United States v. EES Coke Battery, LLC

Opinion

UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF MICHIGAN SOUTHERN DIVISION

UNITED STATES OF AMERICA, Case No. 22-11191 Plaintiff, v. Gershwin A. Drain United States District Judge EES COKE BATTERY, LLC; DTE ENERGY SERVICES, INC.; Curtis Ivy, Jr. DTE ENERGY CO.; and DTE United States Magistrate Judge ENERGY RESOURCES, LLC, Defendants. ____________________________/

OMNIBUS OPINION AND ORDER ON ECF Nos. 145, 156, 158, 161, 164, 165, 166, 221, and 269

Several motions to exclude expert testimony were referred to the undersigned. Defendants’ motion to seal was also referred. The Court heard argument on the motions on May 28, 2025. For the reasons below, • Defendants’ Motion to Exclude Galinsky at ECF No. 145 is GRANTED IN PART to preclude Galinsky from testifying in rebuttal to Mr. van der Poel;

• Defendants’ Motion to Exclude Galinsky at ECF No. 156 is DENIED;

• Defendants’ Motion to Exclude Chinkin in part at ECF No. 164 is GRANTED IN PART to strike the HYSPLIT opinions in the rebuttal report;

• The government’s motion to strike the new argument in two of Defendants’ reply briefs (ECF No. 221) is GRANTED; • Defendants’ Motion to Strike Dr. Schwartz at ECF No. 166) is GRANTED IN PART; only the opinions on years 2023-2028 in the rebuttal report are stricken;

• The government’s Motion to Strike Dr. Kelly at ECF No. 165 is DENIED;

• Sierra Club’s Motion to Strike Dr. Kelly at ECF No. 158 is DENIED AS MOOT;

• The government’s Motion to Strike Leahy at ECF No. 161 is GRANTED; • Defendants’ Motion to Seal at ECF No. 269 is GRANTED.1 A. Defendants’ Motions to exclude the 2d and 3d disclosures from United States’ Expert Virginia Galinsky (ECF Nos. 145, 156)

Ms. Galinsky is an environmental engineer at the Environmental Protection Agency who has been involved in the enforcement action against Defendants since its inception. She was deposed twice, once as a fact witness and once as the EPA’s Rule 30(b)(6) corporate representative. After those depositions, the United States designated Galinsky as a non-retained employee expert witness under Fed. R. Civ. P. 26(b)(2)(C) on April 29, 2024, its deadline for expert disclosures. The government served a second disclosure from Galinsky on November 25, 2024, the date rebuttal reports were due, that included opinions not addressed in the first

1 Several documents cited in this Order are the subject of the motion to seal. This Order, however, cites to those documents as originally filed, not as newly filed on May 12, 2025. To avoid delaying resolution of the motions to exclude any further, the Court is not adjusting the citations to the newly filed documents. The parties and the public now have adequate access to all the documents relevant to this Order and are able to understand the bases for the determinations. 2 disclosure. On December 18, 2024, it served Galinsky’s third disclosure with additional opinions.

Given the timing and subject matter of the second and third disclosures, the impetus for them appears to be the Court’s Order striking the government’s experts Mr. Benson and Dr. Sahu. (ECF No. 191). To state it simply, those would-be

experts opined on Best Available Control Technology (“BACT”) and Lowest Achievable Emission Rate (“LAER”) at EES Coke. Galinsky opines on BACT and LAER in her later disclosures. Defendants bring two motions to exclude Galinsky’s second and third

disclosures. The first motion contests whether Galinsky can rightly be considered a non-retained expert witness who need not provide an in-depth expert report under Rule 26(a)(2)(B). (ECF No. 145). The second motion challenges her qualification

to opine on the areas addressed in the second and third disclosures, as well as the reliability of her analyses. (ECF No. 156). 1. Galinsky’s Designation as a Non-Retained Expert Defendants insist that Galinsky’s opinions in the second and third

disclosures were not formed as part of her normal job responsibilities, so she was acting as a retained expert. (ECF No. 145, PageID.4997-98). Thus, they argue that Galinsky was required to provide detailed disclosures required of retained

