definition is even partially based on the contextual relationship, as opposed to the geographic location, of two activities. Here, our study reveals that the word "adjacent" originated in the early fifteenth century from the combination of syllables "ad" and "jet." Online Etymology Dictionary, available at www.etymonline.com (search "adjacent") (last visited May 16, 2012). "Ad" is a "prefix expressing direction toward or in addition to, ... with regard to, [or] in relation to ... space or time ...." Id. (search "ad-") (last visited May 16, 2012) (emphasis added). As incorporated in "adjacent," the verb "jet" implies an action resulting in a certain physical resting place for an object: "to throw, cast ... with notion of `to cast (oneself) down.'" Id. (search: "adjacent" and "jet") (last visited May 16, 2012). Thus, quite literally, two things are adjacent if they have been "throw[n], cast ... down ... in relation to [one another] ... in space...." See generally Online Etymology Dictionary, supra.

The EPA argues that though they are not incorrect, dictionaries provide an incomplete definition of "adjacent," and that the functional interrelationship of two facilities must be considered because physical distance is meaningless without context. See Resp't Br. at 27 ("It is only by examining the relative context that a reasonable determination can be made as to whether two objects are `close' or `proximate' to one another."). The notion that adjacency expands and contracts with context, true as it is, must not be overextended. In the first instance, it is certainly correct that two states could be adjacent to one another in the context of a country, just as two houses could be adjacent in the context of a neighborhood. It is noted, however, that the contextual element of the overall space within which one asks whether two things are adjacent does not fundamentally change the physical and geographical nature of the question. The EPA makes an impermissible and illogical stretch when it states that one must ask the purpose for which two activities exist in order to consider whether they are adjacent to one another. Resp't Br. at 26-27 ("A distance of one mile between two properties may be considered sufficiently small to be close to each other if one is driving between them; not so if one is trying to throw a softball from one to the other."). Whether the distance between two facilities enables a given relationship to exist between them is immaterial to the concept of adjacency — it merely answers the question of whether a certain activity can or cannot occur between two locations that were, and will continue to be regardless of whether they host the activity, physically distant or physically adjacent.

Case law supports this understanding.

2. Case law supports the idea that adjacency relates only to physical proximity.

Several cases specifically address whether the term "adjacent" is ambiguous. While some courts conclude that "adjacent" is ambiguous in the limited sense of lacking an abstract definition, there is common recognition of the fact that adjacency is a purely physical and geographical, even if case-by-case, determination. United States v. St. Anthony R.R. Co. is an illustrative example. The EPA claims support in the case because of its statement that: "[adjacency] must be defined with reference to the context, at least to some extent[ ]." United States v. St. Anthony R.R. Co., 192 U.S. 524, 530, 24 S.Ct. 333, 48 L.Ed. 548 (1904). Ironically, the EPA disregards the context of the Supreme Court's assertion, which undisputedly cabined any purported ambiguity to purely physical dimensions: "[a]s the word [adjacent] is frequently uncertain and relative as to its meaning," the Court reasoned, "it might ... be regarded as more extended when used with reference to a large object than with reference to a comparatively small one. In other words, it must be defined with reference to the context, at least to some extent." Id. See also, e.g., Padula v. Lally, No. 160313, 1994 WL 16193907, at *2 (Mass.Land Ct. Feb. 9, 1994) (noting that all definitions of the word "adjacent" put before the court "suggest that ... two objects must be in reasonable proximity"); Syufy Enters. v. Home Ins. Co. of Ind., No. 94-0756, 1995 WL 129229, at *2 (N.D.Cal. Mar. 21, 1995) (stating that the "`ordinary and popular' reading of the term `adjacent' denotes a sense of physical proximity").

The most persuasive authority comes from the Supreme Court's decision in Rapanos v. United States, 547 U.S. 715, 126 S.Ct. 2208, 165 L.Ed.2d 159 (2006). Similar to this case, Rapanos concerned a regulation enacted by an administrative agency, the Army Corps of Engineers (Corps), interpreting Congress' legislation in the Clean Water Act (CWA). Id. at 723-24, 126 S.Ct. 2208. Specifically at issue in Rapanos was a provision in the CWA prohibiting the discharge of pollutants in "navigable waters," which the CWA further defined as "the waters of the United States, including the territorial seas." Id. at 723, 126 S.Ct. 2208 (citing 33 U.S.C. § 1362(7)). In its regulations implementing the CWA, the Corps interpreted "the waters of the United States" to include, in pertinent part, "[w]etlands adjacent to ... waters [and tributaries] (other than the waters that are themselves wetlands[ ])." Id. at 724, 126 S.Ct. 2208 (citing 33 C.F.R. § 328.3(a)(7)). In the action that was the impetus for suit, the Corps interpreted their own regulation to exercise jurisdiction over a private owner's "sometimes saturated" parcel of land located eleven to twenty miles from the nearest body of navigable water, a proximity it deemed sufficiently "adjacent" to constitute it part of "the waters of the United States." Id. at 720-21, 126 S.Ct. 2208.

