VINCE CHHABRIA, United States District Judge

Martin S. Schenker, Cooley LLP, Bree Ann Bernwanger, Pro Hac Vice, Emilou MacLean, Hayden Miles Rodarte, Lawyers’ Committee for Civil Rights, Francisco Ugarte, Genna Ellis Beier, Jennifer Taylor Friedman, Kelly Engel Wells, Pro Hac Vice, San Francisco Office of the Public Defender, Sean Connor Riordan, William S. Freeman, Angelica Salceda, ACLU Foundation of Northern California, San Francisco, CA, Amalia Margarete Wille, Judah Ben Lakin, Lakin & Wille LLP, Oakland, CA, Francisco Unger, Pro Hac Vice, Timothy W. Cook, Pro Hac Vice, Cooley LLP, Boston, MA, Jordan Samuel Wells, ACLU of Southern California, Los Angeles, CA, Stephanie Padilla, ACLU of Southern California, Bakersfield, CA, for Plaintiffs Angel de Jesus Zepeda Rivas, Brenda Rubi Ruiz Tovar, Lawrence Kuria Mwaura, Luciano Gonzalo Mendoza Jeronimo, Coraima Yaritza Sanchez Nunez, Javier Alfaro, Dung Tuan Dang.

Emilou MacLean, Jennifer Taylor Friedman, Kelly Engel Wells, San Francisco Public Defenders Office, Martin S. Schenker, Cooley LLP, William S. Freeman, Sean Connor Riordan, ACLU Foundation of Northern California, San Francisco, CA, Timothy W. Cook, Cooley LLP, Boston, MA, for Plaintiffs Rajnish Rajnish, Willian Matias Rauda, Juan Jose Erazo Herrera.

Adrienne Zack, Neal Cho Hong, United States Attorney's Office, San Francsico, CA, Shining J. Hsu, Shiwon Choe, Wendy M. Garbers, United States Attorney's Office Northern District of California, San Francisco, CA, for Defendants David Jennings, US Immigration and Customs Enforcement.

Royal Forest Oakes, Michael A.S. Newman, Hinshaw & Culbertson LLP, Susan Eileen Coleman, Burke, Williams & Sorensen LLP, Los Angeles, CA, David S. Weinstein, Pro Hac Vice, Hinshaw and Culbertson LLP, Coral Gables, FL, for Defendants GEO Group, Inc., Nathan Allen.

ORDER GRANTING MOTION FOR SECOND PRELIMINARY INJUNCTION

Re: Dkt. No. 500

VINCE CHHABRIA, United States District Judge

This lawsuit was filed by immigration detainees at two detention facilities in California. Although the detainees are in ICE custody, ICE does not directly operate either facility. One facility—Mesa Verde Detention Center—is run by a for-profit corporation called GEO Group, pursuant to a contract with ICE. The other facility is a jail in Yuba County run by the local sheriff. The jail holds criminal detainees charged with crimes under state law, but it reserves space for immigration detainees and holds them pursuant to a contract with ICE. The plaintiffs allege that ICE and its contractors have been deliberately indifferent to the health risks posed by the coronavirus pandemic.

In April 2020—shortly after the lawsuit was filed—the Court entered a temporary restraining order which primarily required ICE and its contractors to provide information about the detainees to assist the Court in determining whether some of them could be released on bail while the lawsuit was pending, as a way of alleviating health risks for the detainee population as a whole. Later, the Court entered a preliminary injunction which required the bail process to continue but rejected some of the more intrusive forms of relief sought by the plaintiffs.

As applied to Mesa Verde—the facility operated by GEO Group—the decision to avoid being intrusive turned out to be a mistake. From the start of the public health crisis until now, the conduct of the key ICE and GEO officials in charge of operations at Mesa Verde has been appalling. These officials knew that they needed a clear and detailed plan to minimize the risk of an outbreak (and to contain an outbreak if one occurred), but nine months later they still have not created one. They deliberately avoided testing detainees and staff for fear that the results would require them to take expensive and logistically challenging safety measures. They failed to address the safety concerns created by Mesa Verde's unique layout, which makes it far more dangerous from a contagion standpoint than the typical jail or prison. They opposed bail for detainees on a blanket basis—even for those who clearly posed no danger to the community and were obviously not a flight risk. They gave false testimony several times in these court proceedings, on matters of importance. And at least one ICE official with significant decision-making authority over Mesa Verde obstructed the proceedings by effectively refusing to answer, during his deposition, even the most basic questions about ICE's response to the pandemic.

It should thus come as no surprise that the Mesa Verde facility experienced a severe and prolonged outbreak during the summer—an outbreak that ICE and GEO made no meaningful effort to prevent and were totally unprepared to respond to. The result was an additional, more intrusive temporary restraining order directed solely at the Mesa Verde facility requiring regular detainee and staff testing, preventing new detainees from entering the facility, and imposing other safety measures. Now that the outbreak has subsided, the defendants argue that a preliminary injunction with similarly intrusive requirements should not issue, and that the temporary restraining order directed at Mesa Verde should simply be lifted. This argument is impossible to accept, particularly given the unique layout of the facility and the abominable performance of those who run it. Not to mention the recent surge in Covid-19 cases among facility staff—cases we might still be unaware of if not for the temporary restraining order.

The plaintiffs’ motion for a preliminary injunction is, in large part, granted. This ruling assumes familiarity with the Court's prior rulings. See Zepeda Rivas v. Jennings , 445 F. Supp. 3d 36 (N.D. Cal. 2020) ; Zepeda Rivas v. Jennings , 465 F. Supp. 3d 1028 (N.D. Cal. 2020) ; Zepeda Rivas v. Jennings , 2020 WL 4554646 (N.D. Cal. Aug. 6, 2020). The factual and legal analysis from those rulings are incorporated here, and the evidence submitted in connection with them remains part of the record.

