dissenting:
Eyon Christmas sues on behalf of a class encompassing every, one of Union Pacific’s California-based dispatchers. He names five defendants. Four are local managers at a single Los Angeles railyard. The fifth is one of four major American railroads, worth about $90 billion, with operations reaching, in California alone, from the Mexican border to the San Francisco Bay. Relative to the behemoth, do its local agents have a “significant” role in this statewide case? Observable reality and common sense say no, but Christmas, and the majority, say yes.
The question matters because the Class Action Fairness Act (CAFA)—Christmas’s only path to federal jurisdiction since the parties here are only minimally diverse— deprives district courts of jurisdiction over *891so-called “local controversies.” Assuming that other requirements not in dispute here are met, the local controversy exception applies in cases where there is at least one defendant who 1) is a citizen of the state where the suit was originally filed— that is to say, who is “local”; 2) “from whom significant relief is sought” by the plaintiff class; and 3) “whose alleged conduct forms a significant basis” for the class’s claims. 28 U.S.C. § 1332(d)(4)(A)(i)(II) (emphasis added). The Ninth Circuit has long since settled that Christmas, as the party seeking remand, has the burden of pleading facts proving the exception’s applicability. Serrano v. 180 Connect, Inc., 478 F.3d 1018, 1024 (9th Cir. 2007).
“Significance,” for purposes of CAFA, is a comparative test. That is, the question is whether the relief requested against the local defendant(s) is significant relative to the total relief sought, and whether the factual allegations against the local defendants) are a significant basis for the class’s claims when compared to the overall allegations against all the defendants. See Benko v. Quality Loan Serv. Corp., 789 F.3d 1111, 1118-19 (9th Cir. 2015). The individual defendants here are citizens of California, where the case was originally filed. If the local controversy exception’s two significance tests are satisfied as well, then the district court lacked subject-matter jurisdiction, and we would have to vacate and remand with instructions to return the case to the California courts.
At the threshold, I note the importance of Congress’s decision to limit the local controversy exception to cases with a “significant” local defendant. As the Supreme Court has explained, the whole point of CAFA is to “facilitate adjudication of certain class actions in federal court.” Dart Cherokee Basin Operating Co., LLC v. Owens, — U.S. —, 135 S.Ct. 547, 554, 190 L.Ed.2d 495 (2014) (emphasis added). In keeping with that purpose, courts have consistently held that “exceptions to CAFA are to be strictly interpreted ... and that CAFA should be read with a strong preference that interstate class actions should be heard in a federal court if properly removed.” Allen v. Boeing Co., 784 F.3d 625, 632-33 (9th Cir. 2015) (internal quotation marks omitted).
So understood, vigorous judicial enforcement of the significance requirement vindicates CAFA’s purpose by ensuring that the local controversy exception does not swallow Congress’s directive that interstate class actions should be heard in federal courts. Merely naming a local defendant is easy, and if Congress had required no more than that to defeat CAFA jurisdiction, the statute would have been a practical nullity—plaintiffs wishing to avoid federal court would only have needed to identify a nominal local defendant in order to secure their favored forum. This is a case in point: Christmas has deployed exactly that tactic in a transparent attempt to escape federal jurisdiction—naming four local managers in a suit clearly targeted, for every practical purpose, at their colossus of an employer—and the majority’s decision to shrink from CAFA’s safeguards is letting him get away with it.
That approach is based on a blinkered reading of our decision in Coleman v. Estes Express Lines, in which we held that CAFA directs district courts “to loo.k only to the complaint in deciding whether” the “significant relief’ and “significant basis” requirements of the local controversy exception are satisfied. 631 F.3d 1010, 1015 (9th Cir. 2011). The complaint does not allege in haec verba that Union Pacific operates only a single railyard in the state of California. Nevertheless, that is the only logical inference to be drawn from the complaint’s pointed failure to allege that *892Union Pacific has multiple California locations. Indeed, Christmas responds to one of Union Pacific’s arguments by asking us to draw exactly that inference. See Christmas’s Reply Br. at 9 (“Union Pacific argues that Christmas could not possibly be seeking substantial relief from the local defendants because the location where they worked is only one of many Union Pacific locations in California .... Neither the complaint nor Union Pacific’s removal notice state that Union Pacific has multiple locations in California.”).
