DESERT CITIZENS AGAINST POLLUTION and Sierra Club, Petitioners v. ENVIRONMENTAL PROTECTION AGENCY and Lisa Perez Jackson, Administrator, U.S. Environmental Protection Agency, Respondents Nevada Mining Association, Intervenor, 699 F.3d 524

Opinion for the Court filed by Senior Circuit Judge WILLIAMS. WILLIAMS, Senior Circuit Judge:

Section 112(c)(6) of the Clean Air Act requires action by the Environmental Protection Agency on seven bioaccumulative hazardous air pollutants (“HAPs”), each named specifically by Congress. EPA is to list each pollutant’s sources and to “assur[e] that sources accounting for not less than 90 per centum of the aggregate emissions of each such pollutant are subject to standards under subsection (d)(2) or (d)(4)” of § 112. Clean Air Act § 112(c)(6), 42 U.S.C. § 7412(c)(6).

In a rulemaking effective February 17, 2011, EPA identified gold mine ore processing and production as a source for purposes of emissions of mercury, one of the seven HAPs named in § 112(c)(6). 76 Fed.Reg. 9450/1 (the “Gold Mine Rule”). In its response to comments, EPA took two positions contested here by petitioners Desert Citizens Against Pollution and Sierra Club.

First, EPA rejected the claim that § 112(c)(6)’s cross-reference to § 112(d)(2) (in the instances where (d)(2) rather than (d)(4) applies) requires that EPA subject all HAPs emitted by a § 112(c)(6) source — even those not enumerated in § 112(c)(6) — to standards at the stringency level specified by § 112(d)(2). See 76 Fed.Reg. at 9457. Second, EPA made clear that, despite language in the Gold Mine Rule arguably suggesting that it covered “fugitive emissions” — namely emissions from certain sources such as “tailings ponds, leach fields, and waste rock piles” — in fact the rule did not address such emissions.

Id. at 9457/3-58/1. Petitioners timely challenged the rule-making on both issues. We address these claims in the above order, rejecting both. Does § 112(c)(6) require EPA to impose the same stringency levels in standards for non- § 112(c)

(6) HAPs occur ring at § 112(c)(6) sources that it does for § 112(c)

(6) HAPs? We start with a brief review of the statutory context. In the early years of the Act, Congress left the choice of which HAPs to regulate largely to EPA’s discretion. See New Jersey v. EPA 517 F.3d 574, 578 (D.C.Cir.2008). But in 1990 Congress amended the Act to list 189 specific HAPs, including mercury compounds, 42 U.S.C. § 7412(b)(1), and then prescribed a two-step process whereby EPA would regulate their emission. Under the first step, EPA lists “major” and “area” sources of the HAPs, a distinction we have discussed at length elsewhere. See, e.g., Nat’l Mining Ass’n v.

EPA 59 F.3d 1351, 1353-54 (D.C.Cir.1995). (Briefly, “major sources” are those that emit 10 or more tons of a specific HAP annually, or 25 or more tons of any combination of HAPs, 42 U.S.C. § 7412(a)(1), and are generally “subject to stricter regulatory control than are ‘area sources,’ ” Nat’l Mining Ass’n, 59 F.3d at 1353. An “area source” is “any stationary source of [HAPs] that is not a major source,” 42 U.S.C. § 7412(a)(2); their listing and regulation is more discretionary and context-dependent than is the case for major sources.

For example, under § 112(c)(3), “area sources representing 90 percent of the area source emissions of the 30 [HAPs] that present the greatest threat to public health in the largest number of urban areas” are subject to nondiscretionary listing, whereas EPA “does not have to establish emission standards for unlisted area sources.” Nat’l Mining Ass’n, 59 F.3d at 1353.) In the second step, EPA promulgates emission standards pursuant to the procedures and criteria outlined in various paragraphs of § 112(d), 42 U.S.C. § 7412(d).

In the paragraph at issue here, § 112(c)(6), Congress additionally singled out seven specific persistent, bioaccumulative HAPs — some of them separately listed in § 112(b)(1) — and required EPA to list their sources and promulgate emissions standards.

In doing so, Congress did not employ the distinction between “major” and “area sources”: With respect to [the seven named HAPs] the Administrator shall, not later than 5 years after November 15, 1990, list categories and subcategories of sources assuring that sources accounting for not less than 90 per centum of the aggregate emissions of each such pollutant are subject to standards under subsection (d)(2) or (d)(4) of this section. 42 U.S.C. § 7412(c)(6).

Although, like § 112(c)(3), § 112(c)(6) imposes a special deadline for listing sources that account for 90% of specified emissions (in the case of § 112(c)(3), emissions of the 30 most hazardous HAPs), it is unique in denying EPA any choice in the selection of HAPs chosen for special treatment.

Section 112(d)(2), in turn, sets out a level of stringency known as “maximum achievable control technology” or “MACT”: Emissions standards promulgated under this subsection and applicable to new or existing sources of hazardous air pollutants shall require the maximum degree of reduction in emissions of the hazardous air pollutants subject to this

section (including a prohibition on such emissions, where achievable) that the Administrator, taking into consideration the cost of achieving such emission reduction, and any non-air quality health and environmental impacts and energy requirements, determines is achievable for new or existing sources in the category or subcategory to which such emission standard applies.... 42 U.S.C. § 7412(d)(2) (emphasis added). “[M]ajor sources must comply with ...