I

Plaintiffs seeking a preliminary injunction must show that they are "likely to succeed on the merits, that [they are] likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in [their] favor, and that an injunction is in the public interest." Winter v. Natural Resources Defense Council, Inc. , 555 U.S. 7, 20, 129 S.Ct. 365, 172 L.Ed.2d 249 (2008). Where, as here, the government is a party to the case in which a preliminary injunction is sought, the last two factors merge. Drakes Bay Oyster Co. v. Jewell , 747 F.3d 1073, 1092 (9th Cir. 2014). In addition, when the balance of hardships tips decidedly in the plaintiffs’ favor, they need only show "serious questions going to the merits" of their claims instead of a likelihood of success. Alliance for the Wild Rockies v. Cottrell , 632 F.3d 1127, 1135 (9th Cir. 2011).

In the context of determining whether a preliminary injunction should issue to ensure that conditions at an ICE detention facility do not create an unconstitutional risk of detainees contracting Covid-19, the Ninth Circuit's recent decision in Roman v. Wolf , 977 F.3d 935 (9th Cir. 2020) (per curium), is directly on point. In Roman , the court held that the crowded and unsanitary conditions at the Adelanto detention facility and the government's failure to address these conditions in the face of the pandemic likely violated the detainees’ constitutional rights. Id. at 943-44. The court also held that the detainees were likely to suffer irreparable harm without preliminary relief given Covid-19's high mortality rate, and that the equities and public interest weighed in plaintiffs’ favor. Id. at 944. The Ninth Circuit emphasized that the district court "possesses broad equitable authority" to order any measures necessary to remedy the likely constitutional violations at the facility and to prevent the spread of an outbreak. Id. at 945. And the court instructed that injunctive relief in this context should: (a) not be based solely on the vague and unspecific provisions in the CDC guidelines for correctional and detention facilities; (b) avoid micromanaging the administration of the facility; (c) be based on medical evidence before the court; and (d) reflect the scientific evidence in the record. Id. at 946.

II

The government's main argument against continued interim relief is that the plaintiffs have not shown that the defendants have acted and are continuing to act with deliberate indifference toward their safety. Prior rulings have already identified many reasons why the defendants’ conduct amounted to deliberate indifference in violation of the plaintiffs’ constitutional rights. To briefly recap, they include the defendants’ conscious avoidance of widespread testing for fear that positive tests would require them to take measures to protect the safety of detainees that they preferred not to take; the defendants’ submission of a declaration from Assistant Field Office Director Alexander Pham falsely stating that Mesa Verde was imposing a 14-day quarantine on all new arrivals from facilities with reported cases of Covid-19 when no such quarantine existed; the defendants’ blanket opposition to all bail applications; and the defendants’ failure to take the initiative to release the few women held at Mesa Verde until after the Court asked them why so few female detainees (with such minimal criminal histories) were tying up an entire 100-person dorm that could have been used to allow male detainees to socially distance. See Dkt. Nos. 357, 500.

On the issue of clearing the women's dorm, the defendants have now made additional false statements to the Court. In a declaration dated November 2, 2020, Deputy Field Office Director Erik Bonnar stated that he made the decision to try and empty the women's dorm at Mesa Verde "on or about March 18, 2020." Dkt. No. 786-8 ¶ 8. This assertion was part of an effort to combat the impression that ICE officials were deliberately indifferent to the health risks at Mesa Verde, and to combat the impression that they only decided to clear the women's dorm once the Court raised questions about it. But in a prior declaration submitted at the end of April, Bonnar did not mention emptying the women's dorm as one of the steps ICE was taking to reduce Mesa Verde's population. See Dkt. No. 37-1. And a different ICE official submitted a declaration in mid-May asserting that ICE made the decision to release all female detainees on May 7, 2020, after the Court had raised the issue. See Dkt. No. 166-1 ¶ 9. At the evidentiary hearing for this preliminary injunction motion, the Court—in an effort to discern whether any documentary evidence backed up Bonnar's new assertion that he had decided to clear the women's dorm in March—asked Bonnar whether any emails were exchanged about this alleged decision, and Bonnar said he couldn't find any. Bonnar 11/18/20 Tr. 364-65. When the Court asked whether that was odd, Bonnar said no because the discussions were being held at the time they were all in the office, meaning that discussions were not occurring via email. But Bonnar then stated that he began working from home at the end of March, and confirmed that there were no emails on the topic in April or May. Considering these facts and the entire record in this case, along with Bonnar's demeanor during the hearing, the Court has no choice but to find that Bonnar's testimony is false. These false statements made under oath regarding the women's dorm, along with the false declaration from Pham about how Mesa Verde was handling incoming detainees, further underscore that the defendants cannot be trusted to conduct themselves responsibly as it relates to the safety of the detainees.

There is still more evidence in the record—evidence that was not available at the time of the prior rulings—that confirms the defendants’ deliberate indifference to the safety of the detainees. For example, the defendants were aware from early on that they needed to create a plan to deal with a possible Covid-19 outbreak at Mesa Verde. They knew—and were repeatedly warned—about the issues that would arise if they failed to do so. In March, Dr. Richard Medrano, the regional medical director who oversees healthcare at Mesa Verde, informed both GEO and ICE officials that he "anticipate[s] that as intake continues we will encounter more of these cases [of needing to isolate detainees] so we need to figure out a practical solution to ... isolation and monitoring given the limited space at Mesa Verde." Exh. 78. Dr. Medrano reiterated these concerns in June, recommending that ICE and GEO create a contingency plan specifying where detainees would be housed if various numbers of detainees tested positive. Medrano 11/20/20 Tr. 666.

