Emhart Industries v. CNA Holdings LLC

Opinion

United States Court of Appeals For the First Circuit

No. 19-1563

EMHART INDUSTRIES, INC.,

Plaintiff/Third Party Plaintiff, Appellee,

STATE OF RHODE ISLAND,

by and through the Rhode Island Department of Environmental Management,

Plaintiff, Appellee,

v.

UNITED STATES DEPARTMENT OF THE AIR FORCE, et al.,* Defendants/Third Party Plaintiffs, Appellees, BLACK & DECKER INC.,

Third Party Plaintiff/Third Party Defendant, Appellee, CNA HOLDINGS LLC, f/k/a CNA HOLDINGS, INC, et al., Third Party Defendants, Appellants.

APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF RHODE ISLAND

[Hon. William E. Smith, U.S. District Judge]

* Pursuant to Fed. R. App. P. 43(c)(2), Acting Secretary of the United States Department of the Air Force John P. Roth has been substituted for former Secretary Barbara M. Barrett; Acting Secretary of the United States Department of the Navy Thomas W. Harker has been substituted for former Secretary Kenneth Braithwaite; and Secretary of the United States Department of Defense Lloyd J. Austin III has been substituted for former Acting Secretary David L. Norquist.

Before

Barron, Circuit Judge,

and Saris, District Judge.

Bryan Killian, with whom Duke K. McCall, III, Douglas A.

Hastings, Morgan, Lewis & Bockius LLP, Dan Vineyard, Jennifer Caughey, and Jackson Walker LLP, were on brief, for appellants CNA Holdings LLC, et al.

Joan M. Pepin, Attorney, Environment and Natural Resources Division, United States Department of Justice, with whom Jeffrey Bossert Clark, Assistant Attorney General, Eric Grant, Deputy Assistant Attorney General, Michael T. Gray, Attorney, Jerome W. MacLaughlin, Attorney, Phillip R. Dupré, Attorney, Susan Forcier, Deputy Chief Legal Counsel, Rhode Island Department of Environmental Management, Joy Sun, Eve S. Vaudo, EPA Region 1 Office of General Counsel, Michael L. Casillo, Litigation Attorney, Air Force Legal Operations Agency, and Genifer M. Tarkowski, Attorney, Naval Litigation Office, were on brief, for appellees United States, et al.

Joseph W. Hovermill, with whom Joseph L. Beavers, Alexander P. Creticos, Miles & Stockbridge P.C., Christopher A. Duggan, H. Reed Witherby, and Smith Duggan Buell & Rufo LLP, were on brief, for appellees Emhart Industries, Inc., et al.

February 17, 2021

 Judge Torruella heard oral argument in this matter and participated in the semble, but he did not participate in the issuance of the panel's opinion in this case. The remaining two panelists therefore issued the opinion pursuant to 28 U.S.C. § 46(d).

 Of the District of Massachusetts, sitting by designation.

BARRON, Circuit Judge. This is an appeal by three companies -- CNA Holdings LLC, Exxon Mobil Corporation, and Union Oil Company of California -- that seek to vacate a consent decree ("the Decree") to which they were not parties but that had been entered into by the U.S. Department of Defense, the U.S. Department of the Air Force, and the U.S. Department of the Navy ("the federal agencies"); Emhart Industries; the U.S. Environmental Protection Agency ("EPA"); and the State of Rhode Island. The Decree settled claims involving those parties under the Comprehensive Environmental Response, Compensation, and Liability Act ("CERCLA") and Rhode Island law regarding the responsibility for, and the allocation of the costs of, the cleanup of a contaminated Superfund site located in North Providence, Rhode Island ("the Site"). But, the Decree also purported to do something of direct import for the appellants: bar their own CERCLA claims against Emhart and the federal agencies pertaining to the allocation of the costs of cleaning up the Site. In seeking to overturn the District Court's approval of the Decree, the appellants contend that it was improper as a matter of law and that, in any event, the District Court abused its discretion in approving it, because it failed meaningfully to review it before doing so. We disagree and thus affirm the District Court's ruling approving the Decree.

4

I.

A.

Beginning in the 1940s,1 Metro Atlantic (the corporate predecessor to Emhart Industries)2 manufactured textile chemicals on nine acres on a peninsula in North Providence, Rhode Island ("the Source Area"). Emhart Indus., Inc. v. New Eng. Container Co. (Phase I), 130 F. Supp. 3d 534, 538, 541, 542 n.18 (D.R.I. 2015). During some of those years, the company produced hexachlorophene ("HCP") there, id. at 542, and, in the process of manufacturing it, released 2,3,7,8-tetrachlorodibenzo-p-dioxin ("2,3,7,8-TCDD") into the ground in the Source Area and the nearby Woonasquatucket River, id. at 540.

New England Container Company ("NECC") operated a business reconditioning 55-gallon drums on a portion of the Source Area beginning around 1952. Id. at 542, 547. Various entities, including Metro Atlantic and the Department of Defense, sent drums to NECC to be reconditioned. Id. at 547. The drums often contained

1 Our recitation of the facts is drawn from the District Court's findings of fact and conclusions of law from the first two phases of the bench trial below. See Emhart Indus., Inc. v. New Eng. Container Co. (Phase II), 274 F. Supp. 3d 30 (D.R.I. 2017); Emhart Indus., Inc. v. New Eng. Container Co. (Phase I), 130 F. Supp. 3d 534 (D.R.I. 2015).

2At the time, Metro Atlantic was called the Atlantic Chemical Company. Since the initiation of this litigation, Black & Decker, Inc., Emhart's successor-in-interest, has been added as a party. We will refer to "Emhart" only for simplicity.

5

residues of the chemicals that they had carried prior to their refurbishment. Id.

In 1996, the EPA discovered fish contaminated with dioxin in the Woonasquatucket River. Id. at 541. Of all the contaminants subsequently discovered at the Site, 2,3,7,8-TCDD was not only the most toxic dioxin but also one of the most toxic substances of any kind. Id. at 540 n.11.

In investigating the Site, the EPA identified the Source Area as the epicenter of the contamination. Id. at 541-42. In 1999 and 2000, the EPA issued notices of potential liability for that contamination to NECC and Emhart, respectively, as potentially responsible parties under section 107(a) of CERCLA. In 2000, moreover, the agency placed the Site, which consisted of a three-mile stretch of the Woonasquatucket River and the surrounding area, on its National Priorities List ("NPL") for cleanup under CERCLA. Id. at 541.