The Sixth Circuit deferred to the Corps' interpretation of its regulation, but the Supreme Court reversed, holding that deference was not due to "the Corps' definition of `adjacent,' which ... has been extended beyond reason...." Id. at 746, 126 S.Ct. 2208. In swift refutation of the dissent's view that adjacency should be interpreted expansively, the majority held that the Court had previously used "`adjacent' and `adjoining' interchangeably" and had only previously concluded that a wetland was adjacent to a navigable body of water "`[ ][b]ecause' it `actually abut[ted] on a navigable waterway.'" Id. at 747, 126 S.Ct. 2208 (some emphasis added). Finally, the Court noted that "[h]owever ambiguous the term may be in the abstract, `adjacent'... is not ambiguous between `physically abutting' and merely `nearby.'" Id. at 748, 126 S.Ct. 2208.

Together with the ordinary and dictionary definitions of the term "adjacent," the Rapanos decision, and similar case law, points clearly toward the conclusion that the regulatory requirement that aggregated activities be "located on contiguous or adjacent properties" is unambiguous in the context in which it is here considered. For the foregoing reasons, we adopt this position.

C. The EPA's interpretation of the term "adjacent," to which no deference is owed, runs contrary to its plain meaning.

Having determined that the word "adjacent" is unambiguous, we apply no deference in our review of the EPA's interpretation of it. Christensen, 529 U.S. at 588, 120 S.Ct. 1655. Instead, we must test the agency's interpretation against the plain meaning of its unambiguous regulation. Here, we conclude that the EPA's interpretation of the requirement that activities be "located on contiguous or adjacent properties," i.e., that activities can be adjacent so long as they are functionally related, irrespective of the distance that separates them, undermines the plain meaning of the text, which demands, by definition, that would-be aggregated facilities have physical proximity. The mere fact that in a limited sense, the term "adjacent" is not susceptible to abstract definition does not render it unclear, for "however ambiguous the term may be in the abstract, `adjacent' ... is not ambiguous between `[physically proximite]' and merely `[functionally related].'" See Rapanos, 547 U.S. at 748, 126 S.Ct. 2208. In light of this conclusion, we hold that the EPA has interpreted its own regulatory term in a manner unreasonably inconsistent with its plain meaning and vacate the EPA's stationary source determination hereby directing the EPA to reassess the aggregation of Summit's facilities under the ordinary understanding of its requirement that Summit's plant and wells be located on adjacent, i.e., physically proximate, properties.

1. The EPA's unreasonable interpretation is not entitled to heightened deference merely because it is longstanding.

The EPA opposes this conclusion, claiming that because it has an established history of supplementing the traditional definition of adjacency with the concept of activities' functional relatedness, we must review its interpretation with heightened deference. Resp't Br. at 31-32 (citing Alaska Dep't of Envtl. Conservation, 540 U.S. at 487, 124 S.Ct. 983 (stating that courts "normally accord particular deference to an agency interpretation of `longstanding' duration") (quoting North Haven Bd. of Ed. v. Bell, 456 U.S. 512, 522 n. 12, 102 S.Ct. 1912, 72 L.Ed.2d 299 (1982))). The EPA's appeal to the "longstanding duration" principle is unsurprising, since even a cursory overview demonstrates that the agency has rarely, if ever, considered physical proximity the sine qua non of two pollutant-emitting activities being "adjacent" to one another. See, e.g., Letter From Edward E. Reich, EPA to Clyde B. Eller, EPA Region IX (May 16, 1980), available at www.epa.gov/region7/air/nsr/ nsrmemos/shell.pdf (last visited May 16, 2012) (two facilities 1.8 miles apart and connected by pipeline are a single source because neither facility produces products independently); Letter from Edward E. Reich, EPA to Steve Rothblatt, EPA Region V (June 30, 1981), available at www. epa.gov/region7/air/title5/t5memos/defsrce 2.pdf (last visited May 16, 2012) (two plants one mile apart are a single source because they are connected by railroad and operated together); Letter from Robert G. Kellam to Richard R. Long, EPA Region VIII (Aug. 27, 1996), available at http://www.epa.gov/region07/air/nsr/ nsrmemos/abnt.pdf (last visited May 16, 2012) (landfarm and brewery six miles apart are single source because the landfarm operation is integral to the brewery operation); Letter from Richard R. Long, EPA Region VIII to Lynn R. Menlove, Utah Dept. of Envtl. Quality (Aug. 8, 1997), available at www.epa.gov/region07/ air/nsr/nsrmemos/util-at1.pdf (last visited May 16, 2012) (pump station and production operations 21.5 miles apart and on opposite sides of a lake are a single source because connective channel transports raw materials between them and facilities have a support relationship); Letter from Cheryl L. Newton, EPA Region V, to Donald Sutton, Illinois EPA (Mar. 13, 1998), available at www.epa.gov/region7/air/nsr/ nsrmemos/acme.pdf (last visited May 16, 2012) (steel mill facilities 3.7 miles apart and separated by Lake Calumet, landfills, the I-94 expressway, and the Little Camulet River are a single source because "USEPA considers that the close proximity of the sites, along with the interdependency of the operations and their historical operation as one source ..."); Letter from Richard Long, EPA Region VIII, to Dennis Myers, Colorado Air Pollution Control Div. (Apr. 20, 1999), available at http://www.epa.gov/region07/air/nsr/ nsrmemos/amersoda.pdf (last visited May 16, 2012) ("distance alone" does not preclude a mine and plant thirty-five to forty miles apart but connected by pipeline being considered a single source given "integral connectedness of the[ ] facilities"); Forest Gil, et al. (2001) (EPA guidance that adjacency analysis should include operational, not geographical, inquiries, e.g., whether the facilities are situated to enable working integration, whether materials and people are routinely transported between them, and whether the production process is split in any way between the facilities); Letter from Pamela Blakley, EPA Region V to Don Smith, Minn. Pollution Control Agency (Mar. 23, 2010), available at http://www.epa.gov/region7/air/nsr/ nsrmemos/single.pdf (last visited May 16, 2012) (facilities separated by three miles without any physical connection are a single source because materials are transported between them daily). See also Letter from Winston A. Smith, Region VI, to Randy C. Poole, Mecklenburg Cnty. Dept. of Envtl. Protection (May 19, 1999), available at http://www.epa.gov/region07/ air/nsr/nsrmemos/we1999.pdf (last visited May 16, 2012) (two bulk gas terminals separated by .9 miles are separate sources because although close in proximity, they are not functionally related).