MACT standards.” Nat’l Mining Ass’n, 59 F.3d at 1353. “For listed area sources, EPA may choose to promulgate emission standards requiring only ‘generally available control technologies or management practices,’ ” or GACT. Id. We review the competing statutory constructions under the familiar standards of Chevron, U.S.A., Inc. v. NRDC, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), first determining whether there is a relevant textual ambiguity in the statute, and then, if there is, deciding whether the implementing agency’s construction is reasonable. Nat’l Cable & Telecomm. Ass’n v.

Brand X Internet Servs., 545 U.S. 967, 980, 125 S.Ct. 2688, 162 L.Ed.2d 820 (2005) (citing Chevron, 467 U.S. at 843-44 & n. 11, 104 S.Ct. 2778). Petitioners’ claim turns entirely on § 112(c)(6)’s cross-reference to §§ 112(d)(2) and (d)(4). (More on § 112(d)(4) momentarily.) Looking to the language of (d)(2) emphasized above, petitioners argue that its phrase “emissions of the hazardous air pollutants subject to this section” means that whenever EPA creates MACT standards for § 112(c)

(6) HAPs for a source, it must similarly impose MACT standards for emissions from that source of any HAP listed anywhere in § 112 (“this section”) — including the 189 HAPs listed in § 112(b)(1). Petitioners’ reading of the statute is linguistically possible. After all, § 112(c)(6) directs EPA to assure that “the emissions of each such pollutant [the seven § 112(c)

(6) HAPs] are subject to standards under subsection (d)(2) or (d)(4),” and (d)(2) says that “[e]missions standards promulgated under this subsection” must require MACT reductions “in emissions of the hazardous air pollutants subject to this section,” which would seem to mean all HAPs identified in § 112. Further, we have read subparagraphs (1) and (3) of § 112(

d) to require the regulation of all HAPs listed in § 112(b)(1). See, e.g., Nat’l Lime Ass’n v. EPA, 233 F.3d 625, 633-34 (D.C.Cir.2000), Sierra Club v. EPA, 479 F.3d 875, 883 (D.C.Cir.2007). And, consistently with petitioners’ view of the phrase “subject to this

section ” in § 112(d)(2), we have recently recognized that Congress’s usual “hierarchical scheme in subdividing statutory sections” refers to a

section of the U.S.Code, followed by subsections, paragraphs, subparagraphs, and clauses. U.S. v. Hines, 694 F.3d 112, 118 (D.C.Cir.2012) (citing Koons Buick Pontiac CMC, Inc. v. Nigh, 543 U.S. 50, 60-61, 125 S.Ct. 460, 160 L.Ed.2d 389 (2004)). But however linguistically possible petitioners’

interpretation, it is not unambiguously correct. The textual ambiguity does not arise from § 112(d)(2), but from § 112(c)(6), and lies in the phrase “subject to standards under subsection (d)(2) or (d)(4) of this section.” “Standards under subsection (d)(2)” could be given the construction that petitioners advance, namely, that “standards under” incorporates every word of (d)(2), thereby triggering MACT standards for non- § 112(c)

(6) HAPs emitted by a § 112(c)(6) source. But alternatively Congress may have plausibly intended simply to set MACT as the standard for the seven § 112(c)

(6) HAPs, as opposed to the less restrictive GACT standard specified in § 112(d)(5). This reading makes particular sense given that the usual criterion for selecting MACT versus GACT standards — whether a source is “major” or “area” — is missing from the framework established by § 112(c)(6). As EPA pointed out in the rulemaking, petitioners’

interpretation would have the anomalous effect of changing the required stringency of non- § 112(c)

(6) HAPs at a given area source — from the GACT level to the more demanding MACT level — simply on the fortuity that the non- § 112(c)

(6) HAPs in question shared a source with one or more § 112(c)

(6) HAPs. 76 Fed.Reg. 9457/2. Thus a subsection designed for seven HAPs that Congress thought deserved special attention — a temporal priority and a demanding stringency level — would, under petitioners’ view, require EPA to apply those special rules to a broad array of HAPs when they chanced to occur at a § 112(c)(6) source. A further curiosity of petitioners’

interpretation is that it leaves the cross-reference to (d)(4) hanging. That subsection reads, in full: With respect to pollutants for which a health threshold has been established, the Administrator may consider such threshold level, with an ample margin of safety, when establishing emission standards under this subsection. 42 U.S.C. § 7412(d)(4). Whereas the (d)(2) cross-reference provides a linguistic hook for tightening the required stringency of controls over non- § 112(c)

(6) HAPs at § 112(c)(6) sources, there is no comparable hook in (d)(4) — no language equivalent to (d)(2)’s mandate to cover “the hazardous air pollutants subject to this section.” Thus, petitioners ask us to hold that Congress used (d)(2) to upshift the required stringency for some non- § 112(c)

(6) HAPs, with no similar upshift for § 112(c)