B.

CERCLA "grants the President broad power to command government agencies and private parties to clean up hazardous waste sites." Key Tronic Corp. v. United States, 511 U.S. 809, 814 (1994). The statute provides that when there is a "release or substantial threat of release" of "any hazardous substance," or "any pollutant or contaminant" that "may present an imminent and

6

substantial danger to the public health or welfare," the EPA3 is authorized to "remove or arrange for the removal of, and provide for remedial action relating to such hazardous substance, pollutant, or contaminant at any time . . . or take any other response measure consistent with" the statutory scheme. 42 U.S.C. § 9604(a)(1).

Sites for which the EPA has determined that the need for such a response action is "urgen[t]" are listed on the NPL. See 42 U.S.C. § 9605(a)(8)(A)-(B); United States v. Gen. Elec. Co., 670 F.3d 377, 381 n.3 (1st Cir. 2012); Bd. of Regents of Univ. of Wash. v. EPA, 86 F.3d 1214, 1217 (D.C. Cir. 1996). The NPL must be "revise[d] . . . no less often than annually." 42 U.S.C. § 9605(a)(8)(B).

To select a response action for a site, the EPA conducts both a remedial investigation and a feasibility study. See 40 C.F.R. § 300.430(a)(2); Carson Harbor Vill., Ltd. v. County of Los Angeles, 433 F.3d 1260, 1267-68 (9th Cir. 2006); see also CPC Int'l, Inc. v. Northbrook Excess & Surplus Ins. Co., 962 F.2d 77, 79 (1st Cir. 1992) (describing the remedial investigation and feasibility study as a "predicate to necessary remediation"). In the remedial investigation phase, the EPA evaluates the need for

3Much of the authority granted to the President under CERCLA has been delegated to the EPA. See Kelley v. EPA, 15 F.3d 1100, 1103 (D.C. Cir. 1994) (citing Exec. Order No. 12,580 § 1(b)(1), 52 Fed. Reg. 2923, 2923 (Jan. 23, 1987)).

7

a response action and collects the information necessary to assess the possible response actions that could be taken. 40 C.F.R. § 300.430(d)(1).

The EPA is required to select a response action that is "protective of human health and the environment," will "maintain protection over time," and will "minimize untreated waste." Id. § 300.430(a)(1)(i). To that end, the EPA conducts a feasibility study that, based on the data gathered in the remedial investigation, assesses possible response actions against a range of criteria, including cost, complexity, environmental impact, benefits to human health, and state and community buy-in. See id. § 300.430(e)(7), (9).

In that study, the EPA identifies a preferred response action and opens it to public comment. See id. § 300.430(f)(1)(ii). Based on the public comments that the EPA receives and its own analyses, the EPA then selects a response action. Id. § 300.430(f)(4)(i). The EPA also at that point compiles the documents that formed the basis for its selection of the response action in an administrative record that includes the Record of Decision ("ROD"). Id. §§ 300.430(f)(5)(i), 300.800(a).

The response action that the EPA selects can be carried out by the EPA itself; alternatively, the EPA can order "responsible parties" under section 107(a) of CERCLA to carry it out. Key Tronic Corp., 511 U.S. at 813-14; see also 42 U.S.C.

8

§ 9607(a). CERCLA addresses how the costs of carrying out the response action may be allocated among responsible parties.

For "four broad classes" of responsible parties, including any corporation "who at the time of disposal of any hazardous substance owned or operated any facility at which such hazardous substances were disposed of," "CERCLA imposes strict liability for environmental contamination." Burlington N. & Santa Fe Ry. Co. v. United States, 556 U.S. 599, 608-09, 608 n.5 (2009) (quoting 42 U.S.C. § 9607(a)). Moreover, CERCLA provides that any given responsible party can be held liable by the EPA for the entire cost of carrying out the response action. Id. at 614-15.

Where there are multiple responsible parties, however, a "CERCLA defendant[] seeking to avoid joint and several liability bear[s] the burden of proving that a reasonable basis for apportionment exists." Id. at 614. In addition, even where the harm is not susceptible to apportionment, CERCLA "permit[s] . . . private parties [who are themselves responsible parties] to recover cleanup costs and seek contribution from [other] responsible parties" for carrying out the response action. City of Bangor v. Citizens Commc'ns Co., 532 F.3d 70, 90 (1st Cir. 2008).

Specifically, a responsible party may seek cost recovery under section 107, see 42 U.S.C. § 9607, or contribution under section 113, see 42 U.S.C. § 9613, against any other responsible

9

party. Section 107(a) "allows for full recovery of costs," United States v. Davis, 261 F.3d 1, 29 (1st Cir. 2001), by a responsible party from other responsible parties, unless the harm can be apportioned in a manner that would preclude such full recovery, Burlington N. & Santa Fe Ry. Co., 556 U.S. at 614-15. Section 113(f)(1), by contrast, allows a responsible party to seek contribution from other responsible parties for costs incurred in carrying out a response action subject to the allocation of those costs by a court "using such equitable factors as the court determines are appropriate." 42 U.S.C. § 9613(f)(1). Additionally, a party who is held liable under section 107(a) as a responsible party can petition the President under section 106(b)(2) of CERCLA to recover reasonable costs associated with its carrying out the remedy that the EPA ordered it to perform to clean up a site, id. § 9606(b)(2)(A), if that responsible party can establish that the response action that the EPA required was "arbitrary and capricious or otherwise not in accordance with law," id. § 9606(b)(2)(D).

There is one further piece of this intricate legislative framework that bears on the issues before us. CERCLA contemplates the possibility that one or more responsible parties will choose to settle with the United States. Indeed, "early settlement[] . . . is an integral part of the statutory plan." See United States v. Cannons Eng'g Corp., 899 F.2d 79, 92 (1st

10

Cir. 1990). CERCLA provides in this connection that a party that "has resolved its liability to the United States or a State in a[] . . . judicially approved settlement shall not be liable for claims for contribution regarding matters addressed in the settlement." 42 U.S.C. § 9613(f)(2). When such a settlement is entered, it "reduces the potential liability of the [nonsettling parties] by the amount of the settlement," id., instead of by the settling parties' equitable share of the contamination, and thus can lead to "disproportionate liability" for nonsettlors, Cannons, 899 F.2d at 91.