3 experts under Rule 26(a)(2)(B) (a “(2)(B)” report) rather than the summary report required for non-retained experts under Rule 26(a)(2)(C) (a “(2)(C)” report). (Id.

at PageID.5000). The United States argues that Galinsky’s work duties “include developing and supporting enforcement actions brought on behalf of EPA.” (ECF No. 148,

PageID.5191). Given the nature of her work, the government insists that there is no practical way to parse opinions formed as part of the job from those formed for purposes of the litigation. (Id. at PageID.5191-92). In other words, supporting the EPA’s litigation efforts is her job. But, at the hearing, counsel confirmed that this

is the first time Galinsky has reviewed opposing expert reports and drafted rebuttal opinions. Still, counsel said that environmental engineers at the EPA have, in other enforcement actions, reviewed opposition expert reports and provided

testimony in rebuttal to them. If an expert witness is “retained or specially employed to provide expert testimony in the case or . . . whose duties as the party’s employee regularly involve giving expert testimony,” then the expert disclosure must include a host of

information including a complete statement of all the opinions and the grounds for them. Fed. R. Civ. P. 26(a)(2)(B)(i-vi). A non-retained expert or an employee

4 whose duties do not regularly involve giving expert testimony need only provide a summary of the facts and opinions. Fed. R. Civ. P. 26(a)(2)(C)(i-ii).

As explained in the Court’s order on other motions to exclude experts (ECF No. 259), Rule 26(a)(2)(C) applies to “hybrid witnesses,” i.e., fact witnesses who can also provide expert testimony. Call v. City of Riverside, 2014 WL 2048194, at

*3 (S.D. Ohio May 19, 2014). A common example of a hybrid witness is a treating physician. Physicians are experts because of their education, training, and experience. They can also be fact witnesses when testifying about the care they provided to a patient. Fielden v. CSX Transp., Inc., 482 F.3d 866, 870-71 (6th Cir.

2007) (holding that no (2)(B) report required for treating physician who would give expert testimony about causation learned by providing treatment). When they testify about the facts through their expert lens, they are acting as a hybrid witness.

“Put another way, [hybrid witness] testimony arises not from . . . enlistment as an expert but, rather, from . . . ground-level involvement in the events giving rise to the litigation.” Downey v. Bob’s Disc. Furniture Holdings, Inc., 633 F.3d 1, 6 (1st Cir. 2011) (citing id. at 869).

Deciding Galinsky’s status is not as straight forward as it often is when treating physician testimony is involved because of the nature of her job. There is

5 not always a clear line between supporting an enforcement action as an employee and supporting a party in an enforcement action as a retained expert.

The Court has some guideposts. In determining whether Galinsky should be considered a retained expert witness or a hybrid, non-retained witness, the focus is on the substance of her opinions, not her status as an employee. Mohney v. USA

Hockey, Inc., 300 F. Supp. 2d 556, 560 (N.D. Ohio 2004) (“[T]he application of the Rule 26 disclosure requirements depends on the substance of the treating physician’s testimony rather than his or her status.”) (quoting Hawkins v. Graceland, 210 F.R.D. 210, 211 (W.D. Tenn. 2002)); Fielden, 482 F.3d at 871

(“The determinative issue is the scope of the proposed testimony.”) (citation omitted). If the scope of the expert’s testimony strays from the “core” of their job, then a (2)(B) report will be required. Fielden, 482 F.3d at 870. Another factor to

consider is the kinds of documents or evidence the expert reviewed in developing their opinions. A witness acts like a retained expert when they review materials that they do not rely on in the ordinary course of their employment. Id. at 872. If “the facts which form the basis for [an expert’s] opinions derive from information

learned during the [ordinary course of the job]—as opposed to being subsequently supplied by an attorney,” then no (2)(B) report is required. Mohney, 300 F. Supp. 2d at 560 (quoting Sullivan v. Glock, Inc., 175 F.R.D. 497, 501 (D. Md. 1997)).

