Petitioners are twelve states, three cities, an American territory, and numerous environmental organizations. They are opposed by the Environmental Protection Agency as respondent, and ten states and several trade associations as intervenors. The controversy is about EPA’s denial of a petition asking it to regulate carbon dioxide (CO2) and other greenhouse gas emissions from new motor vehicles under § 202(a)(1) of the Clean Air Act, 42 U.S.C. § 7521(a)(1). EPA concluded that it did not have statutory authority to regulate greenhouse gas emissions from motor vehicles and that, even if it did, it would not exercise the authority at this time. 68 Fed.Reg. 52,922 (Sept. 8, 2003).
I.
We should say a few words about our jurisdiction under the Clean Air Act to review an EPA denial of a petition for rulemaking. Section 307(b)(1), 42 U.S.C. § 7607(b)(1), gives this court exclusive jurisdiction over “nationally applicable regulations promulgated, or final action taken, by the Administrator” under chapter 85 of the Act. The district courts, on the other hand, have jurisdiction over citizen suits to compel EPA to perform nondiscretionary acts or duties. 42 U.S.C. § 7604(a)(2); see Sierra Club v. Thomas, 828 F.2d 783, 787-92 (D.C.Cir.1987). Because EPA refused to promulgate “nationally applicable regulations” after being asked to do so, we have jurisdiction only if EPA thereby engaged in “final action.” We can be sure that its denial of the rulemaking petition was “final.” But did this constitute agency “action”? To answer that question we must consult the Administrative Procedure Act — specifically 5 U.S.C. § 551(13). The term “action” in § 307(b)(1) of the Clean Air Act, like the term “final,” carries its traditional meaning in administrative law. See Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 478, 121 S.Ct. 903, 149 L.Ed.2d 1 (2001); Indep. Equip. Dealers Ass’n v. EPA, 372 F.3d 420, 428 (D.C.Cir.2004); Sierra Club v. Gorsuch, 715 F.2d 653, 656-57 (D.C.Cir.1983). Section 551(13) of the APA defines “agency action” as “the whole or a part of an agency rule, order, license, sanction, relief, or the equivalent or denial thereof, or failure to act” (italics added). While § 307 of the Clean Air Act makes several APA provisions inapplicable— namely, 5 U.S.C. §§ 553-557 & 706 — APA § 551 is not among them. EPA’s denial of the rulemaking petition was therefore “fi*286nal action,” and since the petition sought regulations national in scope, § 307(b)(1) confers jurisdiction on this court to hear these consolidated cases.
Another, related, point needs to be mentioned. Several of the petitions for judicial review treated a memorandum of EPA’s General Counsel, Robert Fabricant, as “final action taken, by the Administrator” under § 307(b)(1). The memorandum, dated August 28, 2003, and addressed to the EPA Administrator, was entitled “EPA’s Authority to Impose Mandatory Controls to Address Global Climate Change under the Clean Air Act.” The General Counsel, after analyzing § 202(a)(1) of the Clean Air Act, and other legislative and executive actions, stated his belief that the Act “does not authorize regulation to address global climate change.” He therefore withdrew a contrary memorandum issued in 1998 by one of his predecessors.
The Fabricant memorandum, consisting of legal advice to the EPA Administrator, did not in itself constitute “final action” of the Administrator. To be sure, the Administrator adopted the “General Counsel’s opinion” and relied on its analysis as one of the alternative grounds for rejecting the rulemaking petition. See 68 Fed.Reg. at 52,925. The Administrator’s explanation incorporated many of the memorandum’s passages verbatim, rephrased and reordered others, and expanded on the General Counsel’s reasoning. Still, it is the Administrator’s denial of the rulemaking petition, with the accompanying explanation, that represents the “final action” of the Administrator subject to judicial review under § 307(b)(1). The significance of the General Counsel’s opinion, as set forth in his memorandum, is the Administrator’s reliance on his reasoning in deciding the matter now before us.
There is an additional jurisdictional issue presented, but not under the Clean Air Act. EPA claims that petitioners lack standing under Article III of the Constitution. Standing exists only if the complainant has suffered an injury in fact, fairly traceable to the challenged action, and likely to be redressed by a favorable decision. See Lujan v. Defenders of Wildlife, 504 U.S. 555, 560, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992). EPA’s argument is that petitioners have not “adequately demonstrated” two elements of standing: that their alleged injuries were “caused by EPA’s decision not to regulate emissions of greenhouse gases from mobile sources”; and that them injuries “can be redressed by a decision in their favor” by this court. Brief for Respondent at 16.