Though the EPA has previously considered the functional relationship between activities in assessing whether they lie on "contiguous or adjacent properties," its request of increased deference to this interpretation rests on the premise that some measure of deference is in fact owed to the EPA's interpretation in the first instance — a premise we reject in light of the unambiguousness of the regulation at issue. See Christensen, 529 U.S. at 588, 120 S.Ct. 1655. Here again, we rely on authority from the Supreme Court in Rapanos:

Rapanos, 547 U.S. at 752, 126 S.Ct. 2208 (internal citations omitted). Like Rapanos, we conclude that an agency may not insulate itself from correction merely because it has not been corrected soon enough, for a longstanding error is still an error. We thus reject the EPA's argument to the contrary and have little hesitation in taking this opportunity — the first our Court has been given in the EPA's own history of "entrenched executive error" — to vacate the agency's unreasonable interpretation of its Title V permitting plan. Id.

D. Assuming arguendo the term "adjacent" is ambiguous, the EPA's interpretation is inconsistent with the regulatory history of the EPA's Title V plan and the EPA-issued guidance memorandums.

Although we consider the unambiguousness of the regulatory term "adjacent" to be quite clear, we conclude that our holding would be the same even if we were to accept the EPA's argument that its regulation is ambiguous.

Where an agency interprets its own ambiguous regulation, we credit that interpretation with great deference. Auer, 519 U.S. at 461, 117 S.Ct. 905. Indeed, only if the agency's interpretation is plainly erroneous or inconsistent will we reject it. Id. Because even deferential review is not inconsequential, however, see Ky. Waterways Alliance, 540 F.3d at 474, we will not grant administrative agencies unbridled discretion to interpret the terms of their own regulations in any conceivable manner. See generally Auer, 519 U.S. at 461, 117 S.Ct. 905; Motor Vehicle Mfrs. Ass'n, 463 U.S. at 42-43, 103 S.Ct. 2856. Here, even if we were to find that the word "adjacent" was ambiguous, we would still find that the EPA's interpretation of its regulatory requirement that Title V single stationary sources are located on "contiguous or adjacent properties" was entirely inconsistent with: (1) the regulatory history of its Title V permitting plan; and (2) its own guidance memorandums regarding the application of its Title V regulations to the oil and gas industry. We would thus conclude that EPA's interpretation was unreasonable under Auer, and our decision in this case would be the same.

1. The EPA's interpretation is inconsistent with the regulatory history of its Title V plan.

On the issue of the EPA's regulatory history interpreting Title V, we first note the incongruity between the EPA's refusal to include a functional relationship test in its single stationary source analysis and its current position that its analysis cannot be completed without it.

In order for multiple activities to be together considered a single "building," "structure," "facility," or "installation" within the meaning of the EPA's Title V plan, the activities must: (1) be under common control; (2) be located on contiguous or adjacent properties; and (3) belong to the same industrial grouping. 40 C.F.R. § 71.2. The EPA first established this stationary source test for use in its Prevention of Significant Deterioration (PSD) program. See 40 C.F.R. § 52.21(b)(6) ("Building, structure, facility, or installation means all of the pollutant-emitting activities which belong to the same industrial grouping, are located on one more contiguous or adjacent properties, and are under the control of the same person (or persons under common control)...."). It developed the test in the wake of the Alabama Power Co. v. Costle decision, a 1979 challenge to the EPA's former PSD regulations, adjudicated by the United States Court of Appeals for the D.C. Circuit. See Alabama Power Co. v. Costle, 636 F.2d 323 (D.C.Cir.1979). Assessing the EPA's then-existing stationary source test, the Alabama Power Court advised that "[t]o allow an entire plant or other appropriate grouping of industrial activity to be subject as a single unit ... EPA should devise regulatory definitions of the terms `structure,' `building,' `facility,' and `installation' to provide for the aggregation, where appropriate, of industrial activities according to considerations such as proximity and ownership." Id. at 397. The court also required that the EPA make its aggregation test sufficiently predictable to give plausibly regulated entities "explicit notice as to whether (and on what statutory authority) the EPA construes the term [stationary] source...." Id.