6 The best place to start is to compare the scope and substance of Galinsky’s first disclosure, which Defendants do not challenge, with her second and third

disclosures. Because the first disclosure is not challenged, the Court assumes that the scope of that disclosure is permissible (2)(C) expert opinion. The first disclosure covers two topics: permitting history and mitigation of excess

omissions. Galinsky opines that the changes in emissions following the 2014 permit are a “major modification.” (ECF No. 145-6, PageID.5034-35). She then addresses several ways EES Coke could mitigate excess omissions, including reducing sulfur dioxide (SO ) beyond what is necessary to comply with BACT and 2 LAER and reducing emissions from outside sources. (Id. at PageID.5035-36). She addresses the same two topics in the second disclosure but expands on them. She discusses emissions changes following the permit including what regulations apply

to the permit and whether changing certain phrasing in the permit would have prevented the enforcement action, and she opines on how Defendants’ expert Mr. Campbell’s permit-related analyses are flawed. (ECF No. 144-3, PageID.4957- 62). Galinsky then addresses her opinion that installation of BACT and LAER are

required for pollution control and the data that informs the BACT/LAER analyses. She also discussed other mitigation options. (Id. at PageID.4962-71). The third disclosure only touches on mitigation. She discusses the determination of BACT

7 and LAER in rebuttal to Defendants’ experts Mr. van der Poel and Mr. Saini, including the effect of compliance with LAER. (ECF No. 144-4, PageID.4974-

76). The only substantive difference between the first disclosure and second and third disclosures is the addition of discussion to rebut one or more of Defendants’

experts. The topics and general substance are the same throughout the opinions. The scope of all three sets of opinions matches the scope of her employment—an environmental engineer who participates in and supports Clean Air Act enforcement actions. Attached to that response brief is a new declaration from

Galinsky wherein she describes her job responsibilities. She says her job responsibilities include: • Identifying sources of potential violations • Conducting inspections • Drafting and reviewing Clean Air Act Section 114 information requests • Reviewing and analyzing case files • Drafting Notices of Violation • Undertaking discussions and negotiations regarding resolution of enforcement actions, and • Drafting documents to memorialize the resolutions

(ECF No. 186-1, PageID.13379-80, at ¶ 2). She adds that, to resolve an enforcement action, the “case team” determines what injunctive relief is necessary to return the source to compliance and what penalty is appropriate to resolve the 8 violations. That injunctive relief should, according to Galinsky, at minimum require BACT and LAER, depending on the case. (Id. at PageID.13380, ¶ 3).

“[T]herefore, as a normal course of an action in [a New Source Review] enforcement case, [she] would determine what BACT or LAER would likely be by following” the “top-down” process from a 1990 EPA manual or any state guidance

on implementing new source review. (Id. at PageID.13380-81, ¶ 4). The work she did to form the opinions in the second and third disclosures appears to be the same kind of work she would do even if this litigation did not exist, except for the addition of addressing Defendants’ experts’ reports within her

opinions. Even there, addressing or rebutting a facility’s experts who oppose the enforcement action fits the bill of “supporting enforcement actions.” The documents Galinsky reviewed while crafting her opinions appear to be

the kind of documents EPA environmental engineers typically review in enforcement actions. In the last paragraph of the declaration, she states, “As part of my work on the EES Coke enforcement case, which began in 2017, I have reviewed thousands of documents.” (ECF No. 186-1, PageID.13382, ¶ 9). No

specificity was given on the kinds of documents she reviewed, though it appears

9 that at least some of those documents were obtained through discovery in this litigation.2 Obviously, included in these are Defendants’ experts’ disclosures.

In the context of a treating physician testifying as an expert, (2)(B) expert disclosures are usually required when the physician reviews documents supplied by counsel to assist in forming an opinion. In Mahoney, for instance, the

physician’s opinions drawn from watching video of the accident after treating the plaintiff were excluded because the physician did not submit a (2)(B) report. 138 F. App’x at 810. “There [was] no evidence that [the physician] reached the same conclusions regarding causation at the time he treated [the plaintiff].” Id. at 811.