In anticipation of this argument, petitioners filed two volumes of declarations with the court, some containing lengthy exhibits. The declarations, from scientists, engineers, state officials, homeowners, users of the nation’s recreational resources, and other individuals, predict catastrophic consequences from global warming caused by greenhouse gases, including loss of or damage to state and private property, frequent intense storm surge floods, and increased health care costs. Brief for Petitioners at 2-4.
For the causation and redressability aspects of standing, petitioners cite two of their declarations. One, from a climatologist, states that reductions in C02 and other greenhouse gases from vehicles in the United States would alone have a meaningful impact and would “delay and moderate many of the adverse impacts of global warming.” He adds .that if EPA took action to reduce such emissions, other countries would likely follow suit. The climatologist bases his predictions about future climate change on climate models and on “quantitative scenarios generated *287by the IPCC” — the Intergovernmental Panel on Climate Change, established in 1988 by the United Nations and the World Meteorological Organization. The other declaration is from a mechanical engineer. He states that, on the basis of his experience with controlling other pollutants, there is “no doubt that establishing emissions standards for pollutants that contribute to global warming would lead to investment in developing improved technologies to reduce those emissions from motor vehicles, and that successful technologies would gradually be mandated by other countries around the world.”
We have held that, to establish standing, a petitioner challenging agency action has the same burden of production as “a plaintiff moving for summary judgment in the district court: it must support each element of its claim to standing ‘by affidavit or other evidence.’ ” Sierra Club v. EPA 292 F.3d 895, 899 (D.C.Cir.2002) (quoting Lujan, 504 U.S. at 561, 112 S.Ct. 2130). Petitioners’ declarations do “support each element” of standing. But supporting an allegation is one thing; proving an allegation is quite another. Lujan holds that when a plaintiffs standing is challenged in a motion for summary judgment, the plaintiff “must ‘set forth’ by affidavit or other evidence ‘specific facts,’ Fed. Rule Civ. Proc. 56(e), which for purposes of the summary judgment motion will be taken as true.” 504 U.S. at 561, 112 S.Ct. 2130. If we were to analogize the situation here to one in which EPA filed such a summary judgment motion, we would conclude that petitioners had submitted enough evidence raising genuine issues of material fact to defeat the motion. See Fed. R. Civ. P. 56(c). But Lujan goes on to hold that at “the final stage” the evidence plaintiff presented at summary judgment “(if controverted) must be ‘supported adequately by the evidence adduced at trial.’ ” 504 U.S. at 561, 112 S.Ct. 2130 (quoting Gladstone, Realtors v. Village of Bellwood, 441 U.S. 91, 115 n. 31, 99 S.Ct. 1601, 60 L.Ed.2d 66 (1979)). One might say that in this case we are at the “final stage.” But the analogy is not entirely apt. As an appellate' court we do not conduct evidentiary hearings in order to make findings of fact. This is why, when Sierra Club spoke of “other evidence” relating to standing, the court had in mind evidence presented to the agency. 292 F.3d at 899. Here, the administrative record contains a wealth of such “other evidence,” and some of it contradicts petitioners’ claim that greenhouse gas emissions from new motor vehicles have caused or will cause a significant change in the global climate. That is partly why EPA decided not to regulate at this time.
Steel Co. v. Citizens for a Better Environment, 523 U.S. 83, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998), instructs federal courts to resolve Article III standing questions before proceeding to the merits of a case. The combination of Lujan, Steel Co., and the factual overlap of the standing issues with EPA’s justifications for not regulating greenhouse gases present us with three options. The first is to refer the standing issues to a special master for a factual determination. This would be, as one commentator has suggested, “folly.” 13A Charles A. Wright et al., Federal Practice and Procedure 2d § 3531.15, at 101 (1984). Such a proceeding would largely duplicate the proceedings on the rulemaking petition and to no good end. Another option would be to remand to EPA for a factual determination of causation and redressability. That too would make no sense. For one thing, judgments about standing are the responsibility of the federal courts. For another, EPA has already reached a decision about the state of the evidence regarding global warming from greenhouse gases. The third option is *288to proceed to the merits with respect to EPA’s alternative decision not to regulate on the grounds, among others, that the effect of greenhouse gases on climate is unclear and that models used to predict climate change might not be accurate.
