KANNE, Circuit Judge.
In 2009, the United States and the State of Wisconsin ("the Governments") filed suit in federal district court against eleven of the potentially responsible parties ("PRPs") in an environmental cleanup, seeking response costs under the Comprehensive Environmental Response, Compensation, and Liability Act, 42 U.S.C. § 9601 et seq. (CERCLA). Shortly thereafter, the Governments filed notice of a de minimis consent decree pursuant to CERCLA § 122(g). Eventually, the Governments moved for settlement. Appleton Papers Inc. and NCR Corporation intervened. The district court granted the settlement motion over the intervenors' opposition. Later, the Governments moved for a de minimis settlement with a twelfth defendant, and the district court granted this motion. Appleton and NCR appealed the grant of both settlement motions. We affirm.
I. BACKGROUND
The Fox River in Wisconsin is heavily contaminated with Polychlorinated biphenyls ("PCBs"). The Governments contend Appleton and NCR are responsible for much of these PCBs. Their predecessors, according to the Governments, contributed significant amounts of Aroclor 1242, the most prevalent PCB in Fox River. The river also contains other PCBs, including Aroclor 1254 and Aroclor 1260. Appleton, NCR, and a few other PRPs are currently paying to clean up Fox River in compliance with a 2007 Environmental Protection Agency order.
Appleton and NCR are seeking contribution, in a separate suit, from many other PRPs. These include the twelve PRPs subject to the consent decrees in this suit: Neenah Foundry Company; Green Bay Metropolitan Sewerage District; the City of De Pere; Procter & Gamble Paper Products Company; Union Pacific Railroad Company; Green Bay Packaging, Incorporated; Heart of the Valley Metropolitan Sewerage District; Lafarge Corporation; Leicht Transfer and Storage Company; Wisconsin Public Service Corporation; International Paper Company; and George A. Whiting Paper Company (the "de minimis defendants").
In 2009, the Governments filed suit against the de minimis defendants under CERCLA §§ 106 and 107. See 42 U.S.C. §§ 9606, 9607. The Governments then filed two separate consent decrees—one for the City of De Pere, the other for the remaining de minimis defendants. De Pere agreed to pay $210,000 to satisfy its liability. The others agreed to pay a combined total of $1,875,000.
The Governments estimated that the total cleanup cost would be, accounting for uncertainty, $1.5 billion. They also estimated that each of the de minimis defendants had discharged no more than 100 kilograms of PCBs and that, in total, 230,000 kilograms of PCBs had been discharged into the Fox River. The total discharge estimate is a conservative one, based on a low-end estimate of Aroclor 1242, not total PCBs. The Governments based their individual contribution estimates on the de minimis defendants' responses to CERCLA § 104(e) information requests ("§ 104(e) requests"),1 on discovery responses from Appleton and NCR's contribution suit, and on statements by the de minimis defendants certifying that they had turned over all information related to their use of PCBs.
The Department of Justice filed notice of the consent decrees in the Federal Register and solicited public comment. Appleton and NCR objected to both settlements, arguing that the settlements underestimated the de minimis defendants' contributions. Appleton and NCR based their objections on studies suggesting that the Governments had underestimated the amount of Aroclor 1254 and 1260 in the Fox River. These studies also provided direct evidence of PCB use by some de minimis defendants. The Governments agreed with NCR on one point: they recognized that Green Bay Metro Sewerage may have discharged more than 100 kilograms of PCBs. Appleton and NCR claimed Green Bay Metro Sewerage had discharged up to 324 kilograms. The Governments made an even more cautious estimate of 480 kilograms, and Green Bay Metro agreed to a corresponding payment of $325,000. The remaining de minimis defendants' payments remained the same.
After the notice and comment process, the Governments moved for settlement in the district court. Appleton and NCR intervened and opposed the motions for settlement. The district court approved both settlement decrees and granted the motions for settlement. Appleton and NCR then appealed.