Having cut Union Pacific down to size by omitting to plead the obvious, Christmas rests his significance analysis on the claim that, in essence, the railroad can only be as big as he says it is. On those artificial terms, the local defendants do seem comparatively significant. Unlike the majority, the district judge rejected this sleight of hand, and I would affirm. Although Coleman limits us to considering Christmas’s complaint, “when we take our seats on the bench we are not struck with blindness, and forbidden to know as judges what we see as men and women.” Rusak v. Holder, 734 F.3d 894, 897 (9th Cir. 2013) (quoting Ho Ah Kow v. Nunan, 12 F. Cas. 252, 255 (C.C.D.Cal. 1879)) (internal modifications omitted).
Indeed, Coleman left a safety valve for just this situation. In holding that the plaintiff in that case sought significant relief from the local defendant, the court relied in part on the fact that there was “nothing in the complaint to suggest that” the local defendant was a “nominal defendant,” or that it had “so few assets” that the plaintiff was “not seeking significant monetary relief from it.” Nor was there anything in the complaint “to suggest ... that the injunctive relief sought [was] itself insignificant.” 631 F.3d at 1020. Put differently, to rely on the complaint alone is not to blind ourselves entirely to the underlying realities of the universe. As Coleman recognized, it would be folly to apply the local controversy exception on the basis of a local defendant who is so obviously “small change,” as a practical matter, that “any experienced lawyer or judge reading the complaint would ... know[ ] that ‘significant relief was not being sought against [it],” Id. at 1019-1020.
The local defendants here are—unambiguously—small change. This is not a case, like Coleman itself, Benko v. Quality Loan Service Corp., 789 F.3d 1111 (9th Cir. 2015), or Allen v. Boeing Co., 821 F.3d 1111 (9th Cir. 2016), in which the local defendant is an active business, and there is no indication that it either lacks significant assets, or is such a flyspeck relative to the out-of-state defendant(s) that its presence in the case is functionally irrelevant. Rather, the local parties are four individuals—local managers at a single railyard—sued alongside a company that employs 46,000 people in 23 states, by a complaint alleging (and seeking at least $5,000,000 in damages stemming from) supposedly systematic conduct that took place throughout California.
Nobody—including, I would venture, counsel for the class—genuinely believes that these individuals are “significant” in any practical sense other than to defeat federal jurisdiction. Indeed, CAFA’s amount-in-controversy requirement gives us a measure of certainty as to the stakes of this case; the notion that these men will ever pay any significant portion of a $5,000,000-plus award verges on delusional.1 Coleman does not require us to *893ignore that reality—or the reality that Union Pacific is named as a defendant because it does have pockets deep enough to satisfy such a judgment—merely because Christmas has carefully plead around it. Where Congress has directed that subject-matter jurisdiction turns on the presence or absence of a federal question, it is uncontroversial that plaintiffs may not defeat removal by artful omissions, see, e.g., JustMed, Inc. v. Byce, 600 F.3d 1118, 1124 (9th Cir. 2010), and although CAFA confers a form of diversity jurisdiction, nothing in the statute or in Coleman suggests a different rule should apply to facile attempts to avoid its application in derogation of Congress’s purposes. So based on the complaint alone, I would deny Christmas’s motion to remand.
