Boynes v. Limetree Bay Ventures, LLC

Opinion

DISTRICT COURT OF THE VIRGIN ISLANDS DIVISION OF ST. CROIX

) CLIFFORD BOYNES, et al., ) ) Civil Action No. 2021-0253 Plaintiffs, ) ) v. ) ) LIMETREE BAY VENTURES, LLC, et al., ) ) Defendants. ) __________________________________________) ) HELEN SHIRLEY, et al., ) ) Civil Action No. 2021-0259 Plaintiffs, ) ) v. ) ) LIMETREE BAY VENTURES, LLC, et al., ) ) Defendants. ) __________________________________________) ) FRANCIS E. CHARLES and THERESA J. ) CHARLES, ) ) Civil Action No. 2021-0260 Plaintiffs, ) ) v. ) ) LIMETREE BAY VENTURES, LLC, et al., ) ) Defendants. ) __________________________________________) ) BEECHER COTTON, et al., ) ) Civil Action No. 2021-0261 Plaintiffs, ) ) v. ) ) LIMETREE BAY VENTURES, LLC, et al., ) ) Attorneys:

John K. Dema, Esq., Vincent A. Colianni, II, Esq., Lee J. Rohn, Esq., Rhea Lawrence, Esq., St. Croix, U.S.V.I. Marina Leonard, Esq., Jennifer Jones, Esq., Lee J. Rohn, Esq., St. Thomas, U.S.V.I. St. Croix, U.S.V.I Hugh P. Lambert, Esq., Shanon Jude Carson, Esq., Brian James Mersman, Esq., Daniel H. Charest, Esq., Kerry J. Miller, Esq., Quinn Burns, Esq., C. Hogan Paschal, Esq., Warren T. Burns, Esq., Paul C. Thibodeaux, Esq., Dallas, Texas Rebekka C. Veith, Esq., John Quin Kerrigan, I, Esq., J. Christopher C. Zainey, Jr., Esq., Doylestown, Pennsylvania New Orleans, Louisiana John Albanese, Esq., For the Boynes Plaintiffs Minneapolis, Minnesota Dena R. Young, Esq., Vincent A. Colianni, II, Esq., Philadelphia, Pennsylvania John-Russell Bart Pate, Esq., Charles Jacob Gower, Esq., Lee J. Rohn, Esq., Hugh. P. Lambert, Esq., St. Croix, U.S.V.I. Harry Richard Yelton, Esq., Warren T. Burns, Esq., New Orleans, Louisiana Daniel H. Charest, Esq., For the Cotton Plaintiffs Dallas, Texas Charles Jacob Gower, Esq., Carl A. Beckstedt, III, Esq., Hugh. P. Lambert, Esq., Robert J. Kuczynski, Esq., Korey A. Nelson, Esq., Ryan C. Stutzman, Esq., Harry Richard Yelton, Esq., St. Croix, U.S.V.I. New Orleans, Louisiana Kevin J. Bruno, Esq., For the Shirley Plaintiffs New York, New York Melanie S. Carter, Esq., John K. Dema, Esq., Philadelphia, Pennsylvania Lee J. Rohn, Esq., For Defendant Limetree Bay St. Croix, U.S.V.I Terminals, LLC Hugh. P. Lambert, Esq., Brian James Mersman, Esq., J. Christopher C. Zainey, Jr., Esq., New Orleans, Louisiana For the Charles Plaintiffs MEMORANDUM OPINION

Lewis, District Judge

THIS MATTER comes before the Court on Plaintiffs’ “Amended Motion[s] for Temporary Restraining Order and Preliminary Injunction,”1 and Plaintiffs’ and Defendant Limetree Bay Terminals’ (“Terminals”) second phase briefing. For the reasons stated in the Court’s First Phase Memorandum Opinion, the Court will grant Plaintiffs’ Motions insofar as Plaintiffs have established their entitlement, during the pendency of this litigation, to a water distribution program by Terminals for Plaintiffs and putative class members who cannot afford to purchase water without trading off other basic necessities. For the reasons discussed herein, the Court will order that the water distribution program be established and operated in the manner set forth in the “Order Implementing Water Distribution Program,” filed contemporaneously herewith. I. BACKGROUND2

St. Croix is home to one of the largest oil refineries in the United States. Terminals and its sister company, Limetree Bay Refining (“Refining”), owned and operated the refinery in the Spring of 2021, when the refinery restarted its operations after nearly a decade of closure. Problems arose almost immediately, as a certain “flare” discharged oil and other toxic gases over the island on four separate occasions. In two of these release events—on February 4, 2021, and May 12, 2021—the flare discharged substantial amounts of oil, at least some of which landed on residents’ homes and infiltrated their cisterns, the primary source of

1 Civ. No. 2021-0253, Dkt. Nos. 141, 142; Civ. No. 2021-0260, Dkt. Nos. 68, 68-1; Civ. No. 2021-0261, Dkt. Nos. 170, 171; Civ. No. 2021-0259, Dkt. No. 41. Plaintiffs filed substantially similar motions in each of the four associated cases, and the parties’ briefs have mostly been cross-filed among the four cases. Unless otherwise noted, all citations herein potable water for many residents. In the other two release events—on April 19, 2021, and from May 5-7, 2021—the flare emitted excessive amounts of hydrogen sulfide, leading the Governor of the Virgin Islands to mobilize the National Guard after residents complained of foul smells and health issues. Two days after the May 12, 2021 release event, the

Environmental Protection Agency (“EPA”) issued an emergency order addressed to Terminals and Refining, requiring—among other things—that the refinery cease operations.3 The refinery has not resumed production since that time. The 44 plaintiffs in these four associated cases allege injuries stemming from the release events, principally relating to contamination of their properties and cisterns with petroleum products. In their “Amended Motion[s] for Temporary Restraining Order and Preliminary Injunction,” Plaintiffs requested that the Court order Terminals to remediate the cisterns and property of, and establish a program to provide free bottled water to, Plaintiffs and other similarly situated residents during the pendency of this litigation (hereinafter,

“remedial” and “programmatic” relief, respectively). (Dkt. No. 142 at 2). By Order dated February 2, 2023, this Court determined that the complexity of Plaintiffs’ Motions warranted dividing the evidentiary hearing in this matter into two phases: a first phase “entitlement” hearing, to determine whether Plaintiffs are entitled to the injunctive relief that they seek; and—provided that the Court determined that any such relief was warranted following the first phase hearing—a second phase hearing to determine the precise scope and structure of such relief and the appropriate bond. (Dkt. No. 136). The first phase hearing was held over four days from March 2-7, 2023. For the reasons set forth in the Court’s April 28, 2023 First Phase Memorandum Opinion, this Court found that Plaintiffs had met their burden to establish that some, but not entitled to such relief. (Dkt. No. 285). Specifically, the Court found that Plaintiffs had satisfied the four preliminary injunction factors with respect to only those Plaintiffs and putative class members who cannot afford to purchase water without trading off other basic necessities. Id. at 8. The Court also found that, while Plaintiffs demonstrated entitlement to a

preliminary injunction affording such programmatic relief, they did not do so with respect to remedial relief. Id. Accordingly, by Order dated April 28, 2023, the Court denied Plaintiffs’ Motions insofar as Plaintiffs sought remedial relief, and also insofar as Plaintiffs sought programmatic relief for individuals other than those Plaintiffs and putative class members who cannot afford to purchase water without trading off basic necessities. (Dkt. No. 284). The Court then scheduled a second phase hearing to address the “precise contours” of the population that the Court found to be entitled to programmatic relief; the logistics surrounding the implementation and operation of a water distribution program; and other issues bearing upon

the provision of relief. (Dkt. No. 289). The second phase hearing was held over four days from June 6-9, 2023. The Court has reviewed the parties’ first and second phase briefing, as well as the testimony and evidence presented at both the first and second phase hearings.4 For the

4 The parties’ first phase briefing included, without limitation: Plaintiffs’ Amended Motions (Dkt. Nos. 141, 142); Terminals’ First Phase Response (Dkt. No. 168); Plaintiffs’ First Phase Reply (Dkt. No. 180); Terminals’ First Phase Supplemental Brief (Dkt. No. 267); and Plaintiffs’ First Phase Supplemental Brief (Dkt. No. 269). The parties second phase briefing includes, without limitation: Plaintiffs’ Second Phase Brief (Dkt. No. 317); Terminals’ Second Phase Response (Dkt. No. 327); Plaintiffs’ Second Phase Reply (Dkt. No. 331); Plaintiffs’ Second Phase Supplemental Brief (Dkt. No. 364); Terminals’ Second Phase Supplemental Brief (Dkt. No. 365); Plaintiffs’ Supplemental Brief Regarding the Security Requirement (Dkt. No. 144); and Terminals’ Supplemental Brief Regarding the Security Requirement (Dkt. No. 323). The parties have also submitted—both jointly and separately— reasons stated herein, the Court will order that the water distribution program—to which the Court found certain Plaintiffs and other similarly situated individuals entitled in its First Phase Memorandum Opinion—shall be established and operated in the manner set forth in the accompanying “Order Implementing Water Distribution Program.”

II. DISCUSSION

In the aftermath of the Court’s First Phase Order and the second phase hearing, the parties reached agreement on many issues bearing upon the implementation and operation of the water distribution program.5 Other issues remain in dispute. These issues include, without limitation, the economic and geographic eligibility criteria that applicants must meet to participate in the program; the appropriate bond under Rule 65(c) of the Federal Rules of Civil Procedure; and numerous logistical issues. The accompanying “Order Implementing Water Distribution Program” details the specifics of the water program that Terminals shall establish, including the precise population entitled to participate in the program and other logistics bearing upon the program’s implementation and operation. In this Memorandum Opinion, the Court resolves the parties’ remaining disputes and addresses the changes that the Court makes to the parties’ June 21, 2023 Joint Proposed Order. (Dkt. No. 363-1).6 A. Economic Eligibility Criteria

As discussed above, the Court previously found that Plaintiffs carried their burden to establish that Plaintiffs and putative class members who cannot afford to purchase water without trading off basic necessities are entitled to relief, but that Plaintiffs failed to carry this

Notice Regarding Economic Criteria (Dkt. No. 374); the parties’ Joint Notice Regarding Proposed Testing Protocol (Dkt. No. 375); and Terminals’ Notice Regarding Proposed Testing Protocol (Dkt. No. 376). burden with respect to any other individuals. (Dkt. No. 285 (First Phase Memo. Op.) at 40- 47). At issue in the second phase proceedings is the question of how to capture the population that the Court has found is entitled to relief in a manner that is both reasonable and administratively feasible.