II. ANALYSIS
In reviewing the consent decrees, we are constrained by a double dose of deference. See United States v. Cannons Eng'g Corp., 899 F.2d 79, 84 (1st Cir.1990). First, the trial court must defer to the expertise of the agency and to the federal policy encouraging settlement. In re Tutu Water Wells CERCLA Litigation, 326 F.3d 201, 207 (3d Cir.2003). Thus, the district court must approve a consent decree if it is reasonable, consistent with CERCLA's goals, and substantively and procedurally fair. Id. We, in turn, defer to the district court's decision—reviewing only for an abuse of discretion. Cannons Eng'g, 899 F.2d at 84.
A. No Rational Basis
The district court concluded the consent decrees were substantively fair. Appleton and NCR argue that this conclusion has no rational basis in the record. A consent decree is substantively fair if its terms are based on comparative fault. Tutu Wells, 326 F.3d at 207; Cannons Eng'g, 899 F.2d at 87. The calculation of comparative fault "should be upheld unless it is arbitrary, capricious, and devoid of a rational basis." Cannons Eng'g, 899 F.2d at 87 ("[W]hat constitutes the best measure of comparative fault . . . should be left largely to the EPA's expertise.").
Rarely does an appellate court conclude the district court had no factual basis to approve a consent decree. Appleton and NCR can point to only one such holding. In United States v. Montrose Chem. Corp., the Ninth Circuit reversed the approval of a consent decree because the record included no information—not even an unsupported estimate—about the total cost of cleanup or the settling parties' comparative fault. 50 F.3d 741, 747 (9th Cir.1995). We need not decide whether an unsupported estimate would be a sufficient factual basis to affirm a consent decree—the Governments' estimate here has adequate support in the record.
According to Appleton and NCR, the only bases for the de minimis defendants' comparative fault are the Governments' unsupported conclusions. In reality, the record includes information about each of the de minimis defendants' discharges of PCBs. Whiting Paper, Green Bay Metro Sewerage, Green Bay Packaging, Heart of the Valley, International Paper, Procter & Gamble, and Union Pacific2 all responded to § 104(e) requests. The de minimis defendants that did not respond to § 104(e) requests provided certified statements about their use of PCBs and about any potential discharges. The record also included deposition transcripts and written discovery responses produced in related litigation. Finally, the record included information drawn from the public comment process.
Contrary to Appleton and NCR's argument, these sources are not devoid of content. In fact, Appleton and NCR used the information from Green Bay Metro Sewerage's § 104(e) responses to demonstrate that Green Bay Metro Sewerage's discharges exceeded the Governments' original estimate. Given the amount of relevant information in the record, we find that the record provides a rational basis on which the district court could conclude the consent decrees were substantively fair.
B. Consideration of non-1242 Aroclors
Appleton and NCR next argue that the consent decrees are not substantively fair because the estimates of the de minimis defendants' comparative fault do not account for non-1242 Aroclors. This argument rests on a false premise. In truth, the estimates of the individual de minimis defendants' comparative fault account for discharges of all PCBs—not just Aroclor 1242. According to Appleton and NCR, the Governments relied on surveys of PCB pollution—which focus on Aroclor 1242—to derive individual estimates. But the Governments actually relied on § 104(e) responses and other direct information about the de minimis defendants' discharges. These sources cover discharges of all PCBs.
The Governments did consider only Aroclor 1242 in their estimate of the total amount of PCBs discharged into the Fox River. But Appleton and NCR wisely do not contest this choice: including non-1242 Aroclors in this estimate would have only decreased the de minimis defendants' comparative fault.
Appleton and NCR's only argument, then, is that the evidence it has presented about the presence of non-1242 Aroclors shows that the consent decrees lack a rational basis. We reiterate that a district court should defer to the Governments' expertise in weighing ambiguous and conflicting evidence of substantive fairness. Cannons Eng'g, 899 F.2d at 88. And we will only disturb the district court's decision if Appleton and NCR can show that the court ignored a material factor or made "a serious mistake in weighing" the relevant factors. Id. at 84.