But even if the only sensible reading of the complaint itself did not support that conclusion, I would reach it by another route—judicial notice of public records establishing that Union Pacific operated numerous California railyards during the class period. A motion to remand for lack of subject-matter jurisdiction is the functional equivalent of a motion to dismiss on the same ground under Rule 12(b)(1), and, as Christmas’s counsel conceded at oral argument, it is settled practice under that rule that courts may consider facts subject to judicial notice. See, e.g., La. Mun. Police Emps. Ret. Sys. v. Wynn, 829 F.3d 1048, 1063 (9th Cir. 2016); CopyTele, Inc. v. E Ink Holdings, Inc., 962 F.Supp.2d 1130, 1135-36 (N.D. Cal. 2013).
Coleman, on which the majority relies, did not alter the ordinary practice of considering, on motions to dismiss, certain material that is “extrinsic” in the sense of not being attached to the complaint itself, but that does not threaten to turn motions to dismiss into mini-trials on the merits. Trials, after all, are there to determine disputed facts, and judicial notice only applies to facts that can “be accurately and readily determined from sources whose accuracy cannot reasonably be questioned.” Fed. R. Evid. 201(b)(2). Put in the doctrine’s terms, I understand CAFA to allow “facial” disputes over subject-matter jurisdiction, but prohibit “factual” ones, and it appears settled within this circuit that consideration of judicially noticeable facts does not turn the former into the latter. See CopyTele, 962 F.Supp.2d at 1135-36 (collecting examples). Where appropriate, we may take judicial notice on appeal and on our own initiative. See Dockray v. Phelps Dodge Corp., 801 F.2d 1149, 1152 n.3 (9th Cir. 1986).
Records of the California Air Resources Board retrieved from that agency’s publicly accessible website, whose authenticity and accuracy cannot reasonably be questioned, indicate that during the class period Union Pacific operated sixteen separate railyards within California. See California Ar Resources Board, Maps of 32 Up and Bnsf Major and Smaller Railyards in California (2013), archived at https:// perma.cc/4XAM-YDNS. Indeed, when asked about these records at oral argument, Christmas did not dispute their existence or accuracy.
Measured against the massive scope of Union Pacific’s California operations, the individual defendants’ role in this case can only be characterized as insignificant. Every one of the class members has a claim against Union Pacific—regardless of where a dispatch worker was stationed during the class period, Union Pacific was responsible for his or her schedule, and liable for any hours-of-service violations.2 *894To use CAFA’s language, Union Pacific’s allegedly systematic conduct is the basis for the claims of every class member. The local defendants’ conduct, however, will only be at issue with respect to the limited set of class members that they actually supervised—employees at the single California railyard (out of sixteen) at which Christmas worked. As we recognized in Coleman, where “liability is premised on” conduct by the diverse defendant that is common to the whole class, “the conduct of the local [defendants] will not form a significant basis for the plaintiffs’ claims.” 631 F.3d at 1018.
By the same token, the relief sought against Union Pacific dwarfs that sought against the local defendants. The math the majority deploys to hold otherwise depends on the provably erroneous assumption that Union Pacific operates only one railyard, at which it employs the local defendants. Consider that counterfactual— the one which Christmas posits and on which the majority relies: Every violation of Cal. Labor Code § 606 would generate a $5000 penalty for the railroad, and a $1000 penalty for each of the four yard managers. Union Pacific could have no additional liability for violations outside the local defendants’ control, because in a one-yard world, there would be no dispatcher scheduling for which those managers were not responsible. Under those circumstances, the local defendants’ combined liability would indeed total 80% ($4000/$5000) of Union Pacific’s—surely a significant number. The problem with the majority’s approach, however, is that Union Pacific operates fifteen other yards, none of which involve the four local defendants.