C.

Between 2000 and 2003, the EPA issued various administrative orders directing NECC, Emhart, and others to undertake response actions at the Site to effectuate its cleanup. In 2006, Emhart filed cost recovery and contribution claims under, respectively, CERCLA sections 107(a) and 113(f)(1) against NECC (as a responsible party) and its insurers. Phase I, 130 F. Supp. 3d at 538. Emhart alleged in those claims that it had incurred various costs in carrying out the EPA's administrative orders setting forth the response actions and that it expected to continue to accrue costs in doing so in future response actions.

In 2011, Emhart filed another set of cost recovery and contribution claims under those respective provisions of CERCLA in connection with the cleanup of the Site. Id. This time, the

11

claims were against the federal agencies, id., as well as the United States as the party that "controls" them. Emhart alleged that the federal agencies were themselves responsible parties under section 107(a) of CERCLA, because they had shipped drums carrying toxins to NECC during the time period in question. See id. at 540-41.

Emhart alleged in its complaint that while the EPA had not yet determined how the Site should ultimately be remediated, it believed that the EPA ultimately would "demand that Emhart undertake additional remedial work at the Site potentially costing hundreds of millions of dollars." The United States counterclaimed, both on behalf of the federal agencies -- as responsible parties in their own right -- for contribution under section 113(f)(1) of CERCLA and on behalf of the EPA for cost recovery under section 107(a) of that statute. Phase I, 130 F. Supp. 3d at 538.

In 2012, Emhart's CERCLA suit against the EPA and the federal agencies was consolidated with its CERCLA suit against NECC and its insurers. Id. at 538 n.3. Later that year, the United States, on behalf of the EPA and the federal agencies, brought CERCLA claims for both cost recovery and contribution in connection with that pending litigation against a collection of third-party defendants -- namely, various companies that had sent

12

drums to NECC for reconditioning at the Site. Those third-party defendants included each of the appellants here.4 Also in 2012, the EPA issued the ROD ("the 2012 ROD")

for the Site. Id. at 601. The 2012 ROD identified the agency's selected response action (which encompassed a number of discrete remedial and removal actions) for the Site going forward. It also described the facts, analyses, and policy considerations the EPA had accounted for in its remedy-selection process. See id.; Emhart Indus., Inc. v. New Eng. Container Co. (Phase II), 274 F. Supp. 3d 30, 42 (D.R.I. 2017). The ROD estimated the total outstanding cost associated with implementing the response action going forward at over $100 million. Phase I, 130 F. Supp. 3d at 601- 02.

In 2014, the EPA issued a Unilateral Administrative Order ("UAO"). Phase II, 274 F. Supp. 3d at 79. The UAO directed Emhart to perform the response action designated in the 2012 ROD. Id. By the time the UAO issued, the parties to the consolidated CERCLA litigation involving the EPA, Emhart, the NECC, the federal agencies, and the other companies accused of sending drums to the Site were already engaged in discovery. The District Court then

Eventually, Emhart also filed CERCLA cross-claims against 4

the third-party defendants, including appellants, each of which has since brought cross-claims against Emhart under CERCLA. CNA Holdings has also filed CERCLA counterclaims against the United States.

13

entered a case management order providing that this litigation would proceed in three phases.

The first phase would determine whether Emhart and NECC were liable under section 107(a) of CERCLA as responsible parties, including whether they were jointly and severally liable or whether the harm that each had caused was subject to apportionment. The second phase would address the response action that the EPA had ordered Emhart to carry out in the UAO and how it would impact parties ultimately deemed to be responsible parties under CERCLA, whomever they might be. The third and final phase, "if required," would address the liability as responsible parties under CERCLA of the third-party defendants -- and thus of the appellants -- as well as the amount of money each would have to pay given the pending CERCLA claims for cost recovery and contribution involving them. The District Court stayed the third-party defendants' obligations to take or submit discovery but provided that the third-party defendants could attend the depositions and review the documents produced in discovery for the first two phases.

In the case management order, the District Court stated that it would "not rule on the liability of the [Department of Defense], or its amount in contribution, if any, until the third phase." Phase I, 130 F. Supp. 3d at 539. Even still, that order instructed that "[a]ll evidence pertaining to the [the Department of Defense's] liability for contamination of the Site [would] be

14

presented during the first phase . . . of the trial." Id. The District Court explained that "the evidence [would] be used solely to determine the liability of Emhart and NECC and whether this liability (if proven) is divisible among the two parties." Id.

D.

Shortly before Phase I of the trial began, NECC settled its claims with the federal parties. Id. The District Court approved and entered a consent decree reflecting that settlement. Id. That consent decree provided NECC with protection against contribution claims under CERCLA by any other responsible party in connection with the Site. It thus resulted in the dismissal of all of the pending CERCLA claims between Emhart, NECC, and NECC's insurers. Id. at 539 & n.6.

In light of these developments, the first phase of the trial focused on Emhart's liability under CERCLA as a responsible party. Id. at 539. That first phase was completed following a twenty-day bench trial in May and June 2015. Id. at 540.

Emhart's frontline position in that trial was that it was not a responsible party under CERCLA. Id. Emhart also advanced an argument in the alternative, however. It contended that, even as a responsible party, it was not jointly and severally liable for the costs of the entire cleanup of the Site under CERCLA because its liability could be apportioned with NECC's. Id. At the same time, though, Emhart continued to maintain its CERCLA

15

claims against the federal agencies for both cost recovery and contribution. Id.

At the end of Phase I, the District Court concluded that Emhart was jointly and severally liable under section 107(a) of CERCLA as a responsible party. Id. at 602. The District Court then went on to hold that Emhart had not proved by a preponderance of the evidence that the barrels that the Department of Defense sent to the Source Area contained toxic substances. Consequently, the District Court rejected Emhart's CERCLA claims for both contribution and cost recovery against the federal agencies. Id.

Phase II of the consolidated litigation then began. The District Court conducted a thirteen-day bench trial that came to an end in January 2017. Phase II, 274 F. Supp. 3d at 37 n.2. At the trial, the District Court admitted evidence, accepted post- trial briefing, and heard oral argument concerning the response action that the EPA had ordered Emhart to carry out in the UAO. Id.