Plaintiffs propose that an applicant be deemed economically eligible for the program if any one of the following six conditions is met: (1) the applicant or a member of the applicant’s household receives public assistance through any one of fifteen enumerated needs-based government programs (Dkt. No. 363-1 (Proposed Order) at § 3.D.i); (2) the applicant or a member of the applicant’s household “would be eligible” to qualify for any one of the fifteen enumerated needs-based government programs, notwithstanding that neither the applicant nor a member of the applicant’s household actually receives said benefits, id. at § 3.D.ii; (3) the applicant or a member of the applicant’s household receives the Earned Income Tax Credit (“EITC”), id. at § 3.D.iii; (4) the applicant or a member of the applicant’s household “would be eligible” to qualify for the EITC based on the EITC’s financial criteria, notwithstanding that neither the applicant nor a member of the applicant’s household actually receives the EITC, id. at § 3.D.iv; (5) “the Applicant’s income (or household income) is below [a] Qualifying Income,” which Plaintiffs define as 150% of the federal poverty guideline plus $11,680, which is Plaintiff’s “estimated annual cost of water,” id. at § 3.D.v; or (6) “[n]either the Applicant nor another related adult in the same house is employed, and at least one adult is unemployed and looking for work within the last four weeks,” id. at § 3.D.vi.

Terminals disagrees with Plaintiffs’ third through sixth proposals, arguing that the EITC, the poverty-line multiple, and unemployment are all poor proxies for the level of need that the Court contemplates here. Id. at §§ 3.D.iii-vi. Additionally, while Terminals agrees with thirteen of the fifteen enumerated needs-based government programs that Plaintiffs propose with respect to the first and second criteria, Terminals argues that two of the programs that Plaintiffs propose—the Low-Income Household Water Assistance Program (“LIHWAP”), 1. Needs-Based Government Programs The parties agree that thirteen of the fifteen proposed needs-based government programs function as adequate proxies for the level of need that the Court contemplates here. (Dkt. No. 413 (Joint Notice) at 2). Following the second phase hearing, the Court ordered the

parties to “submit a joint notice that specifies the ‘financial eligibility criteria’ of each of the fifteen programs enumerated in the parties’ Proposed Order, including without limitation the income eligibility requirements for each program expressed, where applicable, as a multiple of the Federal Poverty Guideline, Local Poverty Guideline, Area Median Income, or comparable metric.” (Dkt. No. 368 (June 30, 2023 Order) at 3 (quoting Dkt. No. 363-1 (Proposed Order) at § 3.D.ii)). The parties submitted this notice in accordance with the Court’s instruction. (Dkt. No. 373 (Joint Notice) at 3-11). Upon review of the financial criteria governing each of the thirteen undisputed programs, the Court agrees with the parties that the thirteen agreed programs serve as

reasonable proxies for the level of need that the Court contemplates here. Each of the thirteen programs provides a well-established and widely accepted basis for identifying individuals who cannot obtain basic necessities without assistance, and eligibility for each program is determined, in part, by household income, with the qualifying income thresholds for these programs varying by relatively low multiples of either the Federal Poverty Guidelines, Local Poverty Level, or other comparable metrics. Id. While Terminals disputes the inclusion of the LIHWAP and the ERAP, it does not contend that the financial criteria necessary for persons to qualify for these two programs is inappropriate. Indeed, the financial eligibility requirements for both programs are comparable to the thirteen agreed programs jointly proposed by the parties. (See Dkt. No. 371-1 (Joint proof that they meet the program’s financial eligibility criteria, but rather may simply attest to their financial resources in lieu of documentary proof. (Dkt. No. 365 (LBT Supp. Br.) at 8). Terminals also objects to the LIHWAP—which “help[s] low-income households pay for their water bill”—on the grounds that households participating in the program “are necessarily

connected to another water source” and should be disqualified from participation in the water program on that basis. Id. at 7-8. Plaintiffs argue that Terminals’ objection to the inclusion of the ERAP is a “red herring” because “[t]he ERA[P]-related attestation is not being proposed as a standalone criterion for eligibility for the [Water] Program, it is simply the manner in which an applicant may apply to the underlying program.” (Dkt. No. 364 (Pls. Supp. Br.) at 6). The Court disagrees. As noted above, the parties’ Proposed Order provides that applicants seeking to demonstrate their economic eligibility for the water program may do so by submitting

documentary proof that they are “eligible” to receive benefits under one of the specified programs notwithstanding that they do not actually receive said benefits. (Dkt. No. 363-1 (Proposed Order) at § 3.D.ii). The parties’ Proposed Order further provides that applicants seeking to demonstrate eligibility in this fashion must submit “documentary proof that is consistent with the documentary proof required by the particular program for which eligibility is being claimed by the Applicant[.]” Id. at § 4.iii. It thus follows that, under the parties’ proposed scheme—which the Court adopts here7—an applicant could, if the ERAP

7 The Court notes that it previously expressed its doubts that the water program’s economic eligibility criteria should encompass those residents who “would be eligible” for the specified programs (pursuant to paragraph 3.D.ii of the parties’ Proposed Order)—as opposed to only were included, demonstrate economic eligibility for the water program simply by attestation, because such attestation is an acceptable means of proof for financial eligibility in the ERAP itself. Likewise, since residents may also demonstrate economic eligibility for the water program by showing that they actually participate in one of the specified programs, id. at §

3.D.i, it follows that including the ERAP in the list of qualifying programs would also permit residents to demonstrate economic eligibility for the water program merely by attestation if, in fact, attestation was the means through which the individual qualified for the ERAP. Terminals’ concern with so lax a requirement is well-founded and shared by the Court. The Court will therefore not include the ERAP as a means of qualifying for the water program. On the other hand, the Court does not view the bare fact that a resident receives water from another source, such as the Virgin Islands Water and Power Authority (“WAPA”), as sufficient grounds upon which to disqualify a resident from participation in the water program. The evidence before the Court indicates that WAPA water is prohibitively costly

for certain residents and that WAPA water is not an adequate substitute for cistern water because it “cannot be used for drinking or many household purposes like cooking, brushing teeth, or doing laundry.” (Dkt. No. 285 (First Phase Memo. Op.) at 27-28 (summarizing testimony of Carolyn Urgent)). Further, the fact that a resident is currently paying for WAPA water does not mean that the resident can afford to pay for water without trading off other basic necessities. Id. Indeed, the fact that the resident is already in a program designed to help with the payment of water bills suggests the opposite. Accordingly, the Court finds that Terminals’ objection to the inclusion of the LIHWAP is without merit, particularly in light of the program’s comparability to the thirteen agreed programs. The Court will therefore include the LIHWAP as a means of qualifying for the water program. 2. Unemployment Plaintiffs argue that residents should be deemed economically eligible for the water program where “neither the [resident] nor another related person in the same house is employed and at least one adult [in the household] is unemployed and [has looked] for work

[within the last four weeks]” because unemployment implies “low (if any) income[.]” (Dkt. No. 317 (Pls. Br.) at 13). In support of this proposition, Plaintiffs rely principally on the testimony of Dr. David Neumark,8 who opines that it is reasonable to assume “that the absence of employed adults, and the presence of at least one adult looking for work, implies that the family has no current earnings from work, and hence would likely have to trade off basic necessities in order to purchase water.” (Dkt. No. 317-8 (Neumark May 19, 2023 Dec.) at 9). The Court notes that Plaintiffs have proposed that a resident be deemed economically eligible under this criterion where the resident meets the definition of unemployment utilized

by the United States Bureau of Labor Statistics (“BLS”). (Dkt. No. 331 (Pls. Reply) at 14- 15). With one exception not relevant here, the BLS definition labels a person as unemployed only if that person is actively seeking out a new job. (Dkt. No. 331-2 (Neumark May 31, 2023 Dec.) at ¶ 8). In this regard, the Court agrees with Plaintiffs that their proposed criterion is not likely to improperly capture retirees, students, or seasonal workers within its scope, as Terminals has argued. (Dkt. No. 327 (LBT Resp.) at 15-16; Dkt. No. 331 (Pls. Reply) at 14- 15). At the same time, however, the BLS definition does not incorporate or reference a person’s financial means whatsoever.9 As Terminals correctly observes, it follows from this

8 The Court accepted Dr. Neumark as an expert in “poverty-related need-based programs,” definition that persons may be formally classified as unemployed while still possessing significant financial resources or even investment income. (Dkt. No. 322 (LBT Br.) at 16). Thus, the Court agrees with Terminals that “unemployment” as Plaintiffs define it here—by incorporating the BLS definition—“is simply not synonymous with being poor or unable to

provide for basic necessities.” (Dkt. No. 365 (LBT Supp. Br.) at 9 (emphasis in original)). Because Plaintiffs’ proposed unemployment criterion would include residents with significant financial resources, and because such residents do not fall within the class of persons for whom the Court contemplates relief, the Court will not include the unemployment criterion as a means of qualifying for the water program. 3. EITC Plaintiffs argue that a resident should be deemed economically eligible where the resident or a member of his or her household receives, or would be eligible to receive, the EITC. (Dkt. No. 331 (Pls. Reply) at 13-14). In essence, Plaintiffs’ argument for the inclusion

of this criterion reduces to the contention that the EITC is a tax credit geared toward low- and moderate-income workers, and that research “suggests that families mostly use the EITC to pay for necessities such as food and housing.” (Dkt. No. 331 (Pls. Reply) at 14; Dkt. No. 331- 2 (Neumark May 31, 2023 Dec.) at ¶¶ 12-13). Terminals, in contrast, argues that the Court should not include this factor because “it is a federal tax credit that is highly subjective because it changes depending on many different criteria, including household income and household size.” (Dkt. No. 374 (LBT Notice) at 2). The Court does not view the fact that the amount of credit to which households are entitled under the EITC varies based on household size and household income as particularly salient. Indeed, given that federal and local poverty levels vary with household size, eligibility for virtually all thirteen of the government assistance programs that the parties have jointly proposed also vary according to these same criteria. (See Dkt. No. 373 (Joint Notice) at ¶¶ 1-6). However, there is no evidence before the Court indicating that the population of low- and moderate-income workers—to whom the EITC allegedly is geared—