Imagine that there was one violation of § 606 at each of Union Pacific’s sixteen railyards. The four named local defendants’ (all employees of Union Pacific) combined maximum liability for the single violation under their watch would be $4000 in civil penalties. Union Pacific’s maximum liability would be $5000 per violation, multiplied by sixteen violations, for a total of $80,000 in penalties—twenty times the total penalties for which the local defendants could be liable. The same dramatic imbalance holds true for Christmas’s request for classwide injunctive relief: an injunction against the four local defendants would affect the workings of a single rail-yard in Los Angeles, an injunction against Union Pacific would bind operations at sixteen yards stretching from Stockton to the City of Industry. Again, as Coleman pointed out, where the diverse defendant “will be liable to every member of the class,” but the local defendants “will only be liable to, at most, the particular class members with whom they interacted ..., it is obvious ... that the great bulk of any damage award is sought from the [diverse defendant] rather than from the [local ones].” 631 F.3d at 1018. Here, the monetary relief against the local defendants would amount to at most 5% of that against Union Pacific, and the scope of injunctive relief would be sixteen times smaller. Under such circumstances, the question of significance answers itself.
The illustrative example above assumes that no local defendant works at a railyard-other than Christmas’s, and that violations of § 606 are not concentrated at Christmas’s yard, but spread systematically (to borrow the complaint’s language) across Union Pacific’s California facilities. The *895majority objects to those assumptions on two grounds, one based on the distribution of defendants and one based on the distribution of violations. First, as to the distribution of defendants, the majority argues that the Doe defendants may very well be the named defendants’ counterparts at Union Pacific’s other 16 railyards spread far and wide across the state. But that would be inconsistent with the complaint’s venue allegations, which state that “DEFENDANTS either reside, maintain offices, transact business, and/or have agents in Los Angeles County.” (emphasis added). The allegation as to residence is clearly the one applicable to the individual defendants, and it beggars belief to imagine that Christmas meant to plead the possible existence of a John Doe Angeleno working as a superintendent in Stockton, Roseville, Oakland, or any other distant Union Pacific facility. Second, as to the distribution of violations, the majority argues that the complaint leaves it no more likely that violations were dispersed across Union Pacific’s railyards than that they were concentrated where Christmas worked. That point is self-defeating: If, indeed, the allegations of the complaint, combined with judicially noticeable facts about Union Pacific’s facilities, make it equally likely that the named local defendants have liability amounting to 80% or 5% of Union Pacific’s, then I fail to see how the complaint meets Christmas’s burden of proving the local controversy exception’s applicability.
The remainder of this case is controlled by settled principles of preemption. The Hours of Service Act (now codified at 49 U.S.C. § 21101-09) dates to 1907, and has long been held to preempt the field of railroad working hours regulation. See S. Ry. Co. v. R.R. Comm’n of Ind., 236 U.S. 439, 447, 35 S.Ct. 304, 59 L.Ed. 661 (1915). Although the Act has been recodified several times, most recently by the Federal Railroad Safety Act (FRSA), its preemptive scope has not changed. In Kurns v. R.R. Friction Products Corp., the Supreme Court explained that the FRSA’s recodification of an existing railroad safety law—the Locomotive Inspection Act—did not change its preemptive effect. 565 U.S. 625, 633, 132 S.Ct. 1261, — L.Ed.2d — (2012) (citing Marshall v. Burlington Northern, Inc., 720 F.2d 1149, 1152-53 (9th Cir. 1983) (Kennedy, J.) (addressing the preemptive force of the Boiler Inspection Act)). Like the Locomotive Inspection Act, the Hours of Service Act “was already in effect when the FRSA was enacted,” and as such, its “[field-preemptive] scope remains unaltered.” Id.
The district court properly retained jurisdiction, and correctly dismissed Christmas’s preempted claims on the merits. I respectfully dissent.
. Although Christmas’s complaint—filed in a state court that did not require him to plead an amount in controversy—does not seek a particular amount of money, he does not challenge the defendants' analysis, in the notice of removal filed with the district court, showing *893that his complaint clearly satisfies CAFA's $5,000,000 jurisdictional minimum.
. At oral argument, Christmas's counsel gave a narrower description of the class, asserting *894that it only includes workers involved, as Christmas himself was, in “intermodal” operations, which only take place at a certain subset of railyards. Paragraph 19 of the amended complaint, however, clearly states that the class includes “[all] persons ... who dispatch, report, transmit, receive or deliver orders pertaining to or affecting train movements.”