The District Court again noted during Phase II that "[t]he necessary contributions, if any, of third-party defendants [would] be addressed in Phase III of the trial." Id. at 38. Accordingly, Phase II focused on whether the EPA's "remedy- selection process," as reflected in the 2012 ROD, for the response action Emhart had been ordered to carry out in the UAO "was

16

arbitrary, capricious, or otherwise not in accordance with law." Id. at 37-38.

That response action required, among other things, removal of buried waste and contaminated sediment from the Source Area, disposal of those materials, installation of a hazardous- waste cap and soil cover, and long-term monitoring and maintenance. Id. at 52-53, 52 n.23. The technical requirements for the landfill cap were derived from Subtitle C of the Resource Conversation and Recovery Act ("RCRA"), 42 U.S.C. § 6901 et seq., and thus called for a specialized landfill cap in order to prevent toxins from seeping into the groundwater and contaminating nearby bodies of water. See Phase II, 274 F. Supp. 3d at 51 n.17. We will refer, following the parties, to this cap as the "RCRA C cap."

The District Court noted that, with respect to the response action that Emhart was ordered to carry out in the UAO, Emhart had "argue[d] that several of EPA's individual actions and decisions along the way were either arbitrary, capricious, or not in accordance with CERCLA." Phase II, 274 F. Supp. 3d at 53. Indeed, Emhart challenged the response action on a host of different grounds. See id. at 53-80. The District Court did find that "[a]s a general matter, . . . EPA followed the basic steps mandated by CERCLA and the [National Contingency Plan] in developing its remedial action for the Site." Id. at 53. Nevertheless, the District Court found that Emhart had

17

successfully established, on the then-current record, as reflected in the 2012 ROD, that several distinct decisions the EPA made in its remedy-selection process were, per CERCLA, arbitrary, capricious, or otherwise not in accordance with law. Id. at 80. Consequently, the District Court stayed the UAO compelling Emhart to carry out the response action documented in the 2012 ROD until "th[ose] matters [were] resolved." Id. at 81.

As to how the EPA could go about solving those problems, the District Court explained that it could "envision several ways EPA could approach the[] deficiencies." Id. One possibility the District Court contemplated was that the EPA might decide the best path forward was to "reopen the remedial investigation and feasibility study process." Id. The District Court ultimately took "no view as to the appropriate course of action" and left it "to EPA to address these issues in the first instance," while retaining jurisdiction over the matter to ensure they were properly handled. Id.

Two months later, the United States, on behalf of the EPA, filed a motion for reconsideration of the District Court's ruling concerning the response action at issue. The United States noted that the EPA had made "hundreds of decisions" in selecting the response action and that, "[a]fter the lengthy trial and months of painstaking analysis, and post-trial briefs totaling over 900 pages," the District Court had "narrowed the number of alleged EPA

18

errors down to three." The United States argued that "the Administrative Record support[ed] EPA's determinations on all three issues" and that, even if those three decisions were erroneous, "the record show[ed] that correcting the alleged errors would have made no difference in the selection of the remedy."

In June 2018, while the EPA's motion for reconsideration was still pending, the State of Rhode Island filed various claims against Emhart under CERCLA and state environmental laws, see R.I. Gen. Laws §§ 23-18.9-10, 23-19.1-22. That action was consolidated with the pending litigation in the District Court on July 6.

E.

Three days later, on July 9, 2018, the United States and the State of Rhode Island lodged a proposed version of the Decree with the District Court. The proposed decree would have resolved all claims regarding the Site by the EPA and the federal agencies against Emhart under CERCLA, by Rhode Island against Emhart under CERCLA and R.I. Gen. Laws §§ 23-18.9-1 et seq., 23-19.1-1 et seq., and 23-19.14-1 et seq., and by Emhart against the EPA and the federal agencies under CERCLA, subject to certain standard reservations.

The proposed decree also required Emhart to pay the governments' unrecovered costs for the past cleanup. Those past costs totaled approximately $42 million; the anticipated costs associated with the remaining cleanup at that point were $96

19

million. Additionally, under the proposed decree, Emhart agreed "to perform the remedy EPA selected in the 2012 ROD for the Site." It was accompanied by a Statement of Work ("SOW"), which added further detail as to how Emhart had promised to implement that response action. For example, under the SOW, Emhart would be permitted "to investigate and propose for EPA's consideration potential modifications to the remedy."

With respect to the federal agencies, the proposed decree provided that they collectively would pay $550,000 to "resolve any liability they may have in connection with the Site." Moreover, both Emhart and the federal agencies would "receive protection from contribution actions or claims as provided in Section 113(f)(2) of CERCLA . . . for the defined matters addressed by the Consent Decree." The United States also indicated that it intended to dismiss its CERCLA claims for cost recovery on behalf of the EPA and its CERCLA claims for contribution on behalf of the federal agencies against the third-party defendants, including appellants (although Emhart made no such commitment).

Then, in September of 2018, after opening the proposed decree for public comment, the United States and Rhode Island jointly filed a motion for entry of the Decree. In the motion, the United States and Rhode Island argued that the Decree was "fair, reasonable, and consistent with the goals of CERCLA."

20

Appellant CNA Holdings filed a motion in opposition, along with other third-party defendants, including the two other appellants here. The third-party defendants objected on two grounds relevant to the issues we must decide on appeal: (1) that the proposed decree called for implementation of a remedy that was the same as the one required by the response action that the District Court had previously held was arbitrary and capricious under CERCLA, thereby rendering the Decree itself inconsistent with CERCLA; and (2) that the proposed decree was "not substantively fair" as a whole because no justification was provided for setting the federal agencies' payment at $550,000, such that the Decree could not be approved.

Emhart and the United States both filed motions in response. The District Court then held a hearing on the question of whether to approve the proposed decree. A few weeks later, on April 8, 2019, the District Court entered an order approving the Decree. The District Court explained that "[a]fter a thorough review of the 2012 Record of Decision, Consent Decree, Statement of Work, the United States' Motion for Reconsideration, all parties' briefing related to the Consent Decree, and the representations made at the March 19, 2019 hearing," it had concluded "that the remedial action described in the ROD, when viewed in light of how the Statement of Work and Consent Decree propose to effectuate that remedial action, is not inconsistent

21

with CERCLA and the National Contingency Plan." (citations omitted). The District Court further held that "none of the arguments presented by Third-Party Defendants in their Oppositions poses an obstacle to approving the . . . Decree at this time." The District Court thus vacated its Phase II decision and approved the Decree. Finally, the District Court found that there was "no just reason for delay" and certified its decision as final within the meaning of Federal Rules of Civil Procedure 54 and 58.