In the preamble to the 1980 amendments to the final PSD rules, the EPA recognized that Alabama Power required a new definition of a single stationary source to, inter alia, "approximate a common sense notion of a `plant[ ]' [and] . . . avoid aggregating pollutant-emitting activities that as a group would not fit within the ordinary meaning of `building,' `structure,' `facility,' or `installation.'" Requirements for Preparation, Adoption, and Submittal of Implementation Plans; April 4, 2012 Approval and Promulgation of Implementation Plans, 45 Fed. Reg. 52,676 (Aug. 7, 1980) (to be codified at 40 C.F.R. pts. 51, 52, and 124). Consistent with this understanding, the EPA declined to adopt the court's proposed dual-prong proximity/ownership test, finding that this "definition [of a single source] would fail to approximate a common sense notion of a `plant,' since in a significant number of cases it would group activities that ordinarily would be regarded as separate." 45 Fed. Reg. at 52,695. The EPA therefore set out to add a third criteria which would narrow the field of aggregated sources to those more in line with the "common sense notion of a plant." Id.

The functional relationship test, which EPA admittedly now uses to determine whether activities are adjacent to one another, was one of the additional criteria the EPA considered and rejected during its post-Alabama Power rule-making process. First, the EPA "asked for comment on whether factors other than proximity and control, such as the functional relationship of one activity to another, should be used [in stationary source determinations]." 45 Fed. Reg. at 52,694. Then, after due consideration, the EPA rejected the option of adding a third requirement (in addition to ownership and proximity) that the activities comprising a single stationary source be functionally related to one another. 45 Fed. Reg. at 52,695. The EPA specifically found that assessing whether activities were sufficiently functionally related to constitute a single source "would be highly subjective" and would make "administration of the definition substantially more difficult, since any attempt to assess those interrelationships would have embroiled the Agency in numerous, fine-grained analyses." Id. In lieu of the rejected functional interrelationship test, the EPA elected to incorporate the "industrial grouping" criterion, a factor which asks whether the multiple activities are engaged in the same type of business, i.e., operate under the same SIC code. See supra n.2; 40 C.F.R. § 71.2.

a. The EPA's mis-characterization of its own regulatory history is unconvincing.

The EPA argues that its prior rejection of the proposal that "functional relatedness" be a third prong of its stationary source test is immaterial because its decision not to use this as a stand-alone factor is not inconsistent with its use of the concept as a supplement to its adjacency analysis. In at least two ways, neither Title V's regulatory history nor common sense supports the EPA's position.

First, the fact that the EPA considered adding to the tests of ownership and proximity with a test of functional relatedness belies its current contention that the factors of proximity and functional relatedness are one in the same. The EPA admits that it retained Alabama Power's suggested ownership and proximity test, seeking only a third test to reduce the number of single stationary sources to those representative of a common plant. If the adopted proximity requirement — articulated by the EPA as the requirement that activities be "located on contiguous or adjacent properties" — suggested a measure of the activities' functional relatedness, there would have been no need to consider adopting this already-included test as a third prong in the EPA's single source analysis. Clearly, at least at the time the EPA adopted its stationary source regulation, the EPA understood adjacency in geographical, rather than operational, terms.

Second, the EPA's decision not to employ a functional relatedness test was categorical and unqualified. Far from advocating the use of a functional relatedness assessment as part but not all of the stationary source analysis, the EPA expressed the sweeping conclusion that any reference to the operational relationship between activities "would be highly subjective," would "make administration of the definition substantially more difficult," and would burden "the Agency in numerous, fine-grained analyses." 45 Fed. Reg. 52,695. This broad condemnation cuts heavily against the EPA's argument that the use of any assessment of functional relatedness is reasonable (even in its own view). Ultimately, the EPA has failed to convince us why the criterion it considered, and then strongly rejected, is now indispensable to its ability to make adequate stationary source determinations. The foregone conclusion, given the absence of notice and comment procedures amending the EPA's Title V plan to include an assessment of activities' functional relatedness, is that the EPA's current test is in practice completely inconsistent with the wording of its own regulation.

2. The EPA's interpretation is inconsistent with its own guidance memorandums.

We next note that just as the EPA's present interpretation of adjacency is inconsistent with the regulatory history of its Title V permitting plan, so too is it inconsistent with the very EPA-issued guidance memorandums on which the EPA relies in the instant case. As discussed supra, the EPA's first two letters to Summit cited the Wehrum Memorandum as primary authority in its decision-making process. The Wehrum Memorandum had particular relevance to the EPA's single source determination for Summit — it was issued to guide permitting authorities specifically on the application of Title V to the aggregation of pollutant-emitting activities in the oil and gas industry. The Wehrum Memorandum rejected the notion that geographically distant oil and natural gas activities could be considered "contiguous" or "adjacent:"

Wehrum Memorandum, supra. We need not strain to conclude that the EPA's source determination for Summit, in which the EPA admittedly found that aggregation of Summit's sweetening plant and sour gas wells was appropriate despite their physical separation, is patently inconsistent with this guidance.

While the EPA withdrew the Wehrum Memorandum and issued the McCarthy Memorandum shortly before the final iteration of its single source determination for Summit, the McCarthy Memorandum does not support the EPA's analysis, either. To be sure, the McCarthy Memorandum criticized the Wehrum Memorandum's emphasis on physical proximity, noting that such was not "a sufficient endpoint in the decision-making process." McCarthy Memorandum, supra. Apart from revoking this prior guidance, however, the McCarthy Memorandum promotes a neutral and plain meaning application of the EPA's Title V plan, directing permitting authorities to "rely foremost on the three regulatory criteria" for aggregating emissions from multiple sources, i.e., "(1) whether the activities are under the control of the same person (or person under common control); (2) whether the activities are located on one or more contiguous or adjacent properties; and (3) whether the activities belong to the same industrial grouping." Id. (citing 40 C.F.R. § 52.21(b)(6)). As discussed supra, the McCarthy Memorandum concedes that at the end of any case-by-case analysis, proximity may well be the overwhelming and determinative factor. If the McCarthy Memorandum says anything about including supplemental and unexpected factors in the Title V stationary source analysis, it cautions against it.