substantially overlaps with the population for whom the Court contemplates relief here. The bare fact that “research” indicates that families “mostly” use the EITC to pay for necessities does not militate against this conclusion. Accordingly, the Court will not include Plaintiffs’ proposed EITC criterion as a means of qualifying for the water program. 4. Poverty Line Multiple Plaintiffs argue that a resident should be deemed economically eligible where the resident’s household income is less than 150% of the Federal Poverty Guideline plus $11,680, which is Plaintiffs’ “estimated annual cost of water” based on an estimated cost of $1.60 per gallon and an estimated consumption of twenty gallons of water per day per

resident. (Dkt. No. 317 (Pls. Br.) at 10). The Court agrees with Terminals that this criterion is unnecessary given that the parties have agreed that not only residents who receive benefits under the specified government programs, but also those residents who “would be eligible” to receive benefits under the financial criteria of those government programs, will be economically eligible for the water program. (Dkt. No. 365 (LBT Supp Br.) at 9). In other words, the Court does not believe that an appreciable fraction of residents who cannot afford to purchase water without trading off other basic necessities would be captured under Plaintiffs’ poverty-line multiple criterion without simultaneously being captured under the other economic eligibility criteria that the Court has approved. Further, as Terminals correctly notes, Plaintiffs’ proposed poverty line multiple criterion is—unlike the parties’ joint demonstrates financial need.” Id. Accordingly, the Court will not include Plaintiffs’ proposed poverty line multiple criterion as a means of qualifying for the water program. 5. Proof of Economic Eligibility The parties’ Proposed Order provides that an applicant must submit “reasonable

proof” of his or her economic eligibility for the water program. (Dkt. No. 363-1 (Proposed Order) at § 3.D). The Proposed Order further provides that applicants may satisfy this requirement by submitting “documentary proof” of participation in a listed needs-based government program, and outlines the forms of “documentary proof” that applicants must provide to demonstrate that they receive benefits under a specified program. Id. at § 4.i. In substance, the parties’ proposed language specifies that “documentary proof” consists of “documents reasonably demonstrating admission into or participation in such program within the previous calendar year,” and provides examples of acceptable documentation. Id. In the case of applicants who do not participate in a listed program, but seek to establish their

eligibility for the water program by demonstrating that they “would be eligible” for a listed program, the Proposed Order requires “proof that is consistent with the documentary proof required by the particular program for which eligibility is being claimed by the Applicant[.]” Id. at § 4.iii. The Court adopts the parties’ language regarding documentary proof, notwithstanding that the language defining documentary proof for applicants seeking to establish that they “would be eligible” is not ideal given its lack of specificity.10 The Court takes the parties at their word insofar as they represent that they “have been researching the proof issue and believe that they, together with the Administrator [of the water program], will be able to [provide] notice [to residents as to] what documentary proof is required for each particular program.” (Dkt. No. 363 (Joint Notice) at 2 n.3). B. Geographic Eligibility Criteria

Throughout the second phase proceedings, Plaintiffs have continually revised the list of estates that they seek to have included in the “Covered Area”—that is, the geographic area that is covered by the Court’s Order Implementing Water Distribution Program. Plaintiffs currently propose that the Covered Area should include 59 estates. (Dkt. No. 363-1 (Proposed Order) at § 1.c). Terminals agrees that 20 of these estates should be included in the Covered Area, but disputes the inclusion of the remaining 39 estates. Id. Plaintiffs contend that six main sources of evidence generally support their proposed Covered Area. (Dkt. No. 253 (Pls. Br.) at 5-8; Dkt. No. 331 (Pls. Reply) at 5-8; Dkt. No. 364 (Pls. Supp. Br.) at 2-3). Specifically, Plaintiffs argue that their proposed Covered Area is consistent with, and supported by:

(1) the testimony of resident witnesses Julio Carino, Joyce Allen Murphy, Margaret Thompson, Carolyn Urgent, Marisela Webster, and Marver Antoine;11 (2) what Brian Moore12 calls the “Zone of Impact” identified by the EPA’s Emergency Order;13 (3) air modeling of the release events performed by Dr. Erno Sajo14 (“Sajo Air Modeling”);15 (4) data collected by Terminals and Refining following the first release event (“Limetree Data”);16

11 See Dkt. No. 285 (First Phase Memo. Op.) at 26-30 (summarizing resident witness testimony); Dkt. No. 364-2 (Pls. Map) at 2 (depicting resident witness testimony in blue). 12 The Court accepted Mr. Moore’s testimony, which involved “mapping geo-location data using geographic information systems.” 13 See Dkt. No. 317-2 (Brian Moore Dec.) at 3, 12 (depicting EPA “Zone of Impact” in lavender). 14 The Court accepted Dr. Sajo, Plaintiffs’ expert, as an expert in “aerosol dispersion air modeling” for purposes of these preliminary injunction proceedings only. (5) data collected by Sedgwick Claims Management Services, a consultant employed by Terminals and/or Refining, following the fourth release event (“Sedgwick Data”);17 and (6) data collected by Dr. Marco Kaltofen18 and Dr. Michael Henry19 following the fourth release event (“Kaltofen Data”).20

Terminals argues that these sources of evidence are either flawed and uninformative, or do not justify so broad a Covered Area as the one Plaintiffs propose here. (Dkt. No. 327 (LBT Resp.) at 3-9; Dkt. No. 365 (LBT Supp. Br.) at 1-7). The Court agrees in substantial part. 1. Six Sources of Evidence

a. Resident Witness Testimony

As an initial matter, Plaintiffs are correct that the resident witnesses’ testimony is “consistent” with Plaintiffs’ proposed Covered Area, insofar as each of the resident witnesses testified that their cistern was contaminated by the release event and each resident witness lives in an estate captured within Plaintiffs’ proposal.21 (See Dkt. No. 285 (First Phase Memo. Op.) at 26-30 (summarizing resident witness’ testimony)). However, the resident witnesses’ testimony applies only to a small fraction of the area that Plaintiffs seek to have included in the Covered Area. Accordingly, the informative value of the resident witnesses’ testimony is limited.

17 See Dkt. No. 364-1 (Pls. Summary of Sedgwick Data) at 2; Dkt. No. 364-2 (Pls. Map (depicting Sedgwick Data in pink) at 2; Dkt. No. 365-3 (LBT Summary of Sedgwick Data) at 2. 18 The Court accepted Dr. Kaltofen, Plaintiffs’ expert, as an expert in “environmental science,” “environmental engineering,” and the “fate and transport of petroleum and chemical releases” for purposes of these preliminary injunction proceedings only. 19 The Court accepted Dr. Henry, Plaintiffs’ expert, as an expert in “physical science and inspections” for purposes of these preliminary injunction proceedings only. 20 See Dkt. No. 285 (First Phase Memo. Op.) at 22-25 (summarizing the Kaltofen Data); Dkt. No. 364-2 (Pls. Map) at 2 (depicting Kaltofen Data in yellow and red). b. Limetree Data

Terminals and Refining employees collected the Limetree Data following the first release event. As Joyce Wakefield, an employee of Terminals/Refining explained, this process involved inspecting over 200 properties in Clifton Hill and sampling the cisterns of at least 160 residences. (See LBT Ex. 2-18 (February Flare Release Tracking Log) at 1). The cistern sampling revealed that at least half of the cisterns had been contaminated by the first release event. Id. Because this sampling was limited to Clifton Hill, the informative value of the Limetree Data is limited to the Clifton Hill estate. (See Dkt. No. 364-2 (Pls. Map) at 2 (depicting the Limetree Data in cyan); LBT Ex. 2-39 (February Flare Release Survey Information). c. Kaltofen Data

The Court discussed the Kaltofen Data in detail in its First Phase Memorandum Opinion. (See Dkt. No. 285 (First Phase Memo. Op.) at 22-25). The Court notes that the Kaltofen Data is comprised of a limited number of observations, and that these observations are principally confined to estates near the southwestern edge of the island. (See Dkt. No. 364-2 (Pls. Map) at 2 (depicting Kaltofen Data in yellow and red)). d. Sajo Air Modeling

Based on meteorological data and his expertise in aerosol dispersion air modeling, Dr. Sajo created models estimating the areas he believes to have likely been impacted by the release events. He concluded, with respect to the first release event, that “significant portions of the [] Southcentral and Northwest [subdistricts of St. Croix] were likely affected” as well as “the southern part of [the] Northcentral [subdistrict] and the northern part of [the] Southwest [subdistrict]. Frederiksted is also directly downwind, and may have been affected as well.”22 (Dkt. No. 322-6 (Sajo Dec.) at 4; id. at 5 (figure overlaying Dr. Sajo’s model of the first release event over a map of western St. Croix)). Similarly, Dr. Sajo concluded, with respect to the fourth release event, that “the downwind impact of the released chemicals likely included communities located at Southcentral (west of the release), Southwest,

Frederiksted, Northwest, and parts of Northcentral.” Id. at 7; see also id. at 8 (figure overlaying Dr. Sajo’s model of the fourth release event over a map of western St. Croix). Terminals argues that Dr. Sajo’s modeling “yields purely speculative results” because Dr. Sajo “conceded” that his “modeling does not depict conditions on the ground” and that “critical pieces of information were unavailable to him in performing his modeling, including the nature, duration, and amount of the Releases” and other information regarding “meteorological conditions and terrain[.]” (Dkt. No. 267 (LBT Supp. Br.) at 18). The Court does not agree that Dr. Sajo’s modeling is purely speculative. It is true that Dr. Sajo acknowledged that his models do not account for various

factors. (Mar. 3 Tr. at 41-44). Indeed, Dr. Sajo testified that he selected the relatively simple type of modeling that he performed because he did not have access to certain input variables, including the exact time and duration of the release events or the precise quantity of the materials released during the events. (Mar. 3 Tr. at 44-45). He testified that, as a result of his lack of access to these variables, the models he created are only accurate out to “6 miles or 10 kilometers” from the refinery, while more “sophisticated” modeling—which would have required him to input variables he did not have—would have permitted him to draw