We have decided to follow the third course. Steel Co. endorses this approach with respect to questions of statutory-standing. The Court explained that “the merits inquiry and the statutory standing inquiry often overlap” and “are sometimes identical, so that it would be exceedingly artificial to draw a distinction between the two.” 528 U.S. at 97 n. 2, 118 S.Ct. 1003. The Court’s distinction of Article III standing cases rested on the premise that there would be no such overlap and that the issue of Article III standing would be entirely separate from the merits. Id. The Court did not say what the proper order of decision should be when, as in this case, that premise does not hold. In this highly unusual circumstance — encountered for the first time in this court — we will follow the statutory standing cases. We will therefore assume arguendo that EPA has statutory authority to regulate greenhouse gases from new motor vehicles.1 The question we address is whether EPA properly declined to exercise that authority.
II.
Greenhouse gases trap energy, much like the glass panels of a greenhouse. The earth’s surface is warmed by absorbing solar energy (visible light). The earth, in turn, radiates infrared energy (heat) back into space. A portion of the infrared radiation is trapped by greenhouse gas molecules,’ resulting in additional warming of the lower atmosphere and the earth’s surface. This “greenhouse effect” is a natural phenomenon, without which the planet would be significantly colder and life as we know it would not be possible. EPA, Global Warming — Climate, at http://yosemite.epa.gov/oar/globalwarming.nsPcontenVclimateJhtml.
Petitioners sought to have EPA regulate, under § 202(a)(1) of the Clean Air Act, carbon dioxide (CO2), and three other greenhouse gases: methane (CH4), nitrous oxide (N2O), and hydrofluorocarbons (HFCs).2 In response to EPA’s request for public comments on the 1999 petition for rulemaking, the agency received nearly 50,000 submissions. 68 Fed.Reg. at 52,-924. Most were short expressions of support for the petition; many were nearly identical. Id. The comment period closed in May 2001. In the same month, the White House requested the National Academy of Sciences to assist the Administration in its review of climate change policy. The Academy “is a private, nonprofit, self-perpetuating society of distinguished scholars engaged in scientific and engineering research .... ” National Research Council, Climate Change Science: An Analysis of Some of the Key Questions, preface (2001). Under its congressional charter, issued in 1863, the Academy has a mandate to advise the federal government on scientific and technical matters when requested. The Academy’s principal oper*289ating agency for providing such advice is its National Research Council. Id.
In denying the rulemaking petition, EPA found that the scientific comments petitioners and others submitted rested on information already in the public domain and did not add significantly to the body of knowledge available to the National Research Council when it prepared the report cited above. Since none of the comments caused EPA to question the Council’s report, EPA decided to rely on the Council’s “objective and independent assessment of the relevant science.” 68 Fed.Reg. at 52,930.
The National Research Council concluded that “a causal linkage” between greenhouse gas emissions and global warming “cannot be unequivocally established.” National Research Council, Climate Change Science, at 17. The earth regularly experiences climate cycles of global cooling, such as an ice age, followed by periods of global warming. Id. at 7. Global temperatures have risen since the industrial revolution, as have atmospheric levels of carbon dioxide. But an increase in carbon dioxide levels is not always accompanied by a corresponding rise in global temperatures. For example, although carbon dioxide levels increased steadily during the twentieth century, global temperatures decreased between 1946 and 1975. Id. at 16. Considering this and other data, the National Research Council concluded that “there is considerable uncertainty in current understanding of how the climate system varies naturally and reacts to emissions of greenhouse gases.” Id. at 1. This uncertainty is compounded by the possibility for error inherent in the assumptions necessary to predict future climate change.3 And, as the National Research Council noted, past assumptions about effects of future greenhouse gas emissions have proven to be erroneously high. Id. at 19.
Relying on Ethyl Corp. v. EPA, 541 F.2d 1 (D.C.Cir.1976) (en banc), petitioners challenge EPA’s decision to forego rule-making “[ujntil more is understood about the causes, extent and significance of climate change and the potential options for addressing it.” 68 Fed.Reg. at 52,931. In our view Ethyl supports EPA, not petitioners. Section 202(a)(1) directs the Administrator to regulate emissions that “in his judgment” “may reasonably be anticipated to endanger public health or welfare.” Section 202(a)(1) was not at issue in Ethyl; the court mentioned an earlier version of that provision, in a footnote, only by way of analogy. 541 F.2d at 20 n. 37. But what the court had to say about § 202(a)(1) is instructive. In requiring the EPA Admin*290istrator to make a threshold “judgment” about whether to regulate, § 202(a)(1) gives the Administrator considerable discretion. Id. Congress does not require the Administrator to exercise his discretion solely on the basis of his assessment of scientific evidence. Id. at 20. What the Ethyl court called “policy judgments” also may be taken into account. By this the court meant the sort of policy judgments Congress makes when it decides whether to enact legislation regulating a particular area. Id. at 26.