Three of the third-party defendants -- CNA Holdings LLC, Exxon Mobil Corporation, and Union Oil Company of California -- filed a timely notice of appeal on May 31, 2019. See Fed. R. App. P. 4(a)(1)(B). We have jurisdiction under 28 U.S.C. § 1291.

II.

The appellants put forth three grounds for rejecting the District Court's approval of the Decree. We first address their assertion that it was an abuse of discretion for the District Court to approve the Decree because it incorporates a remedy that is identical to the response action the EPA selected in the 2012 ROD that the District Court had determined was "arbitrary and capricious" under CERCLA. We then address the appellants' contention that the District Court erred in approving the Decree because the fact that it required the federal agencies to pay only $550,000 to protect themselves from liability rendered it "substantively unfair." Finally, we turn to the appellants'

22

argument that the District Court erred in approving the Decree by failing meaningfully to review it.5 The District Court's approval of the Decree "is encased in a double layer of swaddling." Cannons, 899 F.2d at 84. The first layer inheres in the standard of review a district court must itself apply in deciding whether to approve a consent decree settling claims under CERCLA involving the United States, as that standard "implicates the trial court's deference to the agency's expertise and to the parties' agreement." Id. Under this standard, a district court may approve such a consent decree only if it is "reasonable, faithful to the statute's objectives, and fair (both procedurally and substantively)." United States v. Charles George Trucking, Inc., 34 F.3d 1081, 1084 (1st Cir. 1994).

5 Though no party has argued otherwise, we note that the appellants' claims are justiciable. Given that the appellants have plausibly pled valid cost recovery and contribution claims against Emhart and the federal agencies that will be extinguished pursuant to the Decree, they have a cognizable interest in the Decree's approval and entry that gives them standing to appeal. See City of Bangor, 532 F.3d at 92-93. Moreover, appellants' rights to pursue their claims against Emhart and the federal agencies were directly affected by the District Court's decision. The fact that the precise amount they may one day be required to pay is not yet set does not affect our analysis of the legal dispute before us, which determines whether they will be able to bring those claims at all. Thus, the nature of this dispute is neither "premature" nor "abstract," McInnis-Misenor v. Me. Med. Ctr., 319 F.3d 63, 70 (1st Cir. 2003) (quoting Abbott Lab'ys v. Gardner, 387 U.S. 136, 148 (1967)), but is instead ripe for our review.

23

"The second layer of swaddling derives from the nature of appellate review." Cannons, 899 F.2d at 84. We "will overturn a district court's decision to approve the entry of a CERCLA consent decree in a case involving the United States 'only for manifest abuse of discretion.'" City of Bangor, 532 F.3d at 93- 94 (quoting Charles George Trucking, 34 F.3d at 1085). To meet that standard, the objectors must establish that "the lower court made a serious error of law or suffered a meaningful lapse of judgment." Charles George Trucking, 34 F.3d at 1085.

A.

We start with the appellants' contention that the District Court erred in approving the Decree because it requires Emhart to implement a remedy that is identical to the response action that the EPA selected in the 2012 ROD and that the District Court found to be, at least in part, arbitrary and capricious under CERCLA in Phase II of the consolidated litigation. See 42 U.S.C. § 9613(j)(2). In pressing this argument, the appellants do not contend that, in making that determination, we must confine ourselves to a consideration of only the administrative record that was then in place. Indeed, the appellees reference, in support of there being a sound basis for the remedy, not only the 2012 ROD, but also the explanations for the response action identified in the 2012 ROD that the EPA offered in the course of the Phase II litigation. They also reference the information that

24

the EPA submitted to the District Court when it sought approval of the Decree, as well as the Decree itself. Yet, the appellants make no argument that these additional materials, which were not part of the administrative record at the time of the Phase II litigation, fall outside the relevant record for purposes of our review of the District Court's approval of the Decree.

Accepting this understanding of the relevant record in this appeal, we reject the appellants' challenge to the District Court's decision to approve the Decree insofar as it rests on the District Court's Phase II ruling that the response action there at issue was arbitrary and capricious under CERCLA. We emphasize that, in doing so, we assume, to the appellants' benefit, that the District Court did not err in determining in the Phase II litigation that the response action at issue there was arbitrary and capricious. For, as we will explain, that ruling does not itself provide a basis for concluding that the District Court abused its discretion in holding that the Decree, despite its inclusion of a remedy that mirrored that response action, was "reasonable, fair, and consistent with the purposes that CERCLA is intended to serve.'" Cannons, 899 F.2d at 85 (quoting H.R. Rep. No. 253, pt. 3, at 19 (1985)); see also id. at 85-86 (noting that Congress intended for us to "take a broad view of proposed settlements" and leave "highly technical issues . . . to the discourse between [the settling] parties").

25

1.

The appellants first ask us to focus on the Decree's requirement to install the RCRA C cap. As we have explained, the response action the EPA selected in the 2012 ROD also included such a requirement. The appellants argue that this component of the Decree's remedy is inconsistent with CERCLA, such that the Decree as a whole is, because it is predicated on restoration goals that the EPA set for the Source Area's drinking water during the remedy-selection process documented in the 2012 ROD which the appellants contend are flawed. As support for that argument, they note that the District Court invalidated the requirement to install a RCRCA C cap in the earlier response action as "arbitrary and capricious" during Phase II, because it determined that it was predicated on those restoration goals, which it found were unsupportable.6 But, we are not persuaded by this challenge to the District Court's decision to approve the Decree.

The appellants are right that, in selecting the 2012 response action that included the requirement to install such a RCRA C cap, the EPA classified the groundwater at the Source Area as "a potential source of drinking water." Phase II, 274 F. Supp. 3d at 65. They are also right that, in Phase II of the consolidated

6At oral argument, counsel for appellants confirmed that they only object to the groundwater classification insofar as it influenced the EPA's decision to require a RCRA C cap in the remedy.

26

litigation, the District Court concluded that it was "overwhelmingly clear" that, at that time, "the Source Area groundwater [was] . . . far too contaminated to provide a source of drinking water" and that it would be unrealistic to expect much change on that front. Id. at 66.