22 The subdistricts of St. Croix are principally used for statistical purposes during censuses. Notwithstanding that Dr. Sajo frames his conclusions in terms of these subdistricts, a map depicting these subdistricts is not in evidence. The Court has confirmed that Dr. Sajo’s descriptions of his conclusions (in terms of subdistricts) are accurate insofar as they comport with the figures he has created overlaying his models over maps of St. Croix. See Census conclusions with respect to “longer range dispersions”—that is, to draw conclusions about the likely zone of impact past the six-mile limit of his models. Id. at 47-48. In other words, Dr. Sajo testified that, notwithstanding that his modeling is limited by his lack of variables, it provides a “reasonable estimate” of the likely area of impact within six-miles of the refinery,

but has nothing to say regarding the area past this six-mile limit. Id. at 4. Ultimately, Terminals has not provided the Court with any reason to doubt Dr. Sajo’s conclusion that his model is accurate out to its six-mile limit (“the Sajo Terminus”). (See Dkt. No. 364-2 (Pls. Map) at 2 (depicting Sajo model in orange, with the western boundary of the model reflecting the Sajo Terminus)). The Court therefore takes Dr. Sajo’s air modeling as evidence that the area falling inside the Sajo Terminus was impacted by the first and fourth release events. At the same time, the Court recognizes that Dr. Sajo’s models do not “depict conditions on the ground” (Dkt. No. 267 (LBT Supp. Br.) at 18), and further that, given the lack of input variables, Dr. Sajo’s models do not purport to speak to the area outside the Sajo

Terminus. e. Moore’s “EPA Zone of Impact”

On May 14, 2021, the EPA issued an emergency order requiring that the refinery cease operations. (Pls. Ex. 21 (EPA Emergency Order) at 32-42). The emergency order notes that “[a]ccording to [the] EPA[’s] review of [relevant] National Weather Service data for February through May 2021”—that is, the period from the first through the fourth release events—“the prevailing wind bl[ew] from the east and east-southeast to the west and west- northwest [such that] [w]ind direction prevalence indicates that communities located within Southcentral, Southwest, Frederiksted, Northwest, and portions of Northcentral [were] located downwind of the Facility [during this period].” Id. at 6-7. Mr. Moore created a map depicts the entire western half of the island as the “Zone of Impact” consistent with the EPA’s emergency order. Id. at 3 (map depicting the EPA “Zone of Impact” in lavender). Mr. Moore’s “Zone of Impact” is “consistent” with the EPA’s description of the relevant wind conditions during the four-month period that included the time of the first and

fourth release events.23 However, it offers nothing more. Unlike Dr. Sajo, Mr. Moore is not an expert in aerosol dispersion air modeling. Moreover, in light of the fact that the “Zone of Impact” does no more than reflect the EPA’s summary of some of the relevant wind conditions, Mr. Moore’s “Zone of Impact” is of little value in assessing the areas that may have been impacted by the release events. While the Court considers the relevant wind conditions during the release events—as they are described in the EPA’s emergency order and elsewhere—the Court assigns no special weight to the fact that an area falls inside or outside of Mr. Moore’s “Zone of Impact.” f. Sedgwick Data

In the immediate aftermath of the fourth release event, Refining and/or Terminals enlisted Sedgwick to survey neighborhoods potentially impacted by the event; determine which residences had been impacted; and extend settlement offers to residents whose residences had been impacted. The Sedgwick Data consists of observations regarding over 2,00024 properties west of the refinery, and it represents the most voluminous data set presently before the Court. The parties agree that the Sedgwick Data provides the most reliable and complete picture of the extent of contamination. (Dkt. No. 327 (LBT Resp.) at 9;

23 The EPA order does not specify the wind conditions during each of the first and fourth releases specifically, but rather only the prevailing winds during the four-month period between the first and fourth release events. Additional information regarding the wind conditions during the release events is in evidence. (See, e.g., LBT Exs. 2-14, 2-15 (Windrose Dkt. No. 364 (Pls. Supp. Br.) at 2-3). Nonetheless, the parties have registered a number of disagreements regarding the best method to interpret the data. (See Dkt. No. 364 (Pls. Supp. Br.) at 2-3; Dkt. No. 365 (LBT Supp. Br.) at 4-7). The Court addresses these disagreements before summarizing the Sedgwick Data.

Total Number of Properties that Sedgwick Determined to be Impacted. The parties’ first disagreement concerns the appropriate manner to interpret the number of properties Sedgwick found to be impacted by the release events. Sedgwick created a document summarizing its raw data in spreadsheet form. (See Dkt. No. 317-4 (“Sedgwick Raw Data”)).25 As relevant here, the spreadsheet identifies: (1) the address of properties that Sedgwick visited, id. at cols. 4 (“Address”), 6 (“Inspected Y/N”); (2) whether Sedgwick made an offer to settle with the property owner, id. at col. 7 (“Offer Y/N”); (3) whether the property owner’s claims were settled, id. at col. 8 (“Settled Y/N”); and (4) the amount of the settlement offer in a dollar figure, id. at col. 9 (“Amount”).

The parties’ dispute as to the interpretation of this data arises from the fact that the Sedgwick Raw Data exhibits two primary types of inconsistencies. First, for numerous entries, the spreadsheet specifies a settlement offer amount (in column nine) while simultaneously reflecting that no offer was made (in column seven). See, e.g., id. at 1, row 11. Second, for numerous entries, the spreadsheet reflects that no offer was made (in column seven) while simultaneously reflecting that the property owner’s claim was settled (in column eight), and neglecting to specify a settlement offer amount (in column nine). See, e.g., id. at 9, row 1 (entry with both forms of this type of inconsistency). Plaintiffs contend that, in tallying the number of properties that Sedgwick determined to be impacted in each estate—and, in particular, in comparing the number of impacted properties in each estate to the number of properties Sedgwick inspected in that estate to arrive at a “hit count”26 or portrait of the extent of the contamination within that estate—the Court should count as impacted all properties for which Sedgwick’s spreadsheet reflects that the owner received (column seven) or accepted (column eight) an offer, or lists a settlement

offer amount (column nine). (Dkt. No. 364 (Pls. Supp. Br.) at 2). Terminals contends, on the other hand, that doing so results in a “grossly misleading” hit count, and argues that the Court should count as impacted only those properties for which Sedgwick’s spreadsheet lists a settlement offer amount (column nine). (Dkt. No. 365 (LBT Supp. Br.) at 4). The Court agrees with Plaintiffs. The second type of inconsistency identified above makes it inappropriate to simply tally column nine (“Amount”) to arrive at an accurate total number of properties Sedgwick determined to be impacted. This is because doing so results in an underinclusive tally that neglects to count as impacted: (1) those properties for which Sedgwick represents that it made a settlement offer in column seven (“Offer Y/N”)—which

Sedgwick only did if it determined the property was impacted—but no settlement offer amount was stated; and (2) those properties for which Sedgwick represents that the owner’s claim was settled—which would have been the case only if an offer was made in the first instance because the property was identified as impacted—but again, no settlement offer amount was stated.27 In addition, interpreting the data in the way Plaintiffs urge here results

26 As used here, a “hit count” is the percentage of properties Sedgwick determined to be impacted within a given estate, calculated as the number of properties Sedgwick determined to be contaminated within that estate over the number of total inspections Sedgwick performed in that estate. 27 It is likewise impossible to reach an accurate result by individually tallying either column seven (“Offer Y/N”) or column eight (“Settled Y/N”). The first type of inconsistency identified above means that simply tallying column seven (“Offer Y/N”) results in an underinclusive count because doing so neglects to include those properties for which in figures that comport more closely with the testimony this Court received. Jeffery Charles, Terminals’ Chief Operating Officer, estimated that Sedgwick inspected between 2,300 and 2,400 properties, and that Sedgwick identified 2,200 of those properties as impacted (approximately 92%-96%). Plaintiffs’ interpretation likewise leads to the conclusion that

Sedgwick identified approximately 96% of the properties it inspected as impacted—that is, 1,947 impacted properties of 2,026 (or 2,027) inspections. (See Dkt. No. 364-1 (Pls. Summary of Sedgwick Data) at 2, col. 11). Terminals’ interpretation, in contrast, leads to the conclusion that Sedgwick identified 1,694 of the of 2,026 (or 2,027) properties it inspected as impacted, or approximately 84%. (Dkt. No. 365-3 (LBT Summary of Sedgwick Data) at 2, col. 4). The Court therefore considers Plaintiffs’ totals for the properties Sedgwick determined to be impacted when analyzing the Sedgwick Data, with one exception noted below. (See infra, n.29).

Plaintiffs’ “Lumping” of Sedgwick Data for Certain Estates. Terminals argues that, in five instances, Plaintiffs have artificially inflated the hit count of certain areas by improperly labelling multiple estates as a singular estate and providing the Court with only the hit count for the singular, “umbrella” estate. (Dkt. No. 365 (LBT Supp. Br.) at 6). Specifically, Terminals argues that Plaintiffs have improperly grouped estates as follows: (1) the estates of Frederiksted, Hesselberg, Sandy Point, Smithfield, and Two Brothers under the umbrella term of “Two Brothers”; (2) the estates of Campo Rico, Carlton Land, Good Hope, Hope, Whim, Ruans Bay, and Two Williams under the umbrella term of “Whim”; (3) the estates of Bethlehem Old Works, Castle Burke, Fredensborg, and Jealousy under the umbrella term of “Bethlehem Old Works”; (4) the estates of Betty’s Hope, Envy, Mannings Bay, Negro Bay, and “part of” Golden Grove under the umbrella term of “Airport”; (5) the estates of University of Virgin Islands Campus and Upper Bethlehem under the umbrella term of “Upper Bethlehem”; and (6) the estates of Mars Hill and Wheel of Fortune under the umbrella term of “Wheel of Fortune.”