Dr. Medrano was not the only one identifying this need for a contingency plan. On June 4, Pham wrote to his superior that "if/when" the first Mesa Verde detainee tested positive for Covid-19, there would be potential "immediate logistical issues," and that "if/when" there was a second positive case, "there [would] likely be no logistical way to cohort/isolate" that person. Exh. 13. On June 25, a GEO official chastised Nathan Allen, the GEO employee who serves as Mesa Verde's facility administrator, for the inadequacies of the plan he was proposing at that point, stating that it "falls short of identifying mitigating procedures [he] could put in place for detainees who do not consent to testing" and stating that it "should at a minimum identify specific areas within the unit for new arrivals." Exh. 232. The GEO official further warned that "[w]e cannot just throw up our hands and say there isn't anything we can do," and lamented that even though they had "discussed this multiple times before, [ ] it appears there has been no creative effort to come up with some mitigating strategies." Exh. 232.

The general Covid-19 guidelines for detention centers published by the CDC, ICE, and GEO did not in any way obviate the need to create a Mesa Verde-specific contingency plan. See Roman , 977 F.3d at 946. In fact, the CDC guidance in effect in March instructed detention centers to develop a plan that "include[d] contingencies for multiple locations if numerous cases and/or contacts are identified and require medical isolation or quarantine simultaneously." Exh. 379, pg. 10. And experts for both sides agreed that facilities needed to adapt these general guidelines to create plans geared toward particular facility configurations. See Greifinger 11/17/20 Tr. 293; Medrano 11/20/20 Tr. 634. This need was especially critical for a facility like Mesa Verde, which has an unusual housing layout consisting of four 100-person dormitories with bunk beds, and only five single-person cells (three designated for disciplinary use and two for medical isolation). As experts for the plaintiffs and defendants also agreed in their testimony, Mesa Verde faced especially daunting challenges in dealing with Covid-19 because of this layout: dormitories have a high risk of transmission because of a lack of barriers, and the facility had limited ability to separate detainees if an outbreak occurred. See Greifinger 11/17/20 Tr. 168, 182; Medrano 11/20/20 Tr. 665.

Despite these warnings from as early as March that Mesa Verde needed a facility-specific plan, by the beginning of July no plan was in place. So when a group of new detainees arrived together at Mesa Verde on July 1 and one tested positive for Covid-19, the facility had neither a plan nor a space to cohort and house the detainees who had been exposed to Covid-19. Luckily, another ICE facility agreed to accept and house the exposed detainees. But the crisis that would have ensued if the detainees had needed to stay at Mesa Verde was obvious to those managing the facility, including Mesa Verde's facility administrator, who commented to his superiors that they were "dodging a bullet." Exh. 227. Yet even after this recognition, the defendants remained firmly planted in the line of fire and continued to receive new intakes during the latter half of July. And although Pham submitted a declaration stating that new intakes arriving from facilities with confirmed cases of Covid-19 were quarantining for 14 days before being housed with the general population, no such quarantining was occurring. In fact, the defendants were placing all new intakes into the general population dorms after nothing more than a general symptoms screening, without isolating or quarantining them for any amount of time. See Bonnar 11/18/20 Tr. 371.

A similar story plays out in the context of staff testing. Experts from both sides agree that staff are one of the main ways that Covid-19 can be introduced into a closed setting like Mesa Verde. See Greifinger 11/17/20 Tr. 209; Henderson 11/19/20 Tr. 524. And by the end of June, multiple staff members working at Mesa Verde had tested positive for Covid-19. See Dkt. Nos. 402, 407. But nearly two months after the first staff member reported being Covid-positive, Mesa Verde still had not implemented any form of staff testing. See Exh. 335. At the beginning of August, an ICE compliance officer raised this concern, noting that Mesa Verde had not "instituted mandatory testing for its staff as have other ICE contract facilities." Exh. 335. The defendants have provided no satisfactory explanation for why they never required their staff to be tested for Covid-19, especially when other ICE facilities adopted this seemingly obvious measure. In fact, the evidence in the record suggests that the reason the defendants avoided implementing universal staff testing was the same reason they avoided implementing universal detainee testing—they feared that positive tests would require them to adopt safety measures that (at least from the standpoint of a for-profit company running a detention facility) were undesirable. See Exh. 228 (email from Mesa Verde facility administrator stating that an ICE official overseeing the facility "mentioned [ ] he would rather not have staff testing as that may also impact [ICE] functions, i.e., an asymptomatic person testing positive would require possible dorm cohorts and detainee testing protocols").

The defendants still had no Mesa Verde-specific contingency plan by the end of July when the Covid-19 outbreak at the facility began, which, predictably, exacerbated the spread. See Bonnar 11/18/20 Tr. 370-371, 377. On July 29, a detainee in Dorm B tested positive. The next day, Mesa Verde tested all detainees in Dorm B. But without any plan in place for what to do if multiple detainees tested positive, and despite knowing for months that they needed a plan, officials were, quite literally, relying on hopes and prayers. The acting Health Services Administrator at Mesa Verde (and one of the employees primarily responsible for overseeing the day-to-day medical care at the facility) noted that she "hope[d] [they] all have a plan for the results." Exh. 323. Mesa Verde's facility administrator similarly told his GEO superiors that he was "[p]raying for negatives" and had just reached out to ICE officials "pertaining to what happens if we get multiple positives and no way to isolate or cohort." Exh. 312. One of the GEO superiors responded that they "definitely need a contingency plan in the event we have an unexpected high number of positive cases." Exh. 312. And when Mesa Verde's facility administrator told ICE that all detainees in Dorm B were getting tested, he stated that positive results would create problems because the facility had no room to house them but that "[o]f course, these are the issues we knew would happen if a positive population case." Dkt. No. 845.