There is some dispute between the parties as to which of the EPA's letters declaring Summit's plant and wells a single stationary source was the EPA's "final agency determination." Summit argues that the EPA's October 2010 correspondence was a mere reiteration of its final determination in September 2009; thus, Summit contends, the Wehrum Memorandum applied to the EPA's decision and its emphasis on physical proximity in the adjacency analysis should have caused Summit's facilities to be considered separate emissions sources. Ultimately, whether the Wehrum Memorandum had been superceded by the McCarthy Memorandum at the time of the EPA's final determination would not impact our decision. Although the Wehrum Memorandum made clear the near-singular importance of the physical proximity of two facilities in the EPA's stationary source analysis, the McCarthy Memorandum supports Summit's position with its focus on applying the regulatory text in Title V decisions. In either case, then, the EPA-issued guidance memorandums support the error in the EPA's interpretation of its own regulatory term.4

E. Arguments by Amicus API support our conclusion.

Our decision finds support in many of the arguments raised in the brief of Amicus API.5 First, as a member of the oil and gas industry itself, API urges us to consider the fact that the EPA's interpretation of adjacency to include a measure of functional relationship is subjective and unpredictable, undermining the Alabama Power Court's mandate that the EPA's test give "explicit notice as to whether (and on what statutory authority) the EPA construes the term [stationary] source...." See Alabama Power, 636 F.2d at 397. Relatedly, API criticizes the EPA for supporting its source determination with the concept of a "common sense notion of a plant," reasoning that this general principle is clearly meant to constrain, rather than enlarge, the EPA's definition of a single stationary source.

API also argues that the EPA's interpretation has led in this case, and will continue to lead, to absurd source determinations within the industry. In a context where oil and gas producers often have little say over where emission points are located and nearly every facility is connected to one another by pipeline (and thus, in a sense, functionally related), API argues, the aggregation of activities without regard to their physical proximity is inherently unreasonable. Indeed, with the concept of a "common sense notion of a plant" construed as facilitating the aggregation of operationally related but physically distant activities, there is little foreseeable limit to the aggregation of emission points spread out literally across the country. Moreover, API cautions that source determinations dependent on activities' functional relatedness will be quickly outdated as oil and gas producers add and remove emission points from their production systems, potentially requiring nearly constant renewed source determination requests and decisions.

We are particularly struck by API's final observation — that the EPA's source determination in this case is an ironic showcase of the very fears that caused the agency not to adopt a functional relatedness test for source determinations in the first instance. From Summit and MDEQ's joint source determination request in January 2005 to the final iteration of the EPA's decision, nearly five years passed. The parties engaged in at least twenty-five conference calls and exchanged "a small mountain of paper" within this period. Certainly, the cost for Summit to produce the data and schematics requested by the EPA, and the cost for the EPA to distill and understand the same, were high in terms of both monetary and capital resources. As API points out, this process produced exactly the "fine-grained" and administratively burdensome result the EPA sought to avoid in its drafting of its Title V stationary source test, see 45 Fed. Reg. at 52,695, and is strong evidence that the EPA's interpretation of its own regulation was unreasonable.

III. Conclusion

For the reasons stated, we VACATE the final agency determination of the EPA aggregating Summit's sour gas wells and sweetening plant into a single and major source under its Title V permitting plan. We REMAND this case to the EPA for a reassessment of Summit's Title V source determination request in light of the proper, plain-meaning application of the requirement that Summit's activities be aggregated only if they are located on physically contiguous or adjacent properties.

KAREN NELSON MOORE, Circuit Judge, dissenting.

Congress passed the Clean Air Act ("CAA") "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." 42 U.S.C. § 7401(b)(1). Because the majority hamstrings the Environmental Protection Agency's ("EPA") ability to pursue this mission by refusing to defer to the agency's reasonable interpretation of its own regulation, I respectfully dissent.

I. REVIEW OF AGENCY ACTION

As the majority acknowledges, we defer to an agency's interpretation of its own ambiguous regulation as "controlling" unless that interpretation is "`plainly erroneous or inconsistent with the regulation.'" Auer v. Robbins, 519 U.S. 452, 461, 117 S.Ct. 905, 137 L.Ed.2d 79 (1997) (quoting Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 359, 109 S.Ct. 1835, 104 L.Ed.2d 351 (1989)); Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 413-14, 65 S.Ct. 1215, 89 L.Ed. 1700 (1945); Claiborne-Hughes Health Ctr. v. Sebelius, 609 F.3d 839, 844 (6th Cir.2010). As in the context of agency interpretation of statutes, judicial deference to an agency's interpretation of regulations is premised on the relative institutional expertise and political accountability of an agency as compared to a court; moreover, an agency's authority to interpret its own regulations implementing a statute is incident to the authority to interpret that statute. See Matthew C. Stephenson & Miri Pogoriler, Seminole Rock's Domain, 79 Geo. Wash. L. Rev. 1449, 1456-57 (2011).