(See Dkt. No. 365-6 (LBT Summary Chart) at 2). The Court agrees with Terminals that Plaintiffs’ presentation28 of the Sedgwick Data with respect to these 25 estates is misleading in important part. Plaintiffs appear to attempt to implicitly justify their presentation of these estates by noting that they have “corrected [the data set] to address common knowledge, vernacular, and phonetic concerns.” (Dkt. No. 364-1 (Pls. Summary of Sedgwick Data) at 2). The Court recognizes that the boundaries of St. Croix’s estates are not always precise and well defined, and that certain small estates are frequently referred to, colloquially, as though they were part of a larger neighboring estate. In this respect, the Court finds that the fifth and sixth of Plaintiffs’ umbrella terms, as identified above, are defensible. Plaintiffs’ third and fourth umbrella terms are less so, especially in light of Plaintiffs’ previous references to (for example) Jealousy and Golden Grove as individual estates in the course of arguing for their inclusion during these proceedings. (See, e.g., Dkt. No. 317-1 (Pls. May 23, 2023 Proposed Order) at § 1.c). Plaintiffs’ first and second umbrella terms, however, are simply indefensible. Frederiksted is one of St. Croix’s two main towns, and it is not referred to—colloquially or otherwise—as “Two Brothers,” which is a smaller, neighboring estate. Likewise, the estate of Hope is noncontiguous both from Whim and the six other estates Plaintiffs group under the umbrella term “Whim.” Moreover, Hope is made noncontiguous from these estates by other intervening estates for which Plaintiffs individually identify the relevant Sedgwick Data— namely, Carlton, Cane, and Williams Delight. Because of Plaintiffs’ unjustified umbrella groupings, the fact that Sedgwick performed few or no inspections in certain areas, or received only few “hits” for the inspections it performed, can be easily overlooked in Plaintiffs’ presentation of the data. For purposes of its analysis here, the Court will therefore consider the individual “hit rates” of the

Sedgwick Data for each of the 21 estates comprising Plaintiffs’ first through fourth umbrella terms.29 Estates Where Sedgwick Performed a Limited Number of Inspections. Terminals notes that it “disputes the wholesale inclusion of any estate where there have been fewer than seven inspections, offers, or offers in amounts above zero.” (Dkt. No. 365 (LBT Supp. Br.) at 5). The Court rejects this argument because the Court is not considering the Sedgwick Data alone in determining the inclusion or exclusion of any particular estate or portion thereof. The

29 Unlike Plaintiffs, Terminals subdivides the data by estate in the correct manner. However, Terminals’ presentation of this data raises a different problem. Terminals specifies the relevant totals only in terms of the number of properties that the Sedgwick Raw Data reflects received an offer (in column seven) in each estate, and in terms of the number of properties for which the Sedgwick Raw Data specifies a settlement offer amount (in column nine). (Dkt. No. 365-3 (LBT Summary of Sedgwick Data) at 2). In other words, Terminals’ presentation of the data for the subdivided estates does not specify the total number of properties within each estate for which the Sedgwick Raw Data reflects that there was either an offer made (in column seven) or a specified settlement offer amount (in column nine), nor does it specify the number of properties within each estate that received a settlement (in column eight). The Court therefore counts the highest of the two variables Terminals specifies for purposes of tallying the total number of residences impacted within these 21 areas, which is somewhat underinclusive, per the Court’s discussion above. Thus, the Court considers the following estates to have the following hit rates: Bethlehem Old Works (not lumped): 100% (2/2); Campo Rico: 89% (16/18); Frederiksted: 0% (0/1); Hope: 0% (0/1); Smithfield: 100% (36/36); Two Brothers (not lumped): 88% (14/16); and Whim (not lumped): 79% (440/556). Id. Likewise, the Court considers the following estates to have had no inspections performed: Betty’s Hope; Castle Burke; Envy; Fredensborg; Golden Grove; Good Hope; Hesselberg; Jealousy; Mannings Bay; Negro Bay; Ruans Bay; and Sandy Point. Id. (The Court notes that it considers Carlton Land and Two Williams to be part of Whim and Concordia, respectively, and will not treat Carlton Land and Two Williams as separate estates in which no inspections were performed.) Court agrees that the hit rates for estates with a relatively low number of inspections are of limited evidentiary value, and the Court considers the Sedgwick Data—in combination with all other relevant evidence—cognizant of the number of inspections performed. Summary of Sedgwick Data. Based on the Court’s interpretation of the Sedgwick

Data, as described above, the Court finds that the data shows that the impact of the release events was concentrated in an area that will be defined, very approximately, as an area bounded to the east by East Airport Road, to the north by Centerline Road, and to the west and south by the ocean. The Court will refer to this area as the “Sedgwick Concentrated Area.” The scarcity of major or straight roads on St. Croix makes this definition both somewhat underinclusive and overinclusive. As described below, this definition is overinclusive with respect to its eastern, southern, and western borders, and underinclusive with respect to its northern border. “Core Estates” in the Sedgwick Concentrated Area. The Sedgwick Concentrated

Area captures the following fifteen estates with both a high hit rate and a relatively high number of inspections: Adventure (100%, 34/3430), Cane (97%, 62/64), Cain Carlton (97%, 61/63), Campo Rico (89%, 16/18), Carlton 1 South (90%, 55/61), Carlton 2 (90%, 55/61), Hannah’s Rest (99%, 134/135), Diamond (96%, 74/77), Enfield Green (100%, 115/115), Mount Pleasant (98, 182/185), Paradise (100%, 7/7), Stony Ground (100%, 63/63), Whim (not lumped) (79%, 440/556), Whites Bay (100%, 70/70), and Williams Delight (92%, 325/352). Below, the Court refers to these fifteen estates as the “core” estates in the Sedgwick Concentrated Area.31

30 The first number in each of the parentheticals represents the hit rate expressed in percentage form. The second set of numbers represents the hit rate expressed as the number “Underinclusive Estates” in the Sedgwick Concentrated Area. The Court’s description of the Sedgwick Data is underinclusive with respect to its northern boundary insofar as it excludes estates north of Centerline Road with both a high hit rate and a relatively high number of inspections rendering that hit rate reliable in the first instance.

These estates, which cluster toward the western edge of the island, are: Carlton 1 North (90%, 55/61), Concordia (98%, 125/128), Smithfield (100%, 36/36), Two Brothers (not lumped) (88%, 14/16), and Wheel of Fortune/Mars Hill (100%, 7/7). Below, the Court refers to these five estates as the “underinclusive estates” in the Sedgwick Concentrated Area. “Overinclusive Estates” in the Sedgwick Concentrated Area. The Court’s description of the Sedgwick Data is overinclusive with respect to its western boundary insofar as it captures Sandy Point and Whites Bay 2, two estates in which Sedgwick performed no inspections. It is overinclusive with respect to its eastern boundary insofar as it captures Golden Grove and Upper Bethlehem, both of which are estates in which Sedgwick performed no inspections.32 It is overinclusive with respect to its southern boundary insofar

as it captures Betty’s Hope, Envy, Mannings Bay, and Negro Bay, four estates in which Sedgwick performed no inspections. It is overinclusive with respect to interior estates insofar as it captures Good Hope and Ruans Bay, two estates in which Sedgwick performed no inspections. Below, the Court refers to these ten estates as the “overinclusive estates” in the Sedgwick Concentrated Area.

north of Centerline Road. The Court does not refer to these five estates as “core estates,” but rather as “underinclusive estates.” 2. Reasonable Inferences from Foregoing Evidence

The Court finds that the force of the various sources of evidence described above leads to the conclusion that the appropriate Covered Area is an area roughly approximating, but not precisely equivalent to, the Sedgwick Concentrated Area. Core Estates. As an initial matter, the Court finds that the evidence supports including the fifteen core estates in the Sedgwick Concentrated Area within the Covered Area. First and foremost among this evidence is the Sedgwick Data itself, which reflects that the hit rates for each of these estates is exceptionally high and particularly reliable (given that the number of inspections yielding these rates is also high). The Sajo Air Modeling also supports including the nine (of fifteen) core estates that are east of, or on the threshold of, the Sajo Terminus, namely: Adventure, Cane, Cain Carlton, Carlton 1 South, Diamond, Enfield Green, Mount Pleasant, Paradise, and Williams Delight. Additionally, the Kaltofen Data supports including the six core estates that are west of the Sajo Terminus—Campo Rico,

Carlton 2, Hannah’s Rest, Stony Ground, Whim (not lumped), and Whites Bay—because the Kaltofen Data reflects contamination within these six estates or within estates immediately proximate to them. The resident witness testimony also provides limited support for the inclusion of these six estates.33 Overinclusive Estates: Western Boundary. With respect to the two overinclusive estates at the western boundary—Sandy Point and Whites Bay 2—the Court finds that the evidence supports including these two estates within the Covered Area. Both Sandy Point and Whites Bay 2 are immediately proximate to estates with exceptionally high hit rates that are particularly reliable—namely, Hannah’s Rest (99%, 134/135), Smithfield (100%, 36/36), Stony Ground (100%, 63/63), Two Brothers (not lumped) (88%, 14/16), Wheel of Fortune/Mars Hill (100%, 7/7), and White’s Bay (100%, 70/70). The inclusion of Sandy Point and Whites Bay 2 also finds support in the Kaltofen Data, which reflects contamination within estates immediately proximate to Sandy Point and Whites Bay 2, as well as limited

support in the resident witness testimony. It also finds support in the testimony of Roland Riviere, a Sedgwick employee involved in the inspections, who testified that Sedgwick ceased performing inspections before completing inspections of the western edge of the island due to non-payment of its contract by Terminals and/or Refining, notwithstanding that Sedgwick was still receiving reports of contamination from residents at this time. Overinclusive Estates: Interior. With respect to the two overinclusive estates in the interior of the Sedgwick Concentrated Area—Good Hope and Ruans Bay—the Court finds that the evidence supports including these two estates within the Covered Area. Both Good Hope and Ruans Bay are immediately proximate to one another, and are surrounded by

estates with exceptionally high hit rates that are particularly reliable—namely, Cane (97%, 62/64), Whim (not lumped) (79%, 440/556), and Williams Delight (92%, 325/352). Moreover, both properties are on the threshold of the Sajo Terminus, and there are numerous core estates west—that is, farther from the refinery—of both estates. The same is true for Paradise, the sole core estate that Terminals disputes.34 Paradise is surrounded by estates with exceptionally high hit rates that are particularly reliable—namely, Adventure (100%, 34/34), Diamond (96%, 74/77), and Mount Pleasant (98%, 182/185)—and there are numerous core estates to its west. The inclusion of Paradise is also strongly supported by the Sajo Air Modeling given its proximity to the refinery. Overinclusive Estates: Southern and Eastern Boundaries. With respect to the two overinclusive estates at the eastern boundary—Golden Grove and Upper Bethlehem35—and the four overinclusive estates at the southern boundary—Betty’s Hope, Envy, Mannings Bay, and Negro Bay—the Court finds that the evidence supports including these six estates within

the Covered Area. These estates are immediately proximate to one another and to estates with exceptionally high hit rates that are particularly reliable—namely, Adventure (100%,34/34), Diamond (96%, 74/77), Enfield Green (100%, 115/115), and Mount Pleasant (98%, 182/185).36 Moreover, there are numerous core estates to the west of these estates, and the inclusion of these estates is strongly supported by the Sajo Air Modeling given their proximity to the refinery. Underinclusive Estates. With respect to the five underinclusive estates, which cluster to the north of Centerline Road at the western edge of the island—Carlton 1 North, Concordia, Smithfield, Two Brothers (not lumped), and Wheel of Fortune/Mars Hill—the