The story of this failure to create a contingency plan despite knowing of the need to do so is clear evidence of deliberate indifference. See Roman , 977 F.3d at 943. And the problems stemming from not having a plan compounded as detainees in additional dorms began to test positive. On July 31, a detainee in Dorm C who was experiencing symptoms consistent with Covid-19 was sent to the hospital, where he tested positive. Dkt. No. 786-4 ¶ 7. The next day, ICE instructed Mesa Verde to test all detainees in Dorm C because of the possible exposure. See Exh. 529. But the lack of any testing or contingency plan led to a delay in testing the exposed detainees—Mesa Verde's facility administrator responded to ICE's testing directive by saying that the testing of Dorm C would need to be "postponed until a plan can be approved by ICE and subsequently approved by GEO leadership." Exh. 529. The facility administrator sent this draft plan to ICE for approval on August 3—three days after learning of Dorm C's exposure—and with the caveat that the plan still needed GEO approval before it could be implemented. Exh 530. Testing of detainees in Dorm C thus did not finish until August 4, a full four days after the defendants learned about the possible exposure. See Dkt. No. 799-51. This kind of delay in testing is quite dangerous in congregate settings like detention centers where the virus can spread rapidly, and particularly so in a detention center like Mesa Verde where detainees live in dormitory-style housing. See Greifinger 11/17/20 Tr. 219-220; Dkt. No. 799-58 ¶¶ 11, 21-22. The initial delay in administering the tests was then exacerbated by a nearly two-week delay in receiving results—a delay that the defendants were aware might occur because of testing backlogs. See Exh. 341; Exh. 8 ¶ 10 n.3.

In addition to the undue delays in testing, the lack of a contingency plan resulted in a failure to properly handle symptomatic detainees at the onset of the outbreak. Specifically, the defendants left multiple symptomatic detainees in Dorm C pending their test results, even knowing that those detainees had a good chance of having Covid-19 because of their exposure to the hospitalized patient. See Exh. 322. Similarly, as the defendants struggled to contain the outbreak and find space to house the detainees who tested positive, they initially contemplated mixing symptomatic detainees from Dorm C with asymptomatic detainees from Dorm B. See Dkt. No. 799-50. The defendants knew that this proposal was contrary to the recommendation of their medical staff, but found that "space issues" made it "understandable" that it was not possible to separate symptomatic and asymptomatic detainees at that time. Dkt. No. 799-50. Ultimately, a few detainees from Dorm B also began exhibiting symptoms before the two dorms were combined. See Dkt. No. 799-50. The result was that the defendants combined symptomatic and asymptomatic detainees from Dorm B along with symptomatic and asymptomatic detainees from Dorm C. See Dkt. No. 799-50. All of these actions flew in the face not only of the advice of the facility's medical staff but also of the CDC guidelines for detention facilities in effect at the time, which mandated that detainees with Covid-19 symptoms "should be placed under medical isolation immediately." Exh. 126, pg. 17. The defendants’ own medical expert testified that mixing symptomatic and asymptomatic detainees together in this fashion was likely a "mistake." Henderson 11/19/20 Tr. 476.

Considering all of this conduct together, along with the examples identified in prior rulings and the other evidence in the record, it is obvious that the defendants have repeatedly acted with deliberate indifference to the safety of detainees at Mesa Verde. And, as explained in Roman and this Court's prior orders, the danger of irreparable harm is high and the public interest weighs in favor of an injunction. See Roman , 977 F.3d at 944. A preliminary injunction ensuring that the defendants provide constitutionally adequate protection to detainees at Mesa Verde—not to mention to help protect the community of which Mesa Verde is a part—is thus warranted.

III

The defendants argue that circumstances have changed since the first preliminary injunction and the second TRO, with respect to both the conditions at Mesa Verde and the measures the defendants have taken to protect the detainees housed inside. Thus, their argument goes, even if the plaintiffs have proved that the defendants acted with deliberate indifference in the past, future injunctive relief is unwarranted because there is nothing to suggest they continue to do so or will do so in the future.

The defendants are correct that any injunctive relief must be tailored to present circumstances. See Roman , 977 F.3d at 945. But, as discussed in more detail below, there is no reason to believe that Covid-19 has ceased to present a risk to the health and safety of the detainees; to the contrary, 15 staff members who work at Mesa Verde have reported positive Covid-19 tests in just the last few weeks, after months without a single new positive case. Dkt. No. 864. Relatedly, there is a similar surge in Covid-19 cases reported in Kern County where Mesa Verde sits, as well as in counties all around the state and country, from which new intakes to Mesa Verde could come. See Exh. 459. The current state of the pandemic, both within and without Mesa Verde, thus cuts in favor of further injunctive relief.

Moreover, the primary measure the defendants identify as evidence that they will now provide a constitutionally adequate level of care to the detainees is their so-called Detainee Intake Procedures plan, a two-page document submitted at the eleventh hour in conjunction with their response to the order to show cause why a preliminary injunction should not issue. See Dkt. No. 786-10. There are many reasons why this plan does not preclude the need for additional injunctive relief.