(6) HAPs governed by (d)(4). Given that the language “standards under subsection (d)(2)” might simply reflect Congress’s intention to set the stringency level for § 112(c)

(6) HAPs in a way the architecture of the Act does not otherwise make obvious, and that petitioners’ reading has the effect of tightening the stringency of standards for non- § 112(c)

(6) HAPs from sources that happen to emit § 112(c)

(6) HAPs, the meaning of § 112(e)(6)’s “subject to standards under subsection (d)(2) or (d)(4)” is ambiguous. EPA reasonably resolves the ambiguity by reading the cross-references as simply supplying the level of stringency for § 112(c)(6) standards — either MACT under (d)(2) or “health threshold” under (d)(4). Congruently, it sees the cross-references as saying nothing about the standards governing non- § 112(c)

(6) HAPs when EPA sets out to assure that the seven § 112(c)

(6) HAPs “are subject to standards” of the requisite type. Such an

interpretation is not the only one available, as EPA itself acknowledged. See 76 Fed.Reg. at 9457 (the “language [of § 112(c)(6) ] can reasonably be read to mean standards ... for all HAP emitted by the source.”) But our duty is to accept the agency’s

interpretation if it is “based on a permissible construction of the statute.” Chevron, 467 U.S. at 843, 104 S.Ct. 2778. We further note that petitioners’ view would seriously risk undercutting the priority that Congress obviously assigned the § 112(c)

(6) HAPs. If the § 112(c)(6) cross-references triggered a duty to impose more stringent standards on non- § 112(c)

(6) HAPs at § 112(c)(6) sources, such a triggering would almost certainly precipitate pushback from the operators of such sources and slow the process of imposing MACT standards on the § 112(c)

(6) HAPs. For the reasons noted above, we find EPA’s

interpretation eminently reasonable. Does the Gold Mines Rule embrace fugitive emissions? As we observed at the outset, fugitive emissions are ones from sources such as “tailings ponds, leach fields, and waste rock piles.” 94 Fed.Reg. at 9458/1. In its response to comments EPA made clear that the rule would not address such emissions. Prior to that response, the rule could be said to have left some obscurity as to its coverage.

The regulation declares, “You are subject to this subpart if you own or operate a gold mine ore processing and production facility as defined in § 63.11651, that is an area source.” 40 C.F.R. § 63.11640(a). And § 63.11651 in turn defines such a facility as “any industrial facility engaged in the processing of gold mine ore that uses any of [a number of specified production processes].” Id. § 63.11651. These

definitions paint rather broadly. But in another section, EPA appeared to narrow the rule’s scope, saying first that “[t]his subpart applies to each new or existing affected source,” 76 Fed.

Reg. at 9480, codified at 40 C.F.R. § 63.11640(b) (emphasis added), and then defining “affected sources” as each collection of “ore pretreatment processes” at a gold mine ore processing and production facility, each collection of “carbon processes with mercury retorts” at a gold mine ore processing and production facility, each collection of “carbon processes without mercury retorts” at a gold mine ore processing and production facility, and each collection of “non-carbon concentrate processes” at a gold mine ore processing and production facility, as defined in § 63.11651. Id.

Petitioners do not contend that any of the “affected sources” listed encompasses fugitive emissions. In response to petitioners’ comments advocating the broader definition, EPA resolved any resulting ambiguity in favor of the narrower definition, making the exclusion of “fugitive emissions” from “affected sources” express rather than implicit.

It characterized affected sources more generally as consisting of “the thermal processes that occur after ore crushing, including roasting operations (i.e., ore dry grinding, ore preheating, roasting, and quenching), autoclaves, carbon kilns, electrowinning, preg tanks, mercury retorts, and furnaces,” and excluding “tailings ponds, leach fields and waste rock piles.” 76 Fed.Reg. at 9458. We review EPA’s

interpretation of its previous rules even more deferentially than we review its

interpretation of statutory ambiguity. We must give “controlling weight” to the agency’s

interpretation “unless it is plainly erroneous or inconsistent with the regulation.” Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512, 114 S.Ct. 2381, 129 L.Ed.2d 405 (1994) (internal quotation marks and citations omitted). EPA’s prose may be inelegant, even recognizing that drafting rules for mercury emissions from gold mines seems unlikely to inspire elegance. But the basic structure is plausible: gold mines are a broad concept and those who own or operate them are indeed “subject to” this subpart, as § 63.11640(

a) says. Having set out that broad term, however, EPA can logically state that the subpart applies only to affected sources, and then define such sources as including considerably less than all activities at a gold mine. Thus, even assuming that before the clarification a reader might have supposed the rule to cover fugitive emissions, its resolution of the possible linguistic confusion was not “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) (citation omitted), and is entitled to our deference.

Petitioners assert further that, assuming EPA’s rule actually excluded fugitive emissions (as we have just held), its explanations for doing so were arbitrary and capricious. These arguments are "without merit. EPA reasonably concluded that the record before it provided insufficient information about the quantity of fugitive emissions or available methods of controlling them. * * * The petition for review is therefore Denied . We do not know and need not address how the "may” in (d)(4) is to be construed.