The EPA Administrator’s analysis, although it did not mention Ethyl, is entirely consistent with the case. In addition to the scientific uncertainty about the causal effects of greenhouse gases on the future climate of the earth, the Administrator relied upon many “policy” considerations that, in his judgment, warranted regulatory forbearance at this time. 68 Fed.Reg. at 52,929. New motor vehicles are but one of many sources of greenhouse gas emissions; promulgating regulations under § 202 would “result in an inefficient, piecemeal approach to the climate change issue.” 68 Fed.Reg. at 52,931. The Administrator expressed concern that unilateral regulation of U.S. motor vehicle emissions could weaken efforts to persuade developing countries to reduce the intensity of greenhouse gases thrown off by their economies. Id. Ongoing research into scientific uncertainties and the Administration’s programs to address climate change — including voluntary emission reduction programs and initiatives with private entities to develop new technology — also played a role in the Administrator’s decision not to reguláte. 68 Fed.Reg. at 52,931-33. The Administrator pointed to efforts to promote “fuel cell and hybrid vehicles” and ongoing efforts to develop “hydrogen as a primary fuel for cars and trucks.” 68 Fed.Reg. at 52,931. The Administrator also addressed the matter of remedies. Petitioners offered two ways to reduce C02 from new motor vehicles: reduce gasoline consumption and improve tire performance: As to the first, the Department of Transportation — the agency in charge of fuel efficiency standards — recently issued new standards requiring greater fuel economy, as a result of which millions of metric tons of C02 will never reach the stratosphere. Id. As to tire efficiency, EPA doubted its authority to regulate this subject as an “emission” of an air pollutant. Id. “With respect to the other [greenhouse gases]— CH4, N20, and HFCs — petitioners make no suggestion as to how those emissions might be reduced from motor vehicles.” Id.
It is therefore not accurate to say, as petitioners do, that the EPA Administrator’s refusal to regulate rested entirely on scientific uncertainty, or that EPA’s decision represented an “open-ended invocation of scientific uncertainty to justify refusing to regulate,” Brief for Petitioners at 51. A “determination of endangerment to public health,” the court said in Ethyl, “is necessarily a question of policy that is to be based on an assessment of risks and that should not be bound by either the procedural or the substantive rigor proper for questions of fact.” Ethyl, 541 F.2d at 24. And as we have held,' a reviewing court “will uphold agency conclusions based on policy judgments” “when an agency must resolve issues ‘on the frontiers of scientific knowledge.’ ” Envtl. Def. Fund v. EPA, 598 F.2d 62, 82 (D.C.Cir. 1978).
We thus hold that the EPA Administrator properly exercised his discretion under § 202(a)(1) in denying the petition for rule-making. The petitions for review in Nos. 03-1365, 03-1366, 03-1367, and 03-1368 are dismissed, and the petitions for review *291in Nos. 03-1361, 03-1362, 03-1363, and 03-1364 are denied.
So ordered.
. Relying on FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 120 S.Ct. 1291, 146 L.Ed.2d 121 (2000), EPA concluded that in light of the enormous economic and political consequences of regulating greenhouse gas emissions, Congress would have been far more specific if it had intended to authorize EPA to regulate the subject under § 202(a)(1) of the Clean Air Act. 58 Fed.Reg. at 52,928. We express no view on the validity of EPA's analysis.
. The rulemaking request and the papers submitted to this court focus on the effects of CO2.
. "As the NRC explained, predicting future climate change necessarily involves a complex web of economic and physical factors including: Our ability to predict future global anthropogenic emissions of GHGs and aerosols; the fate of these emissions once they enter the atmosphere (e.g., what percentage are absorbed by vegetation or are taken up by the oceans); the impact of those emissions that remain in the atmosphere on the radiative properties of the atmosphere; changes in critically important climate feedbacks (e.g., changes in cloud cover and ocean circulation); changes in temperature characteristics (e.g., average temperatures, shifts in daytime and evening temperatures); changes in other climatic parameters (e.g., shifts in precipitation, storms); and ultimately the impact of such changes on human health and welfare (e.g., increases or decreases in agricultural productivity, human health impacts). The NRC noted, in particular, that ‘[t]he understanding of the relationships between weather/climate and human health is in its infancy and therefore the health consequences of climate change are poorly understood' (p. 20). Substantial scientific uncertainties limit our ability to assess each of these factors and to separate out those changes resulting from natural variability from those that are directly the result of increases in anthropogenic GHGs.” 68 Fed.Reg. at 52,930.