Court finds that the evidence supports including these estates within the Covered Area. First and foremost among this evidence is the Sedgwick Data itself, which reflects that the hit rates for each of these estates is exceptionally high and particularly reliable. The Kaltofen Data likewise supports the inclusion of these estates, because it reflects contamination within these five estates or within estates immediately proximate to them. The resident witness testimony also provides limited support for the inclusion of these five estates.37

35 The Court notes that, in referring to Upper Bethlehem here, it designates the northwest quadrant of Golden Grove shaded in pink in Plaintiffs’ map, not the estate labelled “Upper Bethlehem” on that map. (See Dkt. No. 264-2 (Pls. Map) at 2). 36 These estates are also immediately proximate to Paradise (100%, 7/7), which has an exceptionally high hit rate but is not as reliable as the other core estates in light of the relatively lower number of inspections within it. Other Estates. The Court now addresses those estates that were not included in its initial definition of the Sedgwick Concentrated Area. Estates East of East Airport Road. With the exception of Clifton Hill, discussed below, Plaintiffs have not provided sufficient evidence for the Court to conclude that any of

Plaintiffs’ proposed estates to the east of East Airport Road should be included in the Covered Area. The estates in this area are supported by the Sajo Air Modeling given their proximity to the refinery, and Plaintiffs also contend that their inclusion is warranted in light of the Sedgwick Data. However, given, as noted above, that the Sajo Air Modeling does not reflect conditions on the ground, the Court does not view the Sajo Air Modeling as sufficient to justify the inclusion of any estates without adequate corroborating data collected from the ground. Such corroborating data is absent here. The Kaltofen Data does not corroborate the Sajo Air Modeling with respect to these estates, nor—contrary to Plaintiffs’ contention—does the Sedgwick Data. Indeed, Sedgwick performed no inspections in the vast majority of

estates in this area. Moreover, for the six estates in this area in which Sedgwick performed inspections, it performed very few.38 Given that these estates are not proximate to any estates with both high hit rates and a high number of inspections, the weight of the evidence is insufficient to justify their inclusion. Clifton Hill. Clifton Hill, as previously noted, is an exception. The Limetree Data reflects a substantial degree of contamination within Clifton Hill. (LBT Ex. 2-18 (February Flare Release Tracking Log) at 1.). While Limetree undertook remediation efforts in Clifton

on the fact that—like Paradise—its 100% hit rate is the result of only seven inspections, the Court’s response is functionally the same as it was for Paradise—Wheel of Fortune/Mars Hill is immediately proximate to numerous estates with exceptionally high hit rates that are particularly reliable. Hill following the first release event, these efforts were incomplete according to Ms. Wakefield. Accordingly, the Court will include the Clifton Hill estate within the Covered Area. Estates North of Centerline Road and West of East Airport Road. With the

exception of the overinclusive estates, discussed above, the one core estate North of Centerline Road—Carlton 1 North—and the four estates discussed below, Plaintiffs have not provided sufficient evidence for the Court to conclude that any of Plaintiffs’ proposed estates north of Centerline Road and west of East Airport Road should be included within the Covered Area. Again, Sedgwick performed no inspections in the vast majority of estates in this area and, for the four estates in this area in which Sedgwick performed inspections, it performed very few.39 The Court is cognizant that a number of the estates in this area are supported by the Sajo Air Modeling, and that Grove Place is supported by two samples in the Kaltofen Data. However, none of this evidence is sufficient to justify inclusion in light of the

overall portrait painted by the evidence. Frederiksted, Hesselberg, Hogensborg, and La Grange. Sedgwick performed no inspections in the Hesselberg and La Grange estates, one inspection leading to a non-detect in Frederiksted, and four inspections leading to detects in Hogensborg. However, these estates are immediately proximate to one another and surrounded by estates with exceptionally high hit rates that are particularly reliable—namely, Concordia (98%, 125/128), Cane (97%, 62/64), Carlton 1 North (90%, 55/61), Hannah’s Rest (99%, 134/135), Smithfield (100%, 36/36), Stony Ground (100%, 63/63), Two Brothers (not lumped) (88%, 14/16), Wheel of Fortune/Mars Hill (100%, 7/7), Williams Delight (92%, 325/352), and White’s Bay (100%, 70/70). The inclusion of these four estates also finds support in the Kaltofen Data, which reflects contamination within estates immediately proximate to these estates, as well as limited support in the resident witness testimony. It also finds support in the testimony of Mr. Riviere, who testified that Sedgwick ceased performing inspections before completing inspections of the western edge of the island due to non-payment of its contract by Terminals

and/or Refining, notwithstanding that Sedgwick was still receiving reports of contamination from residents in and around these estates at the time. The Court concludes that the weight of this evidence warrants inclusion of Frederiksted, Hesselberg, Hogensborg, and La Grange within the Covered Area. 3. Final Geographic Eligibility Criteria

Based on the Court’s analysis in Section 2 above, the following estates are included in the Covered Area for purposes of determining geographic eligibility: Adventure, Betty’s Hope, Cane, Cain Carlton, Campo Rico, Carlton 1 North, Carlton 1 South, Carlton 2, Clifton Hill, Concordia, Diamond, Enfield Green, Envy, Frederiksted, Golden Grove, Good Hope, Hannah’s Rest, Hesselberg, Hogensborg, La Grange, Mannings Bay, Mount Pleasant, Negro Bay, Paradise, Ruans Bay, Sandy Point, Smithfield, Stony Ground, Two Brothers, Upper Bethlehem, Wheel of Fortune/Mars Hill, Whim, Whites Bay, Whites Bay 2, and Williams Delight. While the Court has defined the Covered Area in terms of estates above, the Court recognizes that labeling and defining the precise contours of estates on St. Croix is difficult. Accordingly, the Court includes the following equivalent description: Together with Carlton 1 North, Clifton Hill, and Hogensborg:

(A) Residences that are: (1) west of East Airport Road, and (2) south of Centerline Road; and (B) Residences that are (1) south of Mahogany Road, and (2) would be west of Whim Road if Whim Road was extended consistent with its current path up to Mahogany Road.40

4. Proof of Geographic Eligibility

The parties’ Proposed Order provides that an applicant may demonstrate his or her geographic eligibility for the water program by submitting an identification document (such as a driver’s license) and submitting either “reasonable proof of occupancy” within the Covered Area (such as a copy of a lease) or an attestation to the effect that he or she resides in the Covered Area. ((Dkt. No. 363-1 (Proposed Order) at §§ 3.A.i-iii). As previously expressed, the Court is unwilling to condition applicants’ demonstration of eligibility on so lax a requirement as attestation alone. Accordingly, the Court will require that applicants submit both “reasonable proof of occupancy” and an attestation. C. Disputed Logistical Issues

1. Quantity of Water

The parties dispute the amount of water to which each eligible household should be entitled under the program. Plaintiffs suggest that eligible households should be entitled to twenty gallons of water per day, equivalent to 140 gallons of water per week. (Dkt. No. 364 at 7). Plaintiffs argue that their request is conservative in light of the fact that the United Nations has cautioned that thirteen to 26 gallons of water per day are “needed to ensure that most basic needs are met and few health concerns arise”; that the EPA has estimated that Americans use an average of 82 gallons of water per day; and that the Virgin Islands Department of Planning and National Resources has found that Virgin Islanders with cisterns use an average of 30 gallons of water per day. (See Dkt. No. 117 at 14 (collecting sources); Dkt. No. 164 at 7 (same)). In contrast, Terminals suggests that eligible households should be entitled to four gallons of water per day, with a cap of twenty gallons per week. (Dkt. No. 265 at 10-11). Terminals grounds its proposed figures in the pre-existing water distribution program established by consent of the parties during earlier bankruptcy proceedings, arguing that residents were “satisfied” with this amount of water under the pre-existing program.41

(Dkt. No. 322 (LBT Resp.) at 18). The Court does not believe that the amount of water Terminals proposes here is reasonable. Contrary to Terminals’ assertion, it does not appear that the twenty gallons of water per week afforded to residents under the pre-existing water program provided residents with enough water to meet their daily needs. Multiple residents who participated in that program testified that twenty gallons of water per week was not enough to meet their needs. (See Dkt. No. 264 (Mar. 2 Tr.) at 53 (JoAnne Allen-Murphy); id. at 302 (Carolyn Urgent); Dkt. No. 265 (Mar. 3 Tr.) at 102 (Julio Carino)). At the same time, however, Plaintiffs’ proposal appears excessive. The purpose of this preliminary injunction—like all preliminary

injunctions—is to prevent irreparable harm, not all harm. To the extent that certain households may choose to supplement their water needs over and above the amount provided under the program, the Court is of the view that such expenses will not irreparably harm eligible residents.