Additionally, as the appellants rightly note, the District Court found at that time that, as a result of that classification, the EPA "adopt[ed] stricter cleanup goals for the Source Area groundwater" than it otherwise would have. Id. at 65. The appellants are right as well that it was on those grounds that the District Court then concluded in Phase II that the drinking water classification was arbitrary and capricious because the EPA had "not collected sufficient information or conducted sufficient analysis" to support its conclusion that the groundwater could one day be potable. Id. at 67. And, finally, the appellants are correct that the District Court concluded that the cap was likewise arbitrary and capricious because, judging from the 2012 ROD, the decision to require it appeared to be "inextricably intertwined" with the groundwater goals. Id. at 68.

But, the District Court explained in its Phase II ruling that it did not take issue with the EPA imposing a RCRA C cap requirement per se. Id. Instead, it expressly raised the possibility that the EPA might be able to justify the RCRA C cap

27

requirement by establishing that it was "necessary regardless of EPA's groundwater remediation goals." Id.

That is significant for present purposes because the record that was before the District Court at the time that it approved the Decree -- and thus the one that we now consider -- includes a motion from the EPA highlighting various aspects of its decision-making process that demonstrate that it had concluded as of the time that the District Court reviewed the Decree that the RCRA C cap requirement was appropriate to include notwithstanding the District Court's groundwater ruling. Specifically, the EPA explained that, even though the RCRA C cap would help facilitate groundwater cleanup, that fact "was additive to the already existing reasons that a RCRA C cap was appropriate" and that it intended to require the RCRA C cap for other, independent reasons as well.

The EPA then recounted some of those considerations. It explained that "the justifications for the RCRA C cap include better protection against scour and erosion from flooding, physical containment of contaminated soils, and more reliable and robust long term protection." The EPA also pointed out that, regardless of its plans for the groundwater, it could only leave certain chemicals in place in the Source Area if they were under a cap that "complie[d] with the requirements of Subtitle C of RCRA."

28

Moreover, the EPA reinforced those representations during the hearing that the District Court held on the Decree itself. When the District Court asked the EPA's counsel about whether the existence of the Decree "change[d] the fact that [the court had] found in Phase II" that the 2012 ROD was arbitrary and capricious, the attorney referred the District Court to its motion for reconsideration in which it had listed the independent reasons for requiring the RCRA C cap. The EPA's counsel also stated at that time that it was the EPA's position that the RCRA C cap was "required anyway regardless of the groundwater issues." And, indeed, in response, the District Court noted that, based in part on those representations, "the landscape ha[d] significantly changed from where it was at the close of the evidence" in Phase II.7 We also find it significant that the SOW attached to the Decree provided that the EPA and the settling parties anticipated potentially reclassifying the Source Area groundwater. In general, the EPA defers to state groundwater classification guidelines when developing response actions in states with

The EPA did also argue in its motion for reconsideration 7

and in the hearing on the Decree that its initial 2012 ROD was not flawed. But, the agency's representations there still inform our understanding of its justifications for including the remedy in the Decree that is before us today, even if it separately believed the same remedy was justified on the record contained in the 2012 ROD.

29

approved schemes. See U.S. Env't Prot. Agency, OSWER Directive No. 9283.1-09, The Role of CSGWPPs in EPA Remediation Programs (1997), 1997 WL 1068504. At the time of the Phase II litigation and when the Decree was approved, Rhode Island did not have an approved state classification scheme, so the EPA used the federal one. Thus, the EPA categorized the Source Area's water as a potential source of drinking water even though the Rhode Island Department of Environmental Management "considered the Source Area groundwater to be so contaminated" as to be "unsuitable for potential use as drinking water." Phase II, 274 F. Supp. 3d at 64.

In the SOW, however, the EPA and the settling parties noted that they would consider reclassifying the groundwater pursuant to Rhode Island's scheme in the event that it was approved. Despite contemplating the possibility of that change, the SOW did not provide that the RCRA C cap requirement might be eliminated. Instead, the Decree required the RCRA C cap regardless of whether the groundwater was recategorized.

Thus, we see no basis for concluding that the requirement to include a RCRA C cap in the remedy in the Decree was a function of the EPA's groundwater remediation goals. Indeed, the appellants acknowledge that there are other, valid reasons for which the EPA might have decided to require a RCRA C cap, groundwater goals aside.

30

The appellants do contend that we cannot be sure the EPA would have required the RCRA C cap in the 2012 response action absent the groundwater classification. But, that is beside the point. The question for present purposes is not whether the inclusion of the requirement to install a RCRA C cap in the response action set forth in the 2012 ROD was arbitrary and capricious. The question is whether the inclusion of that requirement in the remedy set forth in the Decree renders it an abuse of discretion for the District Court to have approved it. Given the reasons the record reveals for concluding that, at the time of the inclusion of the RCRA C cap requirement in the remedy in the Decree, that requirement was justified for reasons unrelated to the groundwater goals, we see no basis for concluding that it was.

To be sure, the District Court had earlier determined, on the basis of the record then before it, that it was arbitrary and capricious to include that requirement in the response action. But, that was then, and what matters is what the record that is now before us reveals about the basis for the District Court's approval. Thus, the appellants' first ground for contending that we must vacate the Decree fails.

2.

The appellants next object that the District Court abused its discretion in approving the Decree based on the fact that the EPA,

31

in selecting the response action in the 2012 ROD that the Decree mirrors, used findings about the effect of the contaminants on the fish population in Allendale Pond and other bodies of water in the Source Area that the appellants contend are inaccurate. The appellants point out that the District Court determined in the Phase II litigation that those findings were unsupportable on the record before it, rendering the response action based on them arbitrary and capricious under CERCLA.8 In accord with the structure of the challenge to the Decree that we have just rejected above, see supra Section II.A.1, the

8 In Phase II, Emhart challenged the fish consumption estimates that the EPA employed in its baseline risk assessment in several respects. In doing so, Emhart challenged, among other things, the EPA's assumptions that no largemouth bass would be consumed from Allendale Pond and that certain populations would consume fourteen grams of fish per day from the Site. See Phase II, 274 F. Supp. 3d at 78. In particular, Emhart argued that, by assuming that white suckers and eels, rather than bass, would be consumed from Allendale Pond, the EPA "skew[ed] its risk calculations" because "white suckers are 'bottom dweller[s]' and thus likely to have larger body burdens of chemical contaminants" and that eels likewise contained high concentrations of dioxin relative to other species. Emhart also took issue with the fact that the EPA assumed that individuals who consumed eels and white suckers would eat the entire fish but that individuals who consumed bass would eat just the less-contaminated fillet, further compounding the effects of assuming the absence of bass. The District Court ultimately concluded that "Emhart ha[d] demonstrated that [the EPA's] misstep [in assuming no bass would be consumed] arbitrarily increased the risk calculation for Allendale Pond." Phase II, 274 F. Supp. 3d at 78. The District Court further concluded that while the EPA "may ultimately determine that fourteen grams is the appropriate reasonable maximum consumption rate at the Site," on the record that was before the District Court, the EPA's decision to use that estimate was arbitrary. Id.