First, it was adopted months after the Court imposed the first preliminary injunction back in June, and nearly on the eve of the hearing for the second preliminary injunction. Cf. Wilson v. Williams , 961 F.3d 829, 841 (6th Cir. 2020) (vacating preliminary injunction in part because facility took several key preventative measures before district court issued preliminary injunction). Second, the plan deals only with the treatment of intakes and fails to provide any guidance on what the facility will do if another outbreak occurs. It thus does not address any of the concerns raised above about how the lack of a response plan contributed to and exacerbated the first outbreak. Third, there is a difference between having a plan and having the willingness and/or ability to execute that plan. The defendants’ conduct over the course of this lawsuit sheds considerable doubt on this latter (and equally vital) consideration. Indeed, acting Deputy Field Office Director Moises Becerra testified that one of the ICE officials primarily responsible for implementing the plan on the federal side would be Alexander Pham. Becerra 11/16/20 Tr. 85. Given Pham's misrepresentations to the Court in sworn declarations about topics as important as whether incoming detainees were being quarantined and the distance between bunk beds in the dorms, not to mention his evasive responses during his videotaped deposition testimony that appeared designed to prevent the plaintiffs and the Court from obtaining even the most basic information, it is difficult to put faith in a plan that depends on Pham acting responsibly. Finally, the plan itself is extremely vague. It leaves key decisions—such as where to quarantine incoming detainees who refuse testing or test positive—to the discretion of Nathan Allen, the facility administrator. As already discussed, he is another person who has proven to exercise his discretion in a reckless and indifferent fashion.

During the evidentiary hearing, witnesses for the defendants identified several measures they would take in conjunction with the written plan to address the risk of Covid-19 in the facility. These included things like restarting the quarantine clock for intakes who are cohorted together every time a new intake arrives, and instituting an ongoing sample-size saturation testing plan for both staff and detainees. The adoption of many, if not all, of these measures seems wise, but the plan itself does not include or require them. The plan thus does not undercut the evidence produced by the plaintiffs—or the prior findings of this Court—that the defendants cannot be trusted on their own to provide reasonably safe conditions to detainees at Mesa Verde. And at a more fundamental level, the defendants’ argument that the deliberate indifference they demonstrated in the past has no bearing on whether a preliminary injunction is warranted now is wrong. "[P]ast illegal conduct gives rise to an inference that future violations may occur." United States v. Laerdal Manufacturing Corp. , 73 F.3d 852, 857 (9th Cir. 1995) ; see also Orantes-Hernandez v. Thornburgh , 919 F.2d 549, 564 (9th Cir. 1990) ("[A] district court has ‘broad power to restrain acts which are of the same type or class as unlawful acts which the court has found to have been committed or whose commission in the future, unless enjoined, may be fairly anticipated from the defendant's conduct in the past.’ " (quoting NLRB v. Express Publishing Co. , 312 U.S. 426, 435, 61 S.Ct. 693, 85 L.Ed. 930 (1941) )).

Moreover, the defendants’ deliberate indifference was not isolated; it has been ongoing. Most notable is their handling of the outbreak at the end of July, which was well after the initial TRO and preliminary injunction were entered. It is true that the defendants have taken some affirmative measures to better prevent the spread of Covid-19 at Mesa Verde since the filing of the lawsuit and the entry of the first preliminary injunction. However, "reforms introduced under protest or after the violation was discovered" do not preclude the need for injunctive relief. Laerdal Manufacturing Corp. , 73 F.3d at 857. This is especially true where, as here, the vast majority of reforms were adopted in response to the court orders entered in this case (or the threat of additional court orders), and the violations have been ongoing and serious. See id. at 854-56 ; see also Federal Election Commission v. Furgatch , 869 F.2d 1256, 1262 (9th Cir. 1989). Considering the whole record in this case, there exists, to put it mildly, "some cognizable danger of recurrent violation." Laerdal Manufacturing Corp. , 73 F.3d at 854 (quoting United States v. W.T. Grant Co. , 345 U.S. 629, 633, 73 S.Ct. 894, 97 L.Ed. 1303 (1953) ).

IV

Having determined that a second preliminary injunction is warranted, the question then becomes what the scope of that relief should be. That question, in turn, depends on current conditions, both within Mesa Verde and in the broader community, as well as on the current medical understanding of how to handle the virus. See Roman , 977 F.3d at 945-46.

First, the conditions at Mesa Verde. As already discussed, the unique dormitory style layout at Mesa Verde demands precautions that would perhaps be unnecessary if the facility had a different configuration. But because Mesa Verde has only five individual cells, the facility is virtually hamstrung in its ability to cohort and isolate detainees. Thankfully, the population at Mesa Verde has been reduced substantially since the lawsuit was filed. At the time of this ruling, Mesa Verde is housing 45 detainees, which is about 11% of its capacity, in three of its four dorms. Dkt. No. 863. Twenty-four of the current detainees have tested positive for Covid-19 at some point, but all are now deemed recovered and there are currently no active cases in the facility. Dkt. No. 863. At the height of the outbreak, 57 of the 103 detainees at Mesa Verde tested positive; three were sent to the hospital. See Dkt. No. 580. The last new detainee to test positive did so on August 18. See Dkt. No. 786 pg. 6.

But the rate of Covid-19 among the staff working at Mesa Verde is another important part of the equation. As of November 16, when the evidentiary hearing on the order to show cause began, no staff member had tested positive since August 26. But since November 17, there has been a rapid surge in new positive cases, with 15 staff members reporting new positive tests in just this short time. Dkt. No. 864. This resurgence of new Covid-19 cases reflects a larger trend across Kern County, the state of California, and the country at large. In Kern County, for example, there has been a wave of new cases from the beginning of November, with 350 cases described as "new" as of November 20. Exh. 459. For context, during the first outbreak at Mesa Verde, 29 of 126 staff members working at Mesa Verde tested positive for Covid-19. Dkt. No. 620.