In Adkins v. Marathon Petroleum Co. LP, 105 F.4th 841 (6th Cir. 2024), a physician’s (2)(C) report was stricken after finding that he acted as a retained expert rather than a hybrid witness. The physician did not assess the plaintiff until

months after the plaintiff filed the lawsuit. The plaintiff visited the physician at the request of counsel, sometimes with counsel present, and counsel paid for the office visits. Id. at 849. And the physician reviewed documents provided by the plaintiff’s attorney, which meant that the opinions were not formed in the ordinary

course of treatment. Adkins v. Marathon Petroleum Co. LP, 672 F. Supp. 3d 483,

2 The EPA received what is likely a large number of documents in response to its information request issued April 12, 2018. Documents in response to that request included emissions data and permitting documents. (ECF No. 9-2, PageID.114-15, First Galinsky Declaration). 10 511 (S.D. Ohio 2023). These facts led to the conclusion that the doctor assessed the plaintiff and formed causation opinions at the request of counsel for the

litigation. For these reasons, the court affirmed that the physician was retained for litigation purposes and was required to submit a (2)(B) expert report. Adkins, 105 F.4th at 849. So when the physician’s opinions are based only on the facts learned

while treating the patient and not formed outside the time of that treatment, the physician is a hybrid witness. See Oetjens v. Covidien LP, 2025 WL 227261, at *7 (E.D. Mich. Jan. 16, 2025). The nature of a treating physician’s job allows for an easier distinction

between retained and non-retained expert. “It is within the normal range of duties for a health care provider to develop opinions regarding causation and prognosis during the ordinary course of an examination.” Fielden, 482 F.3d at 870 (citation

omitted). When a doctor steps out of that context and reviews evidence beyond what is presented while treating the patient, the physician changes roles and becomes a more traditional expert witness forming causation opinions based not only on what was learning during treatment, but also on after-the-fact documents

or other evidence that informs the cause of the injury or illness. The nature of Galinsky’s job is entirely different. She does not stop acting in her capacity as an EPA environmental engineer supporting an enforcement

11 action by reviewing documents obtained during litigation, even though some of those documents came from counsel for her employer. Thus, the treating physician

cases are not an exact fit for Galinsky. Still, applying the principles discussed in the cases cited above leads to the conclusion that Galinsky did not go beyond the scope of her employment in reviewing case documents, including expert reports, in

forming her opinions about the merits of the EPA’s enforcement action. The Court finds that Galinsky was properly designated as a (2)(C) expert for all three disclosures. The scope of each disclosure does not go beyond the scope of her employment. That Galinsky reviewed other expert reports and stated her

opinions in opposition to those experts does not change the conclusion. Defendants raised a new argument at the hearing—only Galinsky’s opinions about matters that occurred since February 2017 would be permissible, if any of

her opinions are permitted. This is because Galinsky became aware of a potential issue at EES Coke during February 2017, that is when enforcement work began. Were the Court to agree with Defendant, it would preclude Galinsky from opining on the permitting process or other events occurring before February 2017. In

making this argument, Defendants suggested that Galinsky is just like their non- retained expert Eric Marko. Defendants intended to call Marko, an employee of a consulting firm hired by EES Coke, to testify about the 2012-2014 permitting

12 process. Marko’s employment with the consulting firm did not begin until years after the permits were approved. But with Marko, there was no evidence that his

job duties involved reviewing past permitting and events surrounding permitting. Thus, his testimony about the history of the permitting process went beyond the scope of his employment.

Galinsky is not in the same position as Marko. EPA environmental engineers assist the EPA in enforcement actions. This would reasonably include reviewing EPA permits and what occurred during the permitting process to understand whether the EPA indeed has a case for enforcement against a facility.