32

appellants thus point to this Phase II ruling to argue that the Decree cannot stand. For, they contend, it, like the 2012 response action, is necessarily inconsistent with CERCLA due to its reliance on these same flawed findings. But, we are not persuaded here either.

When the EPA develops a response action, it uses information like the fish population data that the appellants contest to conduct a baseline risk assessment and set remediation goals. See 40 C.F.R. § 300.430(d)(1), (d)(4), (e)(2)(i). Those goals reflect "acceptable exposure levels that are protective of human health and the environment." Id. § 300.430(e)(2)(i). Thus, if the EPA expects that local residents will catch and consume a large number of fish from a given Superfund site, it is likely to target cleaner water quality standards at that site in order to keep those residents' exposure to toxins low.

In some instances where the EPA conducts such an assessment, it will compute preliminary remediation goals and discover that the background levels of the contaminants in the surrounding environment are higher than the EPA's goals. Where that is the case, the EPA will not develop a response action that targets its original remediation goals. See U.S. Env't Prot. Agency, OSWER Directive No. 9285.6-07P, Role of Background in the CERCLA Cleanup Program 8-9 (2002). Instead, it will simply target the background contaminant level, in part out of a concern that

33

the remediated areas will be recontaminated by their surroundings. See id.

Here, in selecting the response action in the 2012 ROD, the EPA computed its remediation goals assuming that there were no largemouth bass in Allendale Pond and that certain populations would consume fourteen grams of fish per day from the Site. See Phase II, 274 F. Supp. 3d at 78. Based on those assumptions, the EPA's remediation goal targeted a level of contamination, 0.43 parts per trillion ("ppt"), that was lower than that present in the background environment, which was 15 ppt. As a result, the response action that the EPA set forth in its 2012 ROD targeted the background level of 15 ppt rather than the lower remediation goal that was premised on the fish consumption estimates.

Thus, while the District Court found the underlying fish consumption estimates unsupportable on the record before it during Phase II, we do not see how that finding in and of itself bears on whether the remedy in the Decree is problematic. Indeed, in its motion for reconsideration in Phase II, the EPA recalculated its remediation goal based on the alternative fish consumption estimates that Emhart pressed in Phase II. The EPA's recalculation resulted in a remediation goal of 2.77 ppt rather than 0.43 ppt. But, that recalculated figure still represents a lower level of contamination than the background level of 15 ppt. That accords

34

with the EPA counsel's representation at the hearing on the Decree that the changed computations "didn't matter."

Thus, the appellants are unable to show that the EPA's purportedly erroneous estimates from the 2012 ROD impacted the Decree's remedy. Accordingly, their contention that the District Court abused its discretion in approving the Decree, because it impermissibly contains a remedy infected by data found to be arbitrary and capricious in the Phase II litigation, fails.

B.

The appellants separately contend that a CERCLA consent decree's "settlement terms must be based upon, and roughly correlated with, some acceptable measure of comparative fault," even if the apportionment of the harm is "imprecise," Cannons, 899 F.2d at 87, and that the Decree does not meet that test here. The appellants rest this contention on the fact that the Decree required the federal agencies to pay only $550,000 to relieve themselves of any future liability in connection with the Site, despite the fact that the EPA-ordered response action was estimated to cost around $100 million. The appellants contend that, given this gap between what the federal agencies agreed to pay for the cleanup and the costs of the cleanup, we must find that the District Court abused its discretion in approving the Decree. We see no error, however.

35

We have previously noted that "no universally correct approach" to comparative fault exists and that "[w]hatever formula or scheme EPA advances for measuring comparative fault and allocating liability should be upheld so long as the agency supplies a plausible explanation for it." Id. We have such a plausible explanation for the allocation of liability reflected in the Decree that is at issue here.

As the EPA argued to the District Court in defending this aspect of the Decree, the District Court in Phase I of the litigation had "already found the [federal agencies] not liable under CERCLA." Thus, the seemingly yawning gap between what the federal agencies had been made to pay and their potential liability was no gap at all. Moreover, when prompted by the District Court as to why the Department of Defense should pay anything whatsoever, counsel for the EPA explained that the United States was as "concerned as any party about litigation risks" and the "litigation costs" associated with continuing to participate in the lawsuit.

The appellants do object that "Emhart's failure to prove the Agencies' liability during Phase I is not substantial evidence that the Agencies' liability is insignificant." They contend that they may still be able to establish liability where Emhart failed and that therefore they "should get to finish discovery," which they say "will provide EPA and the court the necessary context to

36

evaluate the [federal agencies'] proportional contribution to the contamination at the Site."

At the hearing on the Decree, however, the District Court explained that, "having lived through this for all these years," it believed "that th[e] possibility of liability by the [Department of Defense] was litigated into the ground." Indeed, the District Court remarked that it "c[ouldn't] imagine that . . . [any of the third-party defendants, including appellants,] would think for a moment that there was some potential benefit to going after the [Department of Defense] on that theory."

Accordingly, it was neither a "harmful error of law" nor a "meaningful error in judgment," Cannons, 899 F.2d at 84 (quoting Anderson v. Cryovac, Inc., 862 F.2d 910, 923 (1st Cir. 1988)), for the District Court to accept the EPA's explanation that the federal agencies' liability was essentially nonexistent but that the agencies paid a settlement figure in order to precipitate the end of their role in the litigation.9 CERCLA, after all, is designed

9 Although the appellants contended at oral argument before this Court that the District Court's decision in the Phase I proceedings only addressed the agencies' liability as to dioxin, and that they could still make out claims for contribution with regard to other chemicals, the District Court specifically rejected that argument when it ruled on the federal agencies' liability in Phase I. See Phase I, 130 F. Supp. 3d at 609 (rejecting Emhart's contention that only evidence of the agencies' liability as to dioxin had been presented because the case management order had required that "all evidence relating to the [Department of Defense's] liability, and not just evidence

37

to facilitate early settlement, which supplies a key mechanism by which efficient cleanup of Superfund sites occurs. See Davis, 261 F.3d at 27 ("CERCLA . . . seeks to induce settlements at higher amounts by allowing settlors to seek contribution from those who have not yet settled."); Cannons, 899 F.2d at 92 ("Disproportionate liability, a technique which promotes early settlements and deters litigation for litigation's sake, is an integral part of [CERCLA's] statutory plan.").