In addition to the deference we give to administrative interpretations of regulatory language, we will not overturn agency action unless it is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A); see also Motor Vehicle Mfrs. Ass'n, Inc. v. State Farm Mut. Auto. Ins. Co. (State Farm), 463 U.S. 29, 42-43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983).

II. MAJOR-SOURCE AGGREGATION

The EPA will aggregate the emissions of multiple stationary sources to determine if the sources together constitute a major source if those sources (1) "are under common control of the same person," (2) "are located on one or more contiguous or adjacent properties," and (3) "belong[ ] to a single major industrial grouping." 40 C.F.R. § 71.2. In this case, the EPA determined that Summit's sour-gas wells, flares, and gas-sweetening plant, which are connected by a dedicated underground pipeline and are used exclusively as part of a single, interconnected operation, together constituted a major source. As part of its determination that Summit's various stationary sources were located on adjacent properties, the EPA considered the fact that those sources were functionally interrelated.

The majority bases its conclusion that the EPA's interpretation of its own regulation was impermissible on three factors: the plain meaning of the word "adjacent," the regulatory history of CAA major-source aggregation determinations, and policy considerations. In my view, none of these factors prevent us from deferring to the EPA and upholding its decision in this case.

A. The Meaning of "Adjacent"

The majority peruses several dictionaries and concludes that "adjacent" refers to the physical or geographical relationship between two or more objects. The objects must be "not distant," "next to," "close to," or "immediately preceding or following." The EPA's interpretation of its Title V regulations to allow for consideration of functional interrelatedness in adjacency determinations does not contradict this plain meaning of the word "adjacent." The ultimate question remains whether two properties are physically close enough to be considered adjacent for the purposes of aggregation under Title V. The EPA's position is simply that other factors in addition to absolute physical distance can be relevant in determining whether two objects that are a given distance apart are close enough to be considered adjacent in this context. The EPA does not contend that multiple stationary sources are adjacent simply because they are functionally interrelated, and the majority either misunderstands or mischaracterizes the agency's position in so describing it.

In this more limited sense, "adjacent" is ambiguous. The fact that "adjacent" refers to the distance between two objects does not explain what factors go into a determination of whether two objects that are a given distance apart are "next to" each other, "close," or "immediately preceding or following." The cited definitions do not unambiguously foreclose the consideration of factors other than absolute physical distance.1 One object in a sequence "immediately preced[es] or follow[s]" the next item in the sequence, for example. Two objects that are physically connected fall "next to" each other along that connection. Two or more stationary sources that are functionally interrelated can similarly be described as "next to" each other; one follows the other in a common process.

Moreover, functional interrelatedness can inform the determination of whether two objects that are a given distance apart are adjacent. If two properties are close enough to each other to house stationary sources that contribute to the same interrelated operation, and only to that operation, those properties are more likely to be close enough reasonably to be considered adjacent. Likewise, the EPA could reasonably conclude that two or more sources that exist only as part of the same larger process or sequence will likely be close enough to each other to be considered adjacent. As the EPA recognizes, circumstances may exist in which the distance between two stationary sources is too great for those sources to be considered adjacent, even if they are functionally interrelated. This fact does not mean that interrelatedness can never be a factor, but that it will not support a finding of adjacency in that instance.

The consistency of the EPA's approach with the regulation is apparent from the record in this case. The EPA did not ignore the physical dimension of adjacency; the EPA's decision to aggregate Summit's wells, flares, and plant was not, as the majority alleges, made "irrespective of the distance that separates them." Distance was a factor in concluding that the various sources were located on adjacent properties. In its April 26, 2007 letter to Summit, for example, the EPA requested a map showing the location of each emission unit, as well as the potential to emit of each unit located within one mile of the gas-sweetening plant. Pet'r App. at 31. The EPA discussed the distance between Summit's gas-sweetening plant, gas wells, and flares in its October 18, 2010 letter explaining its decision to aggregate each source in Summit's operations as a major source, noting, for example, that "it appears that there are a dozen or more sour gas wells within a one mile-radius of the sweetening plant." Resp. App. at 90-91 & n.21. If the EPA had truly considered only functional interrelatedness, it would have declared Summit's operation to be a major source in its initial April 2007 letter after noting that each source was interrelated and would not have sought additional information regarding the distance between the emission units and the plant. See Pet'r App. at 31.

Moreover, in Summit's case, functional interrelatedness has a physical dimension; each stationary source is connected via a dedicated pipeline. The operation is thus not only functionally interrelated but physically interconnected. Each source (gas well, flare, and sweetening plant) is a stop along a single physically connected process; each well is "next to" and "immediately preced[es]" the following well, for example. Nothing outside of this process either physically or functionally interrupts the traverse of gas from an underground field through the gas wells and past the flares to the sweetening plant. The properties on which each well or flare are located are likewise connected to each other, and to the property on which the plant is located.2

The EPA's position thus comports with the plain meaning of the word "adjacent" as a description of the geographical relationship between two objects. The EPA uses interrelatedness only to determine whether two properties are close enough to each other to be considered adjacent. Functional interrelatedness thus operates only in that realm of adjacency that is ambiguous; it does not replace proximity, but serves as a means of determining proximity. Accordingly, deference is warranted.