41 While Plaintiffs and various debtors associated with the refinery attempted to mediate an agreement in the Bankruptcy Court, the debtors agreed to expand a pre-existing program providing free bottled water to residents during the pendency of the mediation. (See LBT Ex. 5 (Stipulation and Agreed Order Adopting Water Distribution Program, dated October 29, 2021) at 3). Under the expanded program, the debtors agreed to provide free bottled water to the approximately 20,000 persons residing in the Frederiksted, Southcentral, Northcentral, and Southwest subdistricts of St. Croix, subject to limits on the amount of water individual residents could receive and a monthly cap on total expenses, among other conditions. Id. at 2- 5. Both the expanded program and its predecessor program ended in September 2022—days after Plaintiffs terminated the Bankruptcy Court mediation—and this Court subsequently returned the four associated cases to the active trial docket. Although not a debtor in the Bankruptcy Court action, Terminals was involved in establishing the predecessor program Balancing these considerations, the Court concludes that twelve gallons of water per day per household is reasonable, with a weekly cap of 84 gallons per household. 2. Five-Gallon Water Containers and Related Issues

The parties have agreed that, to prevent undue environmental impact on the island, Terminals shall use commercially reasonable best efforts to distribute five-gallon sized reusable water containers, rather than one-gallon plastic containers. (Dkt. No. 363-1 (Proposed Order) at § 17). Plaintiffs additionally urge the Court to order Terminals to provide each resident with a “water pump” for five-gallon water containers. Id. Terminals objects to this requirement. (Dkt. No. 363 at 4). The Court does not agree with the parties’ joint proposal to the extent that it requires the distribution of five-gallon containers if they can be distributed through commercially reasonable best efforts. The weight and size of five-gallon containers will undoubtedly make their transport and use cumbersome and potentially impossible for some residents. As the

same time however, the Court—like the parties—is cognizant of the need to prevent undue environmental impact to the island to the greatest extent practicable. Accordingly, the Court will order Terminals to use commercially reasonable best efforts to distribute five-gallon reusable water containers, consistent with the needs of residents. With regard to the water pump, the parties first apprised the Court of this issue in their Joint Notice following the second phase hearing, id., and they have not offered any evidence addressing the necessity or cost of water pumps. In light of Plaintiffs’ failure to present any evidence in this regard, the Court is unwilling to require Terminals to bear the cost of providing these pumps to eligible residents. At the same time, however, the Court notes that such pumps may be required for residents to make use of any five-gallon containers that at cost to those residents who wish to purchase them. Terminals will not be required to provide the pumps free of charge. 3. Waivers

Terminals argues that any residents who have had their property remediated or have signed waivers of claims following the first or fourth release event should be disqualified from the water program. (Dkt. No. 365 (LBT Supp. Br.) at 7). Plaintiffs disagree, arguing that the waivers are “invalid.” (Dkt. No. 364 (Pls. Supp. Br.) at 6). Plaintiffs further argue that the relevant remediation efforts, and at least some of the waivers at issue, were executed following the first release event but prior to the fourth release event; therefore, residents who were impacted by the fourth release event would be improperly excluded from the program if the Court adopts Terminals’ proposal here. Id. The parties have not presented sufficient evidence regarding the waivers at issue, the circumstances surrounding the execution of the waivers, or any remediation efforts

undertaken by Terminals or Refining. Further, the complexity of this issue is such as to require a full presentation by the parties. Because this issue has not been properly presented to the Court, the Court will not exclude otherwise eligible residents from the program based on the record before it. 4. Recertification

The parties disagree as to the frequency with which eligible residents should be required to submit documentation establishing their continued eligibility for the program. Plaintiffs argue that eligible residents should be required to recertify for the program annually; Terminals argues that eligible residents should be required to recertify for the program every six months. (Dkt. No. 363-1 (Proposed Order) at § 25). While the geographic location of one’s residence may be relatively stable, one’s financial circumstances may be more fluid. Accordingly, to help ensure that only eligible residents participate in the program, the Court will require recertification every six months. 5. Attestation Requirements

As set forth in the accompanying Order Implementing Water Distribution Program, residents seeking to qualify for the program must attest: (1) that they currently reside in the “Covered Area”; (2) that they resided in the Covered Area at the time of at least one of the first or fourth release events; (3) that they believe their current residence was impacted by one or both of the first or fourth release events; (4) that their current residence actively uses cistern water or used the same at the time of the first or fourth release events; and (5) that they meet at least one of the economic eligibility criteria. Terminals argues that residents should additionally be required to attest that they do not “receive potable water from any other source” (Dkt. No. 363-1 (Proposed Order) at § 3.B), while Plaintiffs argue that residents

should additionally be required to attest that they do not “receive free potable water from any other source,” id. (emphasis added). The Court does not see value in either of the parties’ proposed additions. Accordingly, the Court will not require residents to attest to either of the parties’ proposals. D. Special Master

The parties’ Proposed Order reflects that the parties have consented to the appointment of a Special Master whose costs will be shared equally by the parties. (Dkt. No. 363-1 (Proposed Order) at §§ 26-27). Per the parties’ agreement, the Special Master shall have “the initial authority to decide” any “Eligibility Appeals”42 and “Home Delivery Appeals,”43 as well as the “authority to meet separately and together with the Parties to facilitate communications, [and to] act[] as a facilitator of communications between the Parties on any issues relating to Eligibility Appeals, or Home Delivery Appeals.” (Dkt. No. 363-1 (Proposed Order) at §§ 26-27). The Proposed Order further provides that “[a]ny

decisions on Eligibility Appeals, Home Delivery Appeals, or Unreasonable Appeals44 rendered by the Special Master may be appealed to the United States Magistrate Judge for the District Court of the Virgin Islands within 14 days of any such decision,” and that the decision of the Magistrate Judge “relating to Eligibility Appeals and Home Delivery Appeals, or Unreasonable Appeals will be final and shall not be further appealable.” Id. at § 16. The Court is of the view that the authority of the Special Master should extend more broadly than the parties’ Proposed Order provides, to include, for example, appeals of routine, day to day actions of the Administrator, or routine, day to day operations of the water program. Accordingly, the Court has left room for further discussion with the parties on this

subject by including language in the Order Implementing Water Distribution Program that, together with Eligibility, Home Delivery, and Unreasonable Appeals, the Special Master’s responsibilities will also include “any other appeals over which the Special Master may be granted authority by agreement of the parties and approval of the Court.” E. Security

Rule 65(c) of the Federal Rules of Civil Procedure provides that a court “may issue a preliminary injunction or a temporary restraining order only if the movant gives security in an amount that the court considers proper to pay the costs and damages sustained by any party found to have been wrongfully enjoined or restrained.” Fed. R. Civ. P. 65(c). The posting of

43 “Home Delivery Appeals” are defined at § 18 of the parties’ Proposed Order, and involve adequate security is a “condition precedent” to injunctive relief, Hopkins v. Wallin, 179 F.2d 136, 137 (3d Cir. 1949), and the amount of security required is left to the discretion of the trial judge, Scanvec Amiable Ltd. v. Chang, 80 F. App’x 171, 178 (3d Cir. 2003). Once the Court has set the bond, “[t]he applicant then decides whether to accept the preliminary relief

by posting the bond or to withdraw its request.” Sprint Commc’ns Co. L.P. v. CAT Commc’ns Int’l, Inc., 335 F.3d 235, 240 (3d Cir. 2003). Here, Plaintiffs request that the Court waive the bond requirement or impose a nominal bond of $250. (Dkt. No. 144 (Pls. Supp. Br.) at 2). In the alternative, Plaintiffs request that the Court initially impose a bond of $50,000, with this figure to be “adjusted as needed” at a rate of $50 per additional household if “materially” more than 1,000 households participate in the water program. (Dkt. No. 372 (Pls. Not.) at 3). Terminals requests that the Court impose a bond equivalent to the “total annual cost to Terminals to run the [water program],” with the bond to be increased each year that the program remains operational.

(Dkt. No. 323 (LBT Supp. Br.) at 11). As an alternative to the security requirement, Terminals proposes that the Court impose a “testing protocol” that requires each eligible household participating in the water program to pay a sum equivalent to the cost of testing its cistern for contamination, with the household disqualified from participation in the program and the sum forfeited where testing reveals no detectable contamination. Id. The Court begins with the issue of waiver. As the Third Circuit has observed, “important policies undergird[] a strict application of the bond requirement in most injunction granting contexts,” and Rule 65(c) “admits of no exceptions” on its face. Hoxworth v. Blinder, Robinson & Co., 903 F.2d 186, 210 (3d Cir. 1990); see also W.R. Grace & Co. v. Local 759, Int’l Union of United Rubber, Cork, Linoleum & Plastic Workers, 461 U.S. 757, Freight Co. v. C.F. Air Freight, Inc., 882 F.2d 797, 804 (3d Cir. 1989) (noting that the bond requirement “deters rash applications for interlocutory orders,” and that “the bond premium and the chance of liability on it causes plaintiffs to think carefully beforehand”). Accordingly, the Third Circuit has “interpreted the bond requirement very strictly.” Hoxworth, 903 F.2d at

210. “While there are exceptions, the instances in which a bond may not be required are so rare that the requirement is almost mandatory,” Frank’s GMC Truck Center, Inc. v. Gen. Motors Corp., 847 F.2d 100, 110 (3d Cir. 1988). Plaintiffs argue that the conditions necessary to justify one of these “rare” exceptions obtain here—specifically, the exception recognized by the Third Circuit in Temple University v. White. (Dkt. No. 364 (Pls. Supp. Br.) at 9-10 (citing 941 F.2d 201 (3d Cir. 1991))). As the Third Circuit explained in a later case: In Temple University v. White, we explicitly recognized an exception to the Rule 65(c) bond requirement. . . [I]n Temple University we determined that, ‘at least in noncommercial cases, the court should consider the possible loss to the enjoined party together with the hardship that a bond requirement would impose on the applicant.’ Thus, the Temple University exception involves a balance of the equities of the potential hardships that each party would suffer as a result of a preliminary injunction. Where the balance of these equities weighs overwhelmingly in favor of the party seeking the injunction, a district court has the discretion to waive the Rule 65(c) bond requirement.