The appellants do allege that the federal agencies'

settlement figure was "nepotistically brokered" and that the EPA shut them out of settlement negotiations in favor of reaching an agreement with the federal agencies.10 They emphasize in this regard that the District Court's case management order did stay the appellants' ability to take discovery during Phase I and Phase II.

But, the appellants do not dispute that they were privy to all of the discovery Emhart conducted against the agencies in Phase I. They also have made no proffer that could suffice to

relating to the [Department of Defense's] liability for dioxin, needed to be put forward in this phase").

10The United States asserts that this argument is waived because the appellants did not bring a distinct objection under the heading of procedural fairness. But, procedural and substantive fairness are not entirely discrete concepts; it is "appropriate" for us "to consider the adequacy of the process" in evaluating substantive fairness. Cannons, 899 F.2d at 87 n.4.

38

show that they might succeed where others have failed. Nor was this a consent decree between the EPA and the federal agencies only; Emhart was also a party to it. See Charles George Trucking, 34 F.3d at 1088 ("Sophisticated actors know how to protect their own interests, and they are well equipped to evaluate risks and rewards.").

The lone out-of-circuit, unpublished, district court opinion on which the appellants rely in pressing this contention of self-dealing, United States v. Pesses, No. 90-654, 1994 WL 741277 (W.D. Pa. Nov. 7, 1994), does not convince us of its merit. In that case, like here, the EPA entered a consent decree with several federal agencies (as well as private parties). Id. at *4- 5. But, while the district court there refused to approve the decree due in part to concerns about a "'sweetheart' deal" between the EPA and those agencies, id. at *7, *15-17, it emphasized that the EPA had failed adequately to explain the basis for the "preferential treatment," id. at *18. For the reasons that we have given, however, the District Court in the present case did not abuse its discretion in finding that here the EPA did provide such an explanation.

C.

Finally, we must address the appellants' argument that the District Court failed to appropriately scrutinize the Decree and instead merely rubber stamped it. They argue that the fact

39

that the order approving the Decree provides, by way of explanation, only "a one-line assertion that, based on a 'thorough review' of the materials, the remedy 'is not inconsistent with CERCLA'" is evidence that the District Court failed to exercise independent judgment. We disagree.11 We do not take the appellants to contend that the District Court failed thoroughly to explain or consider the aspects of the remedy in the Decree that mirrored the aspects of the response action that it found in the Phase II litigation were not "arbitrary and capricious." Nor would any such contention be viable, given the District Court's meticulous opinion at that stage. See generally Phase II, 274 F. Supp. 3d 30. Instead, we understand the appellants to argue that the District Court failed adequately to explain why the aspects of the remedy embedded in the Decree that mirrored the aspects of the response action that it found were "arbitrary and capricious" in the Phase II litigation were, in fact, appropriate when repackaged in the Decree.

As part of this argument, the appellants contend that the 11

District Court's decision to vacate its Phase II ruling without specifically explaining its reasons for doing so is evidence that the District Court "mistook its task" and did not exercise independent judgment. But, the appellants do not argue that the District Court lacked the power to vacate the earlier interlocutory ruling, and we do not see any reason that the District Court's choice to do so would impose a heightened requirement that the court explain its reasoning. Thus, the fact of vacatur provides no independent basis for concluding that the District Court abused its discretion.

40

First, the appellants understate the quality of the District Court's explanation with respect to the import of its Phase II findings as to the decision whether to approve the Decree. In explaining that decision, the District Court catalogued the extensive evidence that it had reviewed, including the original remedy the EPA selected, the Decree, the SOW, the United States' Motion for Reconsideration, the parties' briefing on the Decree, and the parties' representations at the hearing on the Decree, much of which post-dated the Phase II proceedings. It then explained that it had concluded "that the remedial action . . . when viewed in light of how the Statement of Work and Consent Decree propose to effectuate [it], is not inconsistent with CERCLA and the National Contingency Plan." (emphasis added). Citing to the SOW, moreover, the District Court also specifically noted that "[f]or example, [it] expect[ed] the parties [would] consider and implement the reclassification of groundwater in line with the state groundwater classification system." Then, consistent with those conclusions, the District Court "approve[d] the Consent Decree as fair, reasonable, and consistent with the goals of CERCLA."

These statements indicate that the District Court carefully considered the ways in which the record before it at the time of its decision to approve the Decree differed from the record before it when it ruled in the Phase II litigation on the response

41

action. They reflect, in other words, a reasoned judgment rather than the absence of one.

Moreover, we have previously acknowledged the Supreme Court's guidance that we ought to be "reluctant" to invalidate a district court's entry of a consent decree "solely because the court failed adequately to set forth its reasons or the evidence on which they were based." United States v. Comunidades Unidas Contra La Contaminacion, 204 F.3d 275, 280 (1st Cir. 2000) (quoting Protective Comm. for Indep. Stockholders of TMT Trailer Ferry, Inc. v. Anderson, 390 U.S. 414, 437 (1968)). As we have explained, "unless we sense[] something deeply amiss," even where a district court does not set forth its reasons for approving a consent decree as is advisable, we ask only "whether the record contains adequate facts to support the decision of the district court to approve the proposed compromise." Id.

We see nothing deeply amiss here. The District Court has been living with this litigation for over a decade. It is plainly immersed in the details that bear on the remedy contained in the Decree. See generally Phase I, 130 F. Supp. 3d 534; Phase II, 274 F. Supp. 3d 30. That "gives [us] confidence that a neutral adjudicator, intimately acquainted with the case, has focused on the essential criteria and found them not lacking." See Comunidades Unidas Contra La Contaminacion, 204 F. 3d at 280.

42

III.

For the reasons that we have set forth, we affirm the District Court's decision to approve the Decree.