The majority's cited caselaw does not dictate a contrary result. Indeed, the majority's "most persuasive authority" on the meaning of "adjacent" within the CAA regulations is a Supreme Court decision that involved a different statute and did not purport to provide a general definition of the word "adjacent."3 In Rapanos v. United States, 547 U.S. 715, 740-42, 126 S.Ct. 2208, 165 L.Ed.2d 159 (2006), which dealt with the Clean Water Act ("CWA"), a plurality of the Court concluded that, for a wetland to fall within the meaning of "waters of the United States" and thus be subject to regulation under the CWA, the wetland must actually abut a navigable waterway. In its earlier opinion in United States v. Riverside Bayview Homes, Inc., 474 U.S. 121, 106 S.Ct. 455, 88 L.Ed.2d 419 (1985), the Court had held that the Army Corps of Engineers's interpretation of the CWA to cover certain wetlands was reasonable because the precise boundary between water and land is not always clear and an expert agency is best positioned to decide how far the CWA's regulatory scope should extend in such situations; in Riverside Bayview, the Court had described the wetlands that could be regulated under the CWA as those that were "adjacent" to navigable waters. Rapanos, 547 U.S. at 740-41, 126 S.Ct. 2208. Because only wetlands that actually abut waters implicate the issue of how to determine a land/water boundary, the Court in Rapanos clarified that its use of the word "adjacent" in Riverside Bayview meant only physically abutting and, accordingly, that only those wetlands that physically abut navigable waters could be considered adjacent to such waters for CWA purposes. Id. at 742, 126 S.Ct. 2208. Indeed, the plurality expressly stated that "[i]n expounding the term `adjacent' as used in Riverside Bayview, we are explaining our own prior use of that word to interpret the definitional phrase `the waters of the United States'" and that "[h]owever ambiguous the term may be in the abstract, ... `adjacent' as used in Riverside Bayview is not ambiguous between `physically abutting' and merely `nearby.'" Id. at 747-48, 126 S.Ct. 2208. Given the specific, limited nature of this discussion, it really has no applicability to any inquiry into the meaning of "adjacent" more generally or in any other context. In determining otherwise, the majority mistakes the sea for the air, leaving the fish and the fowl confused.

If anything, Rapanos actually undermines Summit's position by demonstrating that the meaning of "adjacent" is contextual and that the determination of whether two objects are adjacent can depend upon the purpose for which the adjacency decision is being made. In Rapanos, the question was how close a wetland must be to navigable waters to qualify as "waters of the United States," and the Court held that "adjacent" in that context meant physically abutting because only abutting wetlands could reasonably constitute "waters" at all. See id. at 741 n. 10, 126 S.Ct. 2208. Different factors are relevant when the question is how close multiple stationary sources must be to each other to qualify as an aggregated major source.

Because "adjacent" is ambiguous as to how to determine whether two objects that are a given distance apart should be considered adjacent, we should defer to the EPA's decision to consider functional interrelatedness as a means of making that determination. Moreover, the values which deference promotes would be served by doing so. Whether two or more stationary sources are located on adjacent properties goes to whether the EPA should aggregate those individual sources; ultimately, the regulatory term that the EPA is interpreting when considering whether properties are adjacent is "major source." See 40 C.F.R. § 71.2.4 The consequence of aggregating the various stationary sources within an operation is that the operation will be subject to the CAA's Title V permitting requirements, and whether an industrial operation should be subject to CAA regulations is a decision best left to the expert agency tasked with enforcing that law. The EPA's aggregation decisions reflect its institutional expertise regarding how various industrial operations affect air quality. In all its talk of adjacency, the majority forgets the purpose of the adjacency determination. To mix environmental metaphors, the majority's analysis of the EPA's clean-air regulations misses the forest for the trees.

B. Regulatory History

In concluding that it would not defer to the EPA's interpretation of "adjacent" even if it found that term to be ambiguous, the majority relies heavily on the Preamble to the EPA's 1980 regulations defining "stationary source" under the CAA's Prevention of Significant Deterioration program, in which the EPA explained its decision not to adopt functional interrelatedness as an independent factor that applied to all aggregation determinations across all industries. See 45 Fed. Reg. 52,676, 52,693-95 (Aug. 7, 1980). The majority first reasons that the EPA would not have considered adding a separate interrelatedness requirement if adjacency and interrelatedness are the same. This argument is logical, but misconstrues the EPA's position. The EPA does not equate adjacency with interrelatedness; the latter is a factor that in some cases can help determine whether the requirement of the former is met. Because adjacency does not depend upon interrelatedness, the latter would have had independent force as a separate requirement.

Next, the majority characterizes the Preamble as a categorical rejection of any consideration of functional interrelatedness. Although the Preamble rejects interrelatedness as "another abstract factor" to be analyzed in all aggregation determinations, 45 Fed. Reg. at 52,695, nothing in the Preamble forecloses the possibility that interrelatedness could be relevant in a specific aggregation determination or otherwise rejects interrelatedness as a criterion that could be considered within the adjacency analysis. As discussed below, functional interrelatedness in the gas-drilling industry is both particularly pertinent and less likely to involve the kind of "subjective" assessments or "fine-grained analyses" that the EPA cited as reasons not to adopt interrelatedness as an across-the-board requirement, see id. Moreover, the historical record belies the majority's characterization of the Preamble as a categorical rejection of interrelatedness. The EPA considered interrelatedness in its aggregation determinations within ten months of the 1980 Preamble — evidence that the consideration of interrelatedness is not a recent phenomenon that somehow deviates from the agency's original regulatory intent. See Letter from Edward Reich, EPA to Steve Rothblatt, EPA Region V (June 30, 1981) (concluding that "two facilities [that] are approximately one mile apart, have a dedicated railroad line between them and are programmed together to produce one line of automobiles... can be considered adjacent").