Finally, the medical experts who testified during the evidentiary hearing seemed to share the baseline understanding that any plan for how to handle the threat of Covid-19 at Mesa Verde would have to account for the separation of five distinct groups. These five groups include: (1) the general detainee population; (2) detainees who test positive for Covid-19; (3) intakes who need to cohort after arrival; (4) intakes who refuse testing; and (5) symptomatic detainees. See, e.g. , Greifinger 11/17/21 Tr. 191, 236-37; Henderson 11/19/20 Tr. 458, 465-66, 492-93. The injunctive relief in this case must accordingly be tailored to allow for the separation of these five groups, within the confines of Mesa Verde's unique structural layout.

The defendants continue to object to a preliminary injunction in any form. However, the measures that they agreed should be adopted if an injunction did issue overlapped in many ways with the measures requested by the plaintiffs and with the measures that the defendants’ own experts stated were necessary. These include key provisions such as isolating detainees who are symptomatic. But there are a few major aspects of the plaintiffs’ proposed injunction that the defendants argue are unnecessary. The Court addresses both these key areas of agreement and disagreement before outlining the full scope of injunctive relief.

Dorm for detainees who test positive for Covid-19 : The experts who testified at the hearing, along with many of the ICE officials responsible for managing Mesa Verde, seemed to agree that reserving a dorm for Covid-positive detainees is a necessary way to safely manage the facility. See Becerra 11/16/20 Tr. 96-99; Henderson 11/19/20 Tr. 475-76, 494; Medrano 11/20/20 Tr. 667. The defendants also stated that if a preliminary injunction should issue, this would be an acceptable provision. 11/20/20 Tr. 691. Reserving a dorm for Covid-positive detainees thus seems to be a bare minimum for any effective injunction.

Intake procedures: Most of the intake procedures ordered in this injunction are contemplated by the defendants themselves in their proposed Detainee Intake Procedures plan, and everyone appears to agree that those procedures are necessary and/or mandated by the CDC. These include things like screening and providing point-of-care tests to all intakes, quarantining intakes who test positive, and cohorting intakes for a safe period of time after their arrival. Experts on both sides also seem to agree that intakes who refuse testing must be treated similarly to symptomatic detainees and isolated individually. See Greifinger 11/17/20 Tr. 236-39; Henderson 11/19/20 Tr. 458. However, a few key questions on how to handle intakes remain disputed.

First, there is a question about where to cohort intakes who receive a negative point-of-care test upon arrival. All parties appear to agree that there must be some designated space to cohort these people. The plaintiffs have proposed two options: (1) a reserved dormitory, or (2) the five individual cells. At the hearing, the defendants expressed a preference for the dormitory option because it gives the facility more flexibility to use the individual cells to help manage safety issues and non-Covid-19 related health issues that arise in the facility. 11/20/20 Tr. 691-92. Reserving a dormitory for cohorting intakes also provides greater flexibility should another outbreak occur at the facility and the five individual isolation rooms are all occupied, as occurred in July. Cognizant of the need to avoid micromanagement at Mesa Verde, this seems like a preferable and less intrusive option than requiring the defendants to cohort intakes in individual cells. See Roman , 977 F.3d at 946.

In addition, the testimony given by the defendants’ witnesses during the evidentiary hearing, especially in response to the Court's questions about how Mesa Verde could safely separate the five groups in the event of another outbreak, suggests that new intakes must cohort in a dorm. For example, Becerra stated that the facility was "working towards" having two non-general population dorms in order to handle another outbreak. Becerra 11/16/20 Tr. 97. Allen, the facility administrator, similarly suggested that separating those five groups would require using one dorm to cohort intakes. Allen 11/19/20 Tr. 556-57; Allen 11/20/20 Tr. 622-23. And both of the defendants’ medical experts seemed to contemplate a facility plan involving a cohort dorm, especially if the plan needed to account for another outbreak. Henderson 11/19/20 Tr. 492-93; Medrano 11/20/20 Tr. 643. Dr. Henderson did testify that using the individual cells to cohort intakes would be another viable option. But he stated that if there were a sudden increase in symptomatic detainees—as occurred in July and August—these symptomatic detainees would have to be transferred out of Mesa Verde in order to safety isolate them. Henderson 11/19/20 Tr. 466. Given the inherently risky nature of transfers, along with the record evidence showing that other detention facilities have refused to accept detainees from Mesa Verde because of its confirmed cases of Covid-19, a plan dependent on transferring detainees seems dangerous to both the detainees themselves and the staff members responsible for the transfers. See Exh. 256.

Second, there is a question about how many days worth of intakes can be cohorted together. The defendants propose a plan where intakes arriving from Monday through Friday can all become part of one cohort. However, the August GEO Interim Reference Sheet on Covid-19 restricts the cohorting of intakes who arrive separately to "patients arriving within 48 hours of each other." Exh. 623, pg. 4 (emphasis in original). The defendants have offered no compelling reason for why the conditions at Mesa Verde should allow for an upward deviation from this general rule. If anything, the dormitory style layout at the facility would seem to demand a stricter timeline. Detainees can thus be cohorted together if they arrive within 48 hours of each other. And the quarantine clock shall start over whenever a new arrival is added to the dorm, as ICE officials agreed that it must. See, e.g. , Becerra 11/16/20 Tr. 82, 146.