Elliott v. Kiesewetter, 98 F.3d 47, 60 (3d Cir. 1996) (quoting Temple, 941 F.2d at 219) (internal citations omitted). “This exception [is] narrow and may only be invoked by the District Court upon specific findings ‘regarding the relative hardships to each party.’” Scanvec Amiable Ltd. v. Chang, 80 F. App’x 171, 75 (3d Cir. 2003) (quoting Elliott, 98 F.3d at 60). In balancing the hardships, a district court should consider “the possible loss to the enjoined party”; “the hardship that a bond requirement would impose on the applicant”; and or public interests], in order to prevent undue restriction of [such rights and interests].” Temple, 941 F.2d at 219-20. Terminals does not expressly contest that the conditions necessary to trigger the Temple exception are present here. Instead, Terminals argues that Temple is no longer good

law, characterizing the decisions as “stale” in light of the Third Circuit’s subsequent decision in Zambelli Fireworks Manufacturing Company v. Wood. (Dkt. No. 323 (LBT Supp.) at 1 (citing 592 F.3d 412 (3d Cir. 2010))). More specifically, Terminals argues that, in Zambelli, the Third Circuit “made clear [that the] bond requirement may only be waived in the ‘exceptionally narrow circumstance where the nature of the action necessarily precludes any monetary harm to the defendant.’” Id. (quoting Zambelli, 592 F.3d at 426). Zambelli contains language that can certainly be interpreted—as Terminals interprets it here—to limit the circumstances under which the bond requirement may be waived to those in which a defendant will suffer no monetary harm as a result of the injunction. See Zambelli,

592 F.3d at 426 (“We therefore hold that a district court lacks discretion under Rule 65(c) to waive a bond requirement except in the exceptionally narrow circumstance where the nature of the action necessarily precludes any monetary harm to the defendant, and that such bond shall be issued irrespective of any request by the parties.”). However, a number of considerations suggest that it is incorrect to read Temple as “stale” in light of Zambelli. First, Zambelli does not explicitly present as abrogating Temple. Indeed, the Third Circuit acknowledged the existence of the Temple exception only sentences before articulating its holding. See id. at 427 (“We have never excused a District Court from requiring a bond where an injunction prevents commercial, money-making activities. Rather, we have recognized exceptions in other contexts only where ‘the balance of [the] equities Temple, 941 F.2d at 219 n. 26)). Second, Zambelli announces its holding only after first finding that the Temple exception was inapplicable under the circumstances. Id. (“The District Court made no such ‘specific finding’ here in support of its waiver of the requirements of Rule 65(c).”). Third, a panel of the Third Circuit has cited the Temple

exception as good law in at least one (unpublished) decision following Zambelli. See PharMethod, Inc. v. Caserta, 382 F. App’x 214, 222 (3d Cir. 2010) (vacating preliminary injunction where district court did not make specific findings as to applicability of Temple exception and requiring that the district court “address and resolve this tension if it refuses to require a bond on this basis” on remand). District courts in this circuit have also interpreted Temple as good law after Zambelli.45 The Court therefore finds that the Temple exception remains good law. The Court also notes that this case has attributes of the nature that could arguably place it within the class of “certain narrowly drawn circumstances” in which the Third Circuit deemed a waiver of

security to be available in Temple. Temple, 941 F.2d 201 at 219. For the reasons expressed in the Court’s First Phase Memorandum Opinion, the Court does not doubt that the issuance of this injunction is in the public interest or that the balance of the hardships tips overwhelmingly in favor of Plaintiffs here. (See Dkt. No. 285 (First Phase Memo. Op.) at 47-

45 As Terminals notes, many of these district court decisions are consistent with Zambelli insofar as they conclude that the Temple exception warrants waiver of the bond requirement while also finding that the enjoined party would not suffer monetary harm as a result of the injunction. See, e.g., Marshall v. Amuso, 571 F. Supp. 3d 412, 430 (E.D. Pa. 2021) (concluding that the Temple exception warranted waiver of bond requirement where “[t]here is no evidence of risk of monetary loss to [the defendant]” from compliance with preliminary injunction); Gilliam v. United States Dep’t of Agric., 486 F. Supp. 3d 856, 882 (E.D. Pa. 2020) (similar). However, the district court decisions applying the Temple exception post- Zambelli are not limited to such scenarios. See, e.g., Doxzon v. Dep’t of Hum. Servs., No. 20- CV-00236, 2020 WL 3989651, at *12 (M.D. Pa. Jul. 15, 2020) (waiving bond requirement pursuant to Temple exception notwithstanding that “the preliminary injunction will impose 51). Further, this injunction does not “prevent[] commercial, money-making activities,” Zambelli, 592 F.3d at 147, and the Plaintiffs are indigent. (See Dkt. No. 285 (First Phase Memo. Op.) at 40-47 (explaining that the Court will afford preliminary injunctive relief only to those Plaintiffs and putative class members who “cannot afford to purchase water without

trading off basic necessities”)). Notwithstanding the foregoing, the Court does not agree that the Temple exception to the bond requirement should be employed here, particularly in light of Plaintiffs’ recent offer to post a considerable sum notwithstanding their indigency. (See Dkt. No. 372 (Pls. Not.) at 3 (proposing that the Court impose a bond of $50,000 to be “adjusted as needed”). Indeed, the Court’s concern that a waiver of the bond requirement may not be appropriate in this instance is reflected in the Court’s numerous requests for additional briefing on this issue.46 Pursuant to the Court’s direction, a significant portion of the parties’ responsive briefing has been dedicated to exploring the logistics and feasibility of Terminals’ proposal of substituting a “testing protocol” for a standard bond requirement.47 Having reviewed the

parties’ briefs on this issue, the Court has ultimately concluded that Terminals’ proposal should not be adopted.48 At least four intractable issues counsel against adopting Terminals’ proposal here.

46 See, e.g., Dkt. No. 136 (February 2, 2023 Order) at 3 (directing Plaintiffs to submit a supplemental brief on the security issue); Dkt. No. 289 (May 8, 2023 Order) at 2 (directing Terminals to submit a supplemental brief on the security issue); Dkt. No. 367 (June 30, 2023 Order) at 5 (directing the parties to submit additional briefing on the security issue). 47 See Dkt. No. 372 (Plaintiffs’ Notice Regarding Proposed Testing Protocol); Dkt. No. 375 (Joint Notice Regarding Proposed Testing Protocol); Dkt. No. 376 (Terminals’ Notice Regarding Proposed Testing Protocol); cf. Dkt. No. 144 (Plaintiffs’ Supplemental Brief Regarding the Security Requirement); Dkt. No. 323 (Terminals’ Supplemental Brief Regarding the Security Requirement)). 48 The Court would be remiss not to note that Terminals initially made this creative First, and perhaps most obviously, the testing protocol does not function as a true alternative to the bond requirement. In particular, Terminals contemplates that the costs associated with sampling and testing cisterns under the protocol would be passed on to the beneficiaries of the program, only a fraction of which will be Plaintiffs. Unlike the imposition

of a “conventional” bond, the imposition of a testing protocol here would not serve to “deter[] rash applications for interlocutory orders” and lead Plaintiffs “to think carefully beforehand[.]” Instant Air Freight Co., 882 F.2d at 804. Nor is it clear that the imposition of a testing protocol would result in any notable sum that could conceivably revert to Terminals if it were ultimately determined that Terminals had been wrongfully enjoined. Second, notwithstanding that the parties have been able to reach agreement on at least some of the tests that they believe would be appropriate, imposing the costs associated with some or all of those tests—or even a substantial fraction of those costs—on the indigent population afforded relief by this injunction is untenable and would ultimately prove fatal to

the program. Third, the implementation of a testing protocol represents an exceptionally complicated endeavor—starting with the tests to be utilized—as reflected in the parties’ inability to reach agreement on several core aspects of a proposed protocol. Simply stated, the Court is not convinced that the implementation or maintenance of a testing protocol is administratively feasible, and is concerned that a testing protocol would be a recipe for endless tangential litigation. Fourth—and perhaps most importantly—the Court is concerned that households that, as a result of any testing protocol, receive a result that disqualifies them from the program would falsely believe that said result means that the Court has determined that their cistern especially salient by the fact that the Court is not presently in a position to determine the scientific reliability of any particular testing regime, or to determine to what extent any such regime would or could ultimately establish the safety (or not) of the sampled water. Consequently, the Court is reluctant to sanction any particular protocol, and is especially

concerned that the implementation of any testing protocol could lead to unintended consequences for the very population that this injunction is intended to protect and benefit. In view of the foregoing, the Court will not implement a testing protocol here. Instead, the Court will require Plaintiffs to post an initial bond of $50,000. Further, the Court will require that the parties file a joint notice every six months to update the Court as to the number of households participating in the program, and will require Plaintiffs to post an additional $50 for each household that participates in the program over and above 1000 households. The Court believes this figure to be appropriate notwithstanding the Plaintiffs’ indigency in light of the need to make Plaintiffs “think carefully” before accepting

preliminary relief of so substantial a magnitude. Further, while the bond requirement the Court imposes here is certainly minimal relative to the anticipated costs associated with the water program, the Court believes it is more appropriate than an outright waiver of the bond under Temple. Cf. Doe 1 v. Perkiomen Valley Sch. Dist., 585 F. Supp. 3d 668, 705 (E.D. Pa. 2022) (noting that “a district court may require a nominal bond even absent an exception to Rule 65”) (citing Temple, 941 F.2d at 220 n.28). The Court will provide Plaintiffs up to an including August 4, 2023, within which to post an initial bond of $50,000 with the Clerk of Court, or, alternatively, to file a notice informing the Court that Plaintiffs withdraw their request for preliminary injunctive relief. III. CONCLUSION Injunction” insofar as Plaintiffs have established their entitlement, during the pendency of this litigation, to a water distribution program by Terminals for Plaintiffs and putative class members who cannot afford to purchase water without trading off other basic necessities. For the reasons discussed herein, the Court will order that the water distribution program be

established and operated as set forth in the “Order Implementing Water Distribution Program,” filed contemporaneously herewith. Date: July 20, 2023 _______/s/_______ WILMA A. LEWIS District Judge