The Preamble also makes clear that aggregation as a major source is warranted if an operation reflects the "common sense notion of [a] plant," and that the three requirements for aggregation — common ownership, adjacency, and a shared industrial grouping — are intended to answer that question. See 45 Fed. Reg. at 52,694-95. Different industries are structured differently. In the oil and gas industry, activities that in other contexts might be housed within a single building may instead be spread out across different parcels of land. Viewing the distance between two facilities in a drilling operation in light of their functional interrelatedness thus serves the purpose of identifying operations that reflect the "common sense notion of [a] plant."

Finally, the fact that the EPA's aggregation determination in this case is somewhat inconsistent with the policy preferences expressed in the withdrawn Wehrum Memo does not render that determination impermissible. The Wehrum Memo relies on policy reasons for focusing on distance rather than interrelatedness, not an argument based on the meaning of "adjacent." Similarly, the Wehrum Memo would not need to identify physical proximity as "the most informative factor" in an adjacency determination if it was the only possible factor. See Pet'r App. at 35.

C. Policy Considerations

Our role is not to judge the wisdom of agency action, but its permissibility. To the extent that Summit's policy-based challenges to the EPA's interpretation of its own Title V regulations are even relevant, they do not hold up to scrutiny.

Summit and the majority stress the value of predictability in a regulatory regime and emphasize that the EPA itself recognized the danger of having to conduct "numerous, fine-grained analyses" in aggregation determinations. See 45 Fed. Reg. at 52,695. What Summit and the majority fail to recognize, however, is that such concerns are not implicated in this case. Summit's operations are clearly interrelated. Each well supplies sour gas only to Summit's gas-sweetening plant. The plant does not receive sour gas from any other source. Each unit is physically connected by dedicated pipeline. Summit does not contend otherwise. No "highly subjective" or "fine-grained analys[i]s" was required.5

Moreover, the EPA has long considered interrelatedness as a factor in major-source aggregation determinations, including in the oil and gas industry, so Summit cannot credibly claim surprise that the agency did so in this case. The concern that deference to an agency's interpretation of its own regulations would "require regulated parties to divine the agency's interpretations in advance or else be held liable when the agency announces its interpretations for the first time in an enforcement proceeding and demands deference" is not implicated. Christopher v. SmithKline Beecham Corp., ___ U.S. ___, 132 S.Ct. 2156, 2168, 183 L.Ed.2d 153 (2012). The majority is correct that a longstanding agency interpretation is not worthy of deference if that interpretation conflicts with the regulation; absent such conflict, however, the longstanding duration of an agency's interpretation certainly weighs towards deference. See Barnhart v. Walton, 535 U.S. 212, 219-20, 122 S.Ct. 1265, 152 L.Ed.2d 330 (2002).

Further, as the majority recognizes, "adjacent" does not have a set meaning as to any specific distance. Absent a bright-line rule as to how far is too far for numerous sources to be considered adjacent, any aggregation determination that considers distance alone could involve the very "subjective" assessments and "fine-grained analyses" that Summit fears. Stripped of the ability to consider the interrelatedness of Summit's wells and plant, the EPA would have had to consider the distance between each individual well, along with the distance between each well and the plant, and decide whether each distance was too far to be considered adjacent. In Summit's case, this process could require over one hundred separate adjacency determinations. Other drilling operations that are similarly structured but cover different distances or traverse different landscapes might not be able to rely on the determination in Summit's case in predicting whether they would be subject to aggregation.

The majority's adoption of Summit's position raises its own policy concerns. Primarily, today's ruling frees the oil and gas industry to gerrymander its way out of Title V regulation. So long as sufficient distance exists between each well (so that they are not "adjacent" as the majority defines that term), or someone other than the drilling company owns parcels of land in between each well (so that they are not "contiguous"), the drilling operation cannot be classified as a major source through aggregation. Unlike the CAA provisions governing hazardous air pollutant emissions, see 42 U.S.C. § 7412(n)(4), Title V does not grant the oil and gas industry immunity from aggregation; this court should not effectively create such a provision when Congress has not done so.

III. CONCLUSION

Because I believe that the EPA's consideration of functional interrelatedness as a factor along with physical distance in its adjacency determination was both reasonable (and thus worthy of deference) and correct, I would affirm its decision to aggregate the various stationary sources in Summit's drilling operation as a major source. Recognizing that the majority concludes otherwise, I note in closing that, aside from essentially holding that functional interrelatedness is an impermissible factor to consider, the majority does not find that any other aspect of the EPA's aggregation determination was flawed under the Administrative Procedure Act and the standards for agency decisionmaking articulated in State Farm, 463 U.S. at 42-43, 103 S.Ct. 2856. On remand, then, the EPA is free to reach the same conclusion that Summit's operations should be aggregated as a major source for Title V permitting purposes, so long as it bases that conclusion on the considerations that the majority today deems appropriate. See Sec. & Exch. Comm'n v. Chenery Corp., 332 U.S. 194, 196, 200-01, 67 S.Ct. 1575, 91 L.Ed. 1995 (1947).