Finally, the plaintiffs ask that intakes be tested at the end of their quarantine before being moved into a general population dorm. At least on this record, that does not seem necessary. Dr. Henderson testified that the CDC does not recommend exit testing after intake cohorting, and that he has not seen this strategy implemented at other facilities. Henderson 11/19/20 Tr. 490. The intakes in quarantine will also be tested every week along with the general detainee population, as discussed below, such that they will be tested before being moved into the general population dorms.

Dorm population limit: The parties disagree on the number of detainees that can be safely housed in each general population dorm. As a preliminary note, all agree that social distancing (and the ability to remain six feet apart) remains critical. At Mesa Verde, each bunk bed is approximately 38 inches, or just over 3 feet, apart. Exh. 633, pgs. 31-32. The defendants assert that a 50% cap on dorms provides for adequate social distancing because if the detainees alternate between top and bottom bunks and alternate the direction in which they sleep, the distance between their heads is over six feet. But the detainees still must get in to and out of their bunks, which will inevitably require them to be sharing the 38-inch space between the beds. Exh. 121 ¶ 35; Greifinger 11/17/20 Tr. 173-75. In addition, as the plaintiffs’ expert testified, the droplets carrying the virus can travel up to six feet in any direction, meaning that the diagonal difference between the heads of the sleeping detainees in adjacent bunks is not the relevant measurement—the lateral distance between the two bunks is. Greifinger 11/17/20 Tr. 175-76. The defendants’ expert stated that he didn't disagree with this assessment. See Henderson 11/19/20 Tr. 515-16. And it's worth noting that the earlier outbreak occurred while the facility was at about a third of its capacity, and, as recounted above, the outbreak was still severe. A population cap for each dorm that allows the detainees to sleep in beds at least six feet apart is thus needed to allow for adequate social distancing. At present, this means each dorm is limited to 26 detainees. The population cap does not apply to the covid-positive dorm, which may house more detainees if the number of positive cases exceeds the cap.

Saturation testing for detainees and staff: Another key area of disagreement is whether periodic saturation testing of detainees and staff should occur. With respect to the need for mass testing generally, experts on both sides agreed that some level of saturation testing is appropriate to protect against the spread of the virus. Dr. Greifinger, the plaintiffs’ expert, strongly recommended saturation testing for institutions such as Mesa Verde, where the possibility of rapid contagion is especially high. Greifinger 11/17/20 Tr. 207-08. Dr. Medrano, a percipient witness who also served as an expert for the defendants, testified that he and GEO's Chief Medical Officer agreed that weekly testing of both detainees and staff was recommended, but in a sample-size, scaled down way. Medrano 11/20/20 Tr. 642-44. It is true that Dr. Henderson, another defense expert, recommended against saturation testing of detainees in detention settings, but the apparent reason for this recommendation was to avoid dealing with the logistical difficulties that would arise from a large number of positive cases. Henderson 11/19/20 Tr. 523-24. This is not an appropriate reason to avoid testing.

With respect to frequency, weekly testing is necessary to ensure the continued safety of the detainees. This level of testing is in accordance with the recommendation from one of the plaintiffs’ experts. See Dkt. No. 545-1 ¶ 9. It is true that Dr. Greifinger, one of the plaintiffs’ other experts, testified that two of the jails he is monitoring conduct mass testing only every three weeks. See Greifinger 11/17/20 Tr. 208. But these jails do not have the dormitory-style housing that Mesa Verde has. Moreover, the rate at which new staff members at Mesa Verde have begun testing positive for Covid-19 in recent days suggests that testing more frequently than every three weeks is necessary to stay ahead of any potential outbreak.

In addition, staff testing remains just as important as detainee testing. Experts on both sides agreed that staff is one of the main ways that the virus enters closed environments like detention centers. Greifinger 11/17/20 Tr. 209; Henderson 11/19/20 Tr. 524. In fact, Dr. Greifinger asserted that every detention center outbreak he has observed began with staff. Greifinger 11/17/20 Tr. 209. Moreover, the recent surge in positive cases among Mesa Verde staff—which was precisely what occurred in July before the outbreak among Mesa Verde detainees—highlights the need for continued staff testing.

The defendants counter that CDC guidance doesn't call for strict saturation testing. But it bears repeating that these are general guidelines that need to be tailored to Mesa Verde specifically. Given the defendants’ failure to implement testing in the past, the rising rates of Covid-19 among staff members and in the surrounding county, and the spatial particularities at Mesa Verde that make the early detection of an outbreak especially important, mass testing is needed to ensure the detainees’ safety. Moreover, the only way foregoing saturation testing at a facility like Mesa Verde could conceivably work is if the officials on the ground were competently and responsibly watching for warning signs of an outbreak. As should be clear, there's no reason to believe that would occur here.

Symptomatic detainees: CDC guidelines and medical experts on both sides agree that isolating symptomatic detainees is crucial to preventing the spread of Covid-19. See Exh. 381, pg. 19; Greifinger 11/17/20 Tr. 185-86; Henderson 11/19/20 Tr. 462. The defendants’ experts also agreed that a rapid point-of-care test should be offered to any detainee experiencing symptoms consistent with Covid-19 because these are precisely the circumstances for which the rapid tests are designed. Henderson 11/19/20 Tr. 528-29; Medrano 11/20/20 Tr. 647. Using the rapid test has the added benefit of potentially freeing up one of the few individual cells more quickly because if the test comes back positive, the detainee can be moved out of isolation and into the positive dorm.