Appleton and NCR have not met this heavy burden. They point to studies suggesting that Aroclors 1254 and 1260 are more toxic than Aroclor 1242. This, they argue, shows the consent decrees lack rational basis because the non-1242 Aroclors do not weigh more heavily in the decrees' comparative fault calculations. But the Governments point to evidence suggesting that Aroclor 1242 is just as toxic as Aroclors 1254 and 1260. The district court considered all the relevant evidence and decided the Governments' approach was rational. We are poorly suited to evaluate the merits of the conflicting positions. See Kalamazoo River Study Grp. v. Rockwell Int'l Corp., 274 F.3d 1043, 1051 (6th Cir. 2001) (upholding district court's decision to defer to the EPA's conclusion that Aroclors 1242 and 1254 are equally toxic). The district court did not abuse its discretion by deeming the Governments' toxicity calculations reasonable.
Appleton and NCR also point to evidence suggesting the Governments have underestimated the amount of non-1242 Aroclors discharged into Fox River. Even if we were to disregard the Governments' evidence of the amount of non-1242 Aroclors in Fox River, Appleton and NCR's evidence would not demonstrate a lack of substantive fairness. Appleton and NCR cannot show that the de minimis defendants—rather than any of the numerous other PRPs—are responsible for the purportedly uncounted non-1242 Aroclors.
C. Unresolved Issue of Divisibility
We note that Appleton and NCR appeal only the district court's approval of the consent decrees at issue here. Any divisibility decision made—or not yet made—in related litigation is beyond the scope of this appeal. The only relevant issue, then, is whether the district court abused its discretion by affirming the consent decrees before deciding whether the de minimis defendants' liability was divisible from that of other PRPs.
By its nature, a consent decree eliminates many possible outcomes that would have been better for one side or the other. See United States v. Armour & Co., 402 U.S. 673, 681, 91 S.Ct. 1752, 29 L.Ed.2d 256 (1971) ("[I]n exchange for the saving of cost and elimination of risk, the parties each give up something they might have won had they proceeded with the litigation."). Appleton and NCR have not shown that the settlement amounts do not account for the risk of divisibility. Under Appleton and NCR's theory, parties could never negotiate away the risk of which side would prevail in a divisibility dispute, even when—as here—the cost of resolving that dispute might exceed the total settlement amount.
Neither the Governments nor the de minimis defendants have an interest in disputing divisibility here. Appleton and NCR are free to dispute the divisibility of their own liability in the appropriate suit. Accordingly, the district court did not abuse its discretion by approving the consent decrees before a divisibility determination.
D. Insufficient Discovery
Appleton and NCR claim more discovery was needed before the district court could approve the settlement agreement. They do not make clear when the discovery should have taken place or who should have been involved. Appleton and NCR did not move for discovery in this case, so they cannot appeal the denial of any such motion. To the extent they argue that the Governments and the de minimis defendants should have engaged in discovery in order to establish a factual basis for the consent decree, we have already addressed their concerns. To the extent Appleton and NCR challenge discovery limitations in separate litigation, their argument is beyond the scope of this appeal.
E. Improper Consideration of Equitable Factors
Appleton and NCR argue that the district court, in its approval of the consent decrees, considered equitable factors in violation of CERCLA § 122(g). By not making this argument until oral argument, they forfeited it. See Ceta v. Mukasey, 535 F.3d 639, 649 n. 16 (7th Cir.2008).
Even if not forfeited, Appleton and NCR's argument has no merit. Section 122(g) puts forth criteria for identifying de minimis defendants. It does not limit the factors a district court can consider in determining whether a settlement decree is fair, reasonable, and consistent with CERCLA. See 42 U.S.C. § 9622(g). Moreover, while the district court's order does mention comparative liability, an equitable factor, its decision rests on its appraisal of comparative fault—an appropriate and necessary factor for consideration. See Tutu Water Wells, 326 F.3d at 207; Cannons Eng'g, 899 F.2d at 87.
III. CONCLUSION
Because the district court did not abuse its discretion in approving the consent decrees between the Governments and the de minimis defendants, we AFFIRM its decisions.