Transfers: The plaintiffs request that the Court continue to prohibit the defendants from transferring detainees out of Mesa Verde without the consent of class counsel or the Court, other than for purposes of removal or release. However, Dr. Henderson testified repeatedly that the ability to transfer detainees among facilities—taking, of course, the requisite safety precautions—is crucial to a facility's ability to effectively manage the threat of an outbreak. See, e.g. , Henderson 11/19/20 Tr. 466, 492. Given this testimony, a restriction of the type requested by the plaintiffs would wade too far into micromanagement of ICE's system for detaining people pending removal. In addition, Mesa Verde will be limited by ICE's Pandemic Response Requirements plan, which has discontinued transfers "unless necessary for medical evaluation, medical isolation/quarantine, clinical care, extenuating security concerns, release or removal, or to prevent overcrowding." Exh. 119, pg. 4.

For similar reasons, the plaintiffs’ proposal to limit all transfers into Mesa Verde, even from non-ICE jurisdictions, unless necessary for medical, security, removal or overcrowding reasons is rejected. This restriction would also wade too far into micromanagement of ICE's processing system and is unnecessary in light of the other measures the defendants are ordered to adopt.

V

It is thus ordered as follows:

Dorm for positives: The defendants shall continue to reserve a dorm at Mesa Verde for the purpose of isolating as a cohort any detainees who test positive for Covid-19. This dorm must remain empty except to house positive detainees. After a detainee has met the CDC criteria for non-contagion, the defendants can move the detainee from the positive dorm into the general population.

Intake procedures: The defendants shall perform a pre-intake screening for Covid-19 symptoms, including a temperature check and verbal symptoms check, for all new intakes arriving to Mesa Verde before the intakes enter the facility or just inside the facility.

The defendants shall offer a point-of-care test to all intakes. If the intake tests positive, the defendants can move that intake directly into the positive dorm. If the intake refuses, the defendants shall report that to class counsel within two hours so that class counsel can discuss the need for testing with the intake. After class counsel has spoken with the intake, the defendants shall re-offer the Covid-19 test. If the intake continues to refuse testing, the defendants shall isolate that intake in one of the individual cells.

The defendants shall reserve a dorm at Mesa Verde for the purpose of quarantining intakes who test negative upon arrival. Intakes arriving within 48 hours of each other can be cohorted together. The quarantine shall last for 14 days, and the time shall begin from the last intake's arrival. If the defendants are accepting new intakes, the intake quarantine dorm must remain empty at all times except for purposes of quarantining. If the defendants are not accepting new intakes, they do not need to keep a quarantine dorm empty, and are free to use three dorms to house detainees in the general population.

Per-dorm population limit: The defendants shall limit the population of each dorm in Mesa Verde to allow detainees to sleep in bunk beds that are at least six feet apart. This population cap does not apply to the dorm used for positive cases.

Saturation testing: The defendants shall administer a Covid-19 test that returns rapid results to all detainees at Mesa Verde and to Mesa Verde staff every week. This includes testing detainees who have been admitted through intake and are cohorting in the intake cohort dorm. The defendants are not required to test detainees or staff who have tested positive for Covid-19 within the past 90 days, unless they begin presenting symptoms.

If a Mesa Verde staff member tests positive, the defendants shall ensure that he or she does not return to work for as long as necessary to meet the CDC requirements for discontinuing home isolation. The defendants shall immediately offer a Covid-19 test to any detainee who had close contact with the staff member who tested positive, and quarantine that detainee for 14 days. If the detainee refuses testing, the defendants shall report that fact to class counsel within two hours, and shall re-offer a test to that detainee after communicating with class counsel.

If a detainee tests positive, the defendants shall immediately move the detainee to the positive dorm. After the detainee is deemed to have recovered under the relevant CDC criteria, the detainee may be moved back to a general population dorm.

If a detainee tests positive and is in a dorm, the defendants shall immediately offer tests to all other detainees in that dorm, as well as any other detainees who had close contact with the detainee who tested positive. The dorm, and any close contacts, shall also be quarantined for 14 days. If any detainee refuses testing, the defendants shall report that fact to class counsel within two hours, and shall re-offer a test to that detainee after communicating with class counsel.

Symptomatic detainees: The defendants shall individually isolate in an individual cell any detainee presenting symptoms of Covid-19 and offer that detainee a point-of-care test. If the detainee tests positive, the detainee can be moved into the positive dorm. If the detainee tests negative, the defendants shall keep the detainee in isolation until the negative point-of-care test is confirmed by a laboratory test.

Transfers: The defendants shall not transfer detainees out of Mesa Verde unless necessary for medical evaluation, medical isolation/quarantine, clinical care, extenuating security concerns, release or removal, or to prevent overcrowding.

Reporting: The defendants shall report to class counsel and the Court any new positive Covid-19 test from detainees or Mesa Verde staff on the day that the defendants learn of it. In addition, the defendants shall file a weekly status report on the docket every Monday by 5 p.m. with staff testing results (positive tests that testing period and cumulative positive tests since the start of the pandemic); the number of staff not at work because of Covid-19; intake and detainee testing results (positive tests that testing period and cumulative positive tests since the start of the pandemic); the housing location of each detainee; descriptive information about any active Covid-19 cases, including information about severe cases and hospitalizations; and the number and identity of new intakes.

VI

The plaintiffs’ motion for a second preliminary injunction is granted in part. The defendants, and all of their officers, agents, servants, employees, and attorneys, are ordered while this case is pending to comply with the requirements set forth in Section V to protect against the spread of Covid-19 at Mesa Verde. This ruling does not disturb the earlier preliminary injunctive relief issued, which remains in effect.

IT IS SO ORDERED.