So setting aside the fact that this argument was raised for the first time at the hearing, the Court finds no merit in the argument. Galinsky will not be limited only to matters occurring February 2017 and after.

2. Timeliness of Reports Defendants argue that the second and third disclosures are untimely because (1) the second disclosure, served on the deadline for rebuttal reports, is not a rebuttal opinion and (2) the third disclosure was filed after the rebuttal disclosure

deadline. The government contends that the second disclosure is a rebuttal report served on the rebuttal expert disclosure deadline. Defendants disagree because the

13 substance of the report supports the government’s case in chief. (ECF No. 152, PageID.5390). When pressed that Galinsky expressly responds to Defendants’

experts in the second disclosure, counsel further explained the argument— Defendants’ experts rebut the government’s expert, so Galinsky’s “rebuttal” to the rebuttals is in effect an opinion that supports the case in chief.

The government has the better of the argument on the second disclosure. Rule 26(a)(2)(D)(ii) contemplates that an opinion “intended solely to contradict or rebut evidence on the same subject matter” as an opponent’s expert will be disclosed within 30 days after the opponent’s disclosure or on a date set by the

Court. The Court set the rebuttal disclosure deadline for November 25, 2024. (ECF No. 113). Rebuttal reports due after Defendants’ expert reports would necessarily, or at least in the normal course, rebut Defendants’ experts who rebut

Plaintiff’s experts. Galinsky purports to opine on the alleged flaws in Defendants’ experts’ reports, which is rebuttal report content. See Telepak Networks, Inc. v. City of Memphis, 2014 WL 5795499, at *2 (W.D. Tenn. Nov. 6, 2014). Simply because Galinsky has an initial report and a rebuttal report does not convert the

rebuttal to an initial report. At trial, she will not be permitted to testify in rebuttal to Defendants’ experts until Defendants’ experts have testified (absent some scheduling agreement). With that guardrail in place, her rebuttal in the second

14 disclosure is not converted to case-in-chief expert testimony. The second disclosure was properly disclosed as a rebuttal disclosure on the rebuttal report

deadline. That said, the District Judge will be in the best position to assess whether Galinsky’s testimony constitutes rebuttal and can rule on objections at that time. As for the third disclosure, Defendants do not appear to contest that this is a

rebuttal, but they insist that it was untimely. The third disclosure was served on December 20, 2024, 18 days after the Court struck Benson’s and Dr. Sahu’s reports and a few weeks after the rebuttal disclosure deadline. Defendants served their expert reports on October 7, 2024. The December 18th rebuttal opinion was

not disclosed within 30 days of Defendants’ expert disclosures, so the third disclosure is untimely by application of the Federal Rules of Civil Procedure. The United States cites the Court’s Order striking Benson and Dr. Sahu as

giving permission to submit a rebuttal disclosure so long as doing so would not impact case deadlines. (ECF No. 148, PageID.5186). In that December 2, 2024, Order, the Court noted that the dispositive motion deadline was February 18, 2025, so “Plaintiff [had] an opportunity to provide a new expert report covering the areas

stricken” in that Order. (ECF No. 134, PageID.4766). The Court said that the government “should have had a contingency plan to address” the potential of having its experts stricken. (Id.). By acknowledging that there was time before

15 dispositive motions were due for the government to identify a witness to cover the loss of Benson and Sahu, the Court was not extending the government’s opening or

rebuttal expert deadline. The government needed the Court’s permission to submit an expert disclosure beyond Court-created deadlines, certainly absent consent from Defendants.

The third disclosure is untimely. Rule 37(c)(1) requires exclusion of an expert report that is not properly disclosed under Rule 26(a), unless the failure was substantially justified or harmless. A harmless violation “involves an honest mistake on the part of a party coupled with sufficient knowledge on the part of the

other party.” Howe v. City of Akron, 801 F.3d 718, 747 (6th Cir. 2015) (internal quotations and citations omitted). “Substantially justified means justified to a degree that could satisfy a reasonable person.” Eagle v. Hurley Med. Ctr., 292

F.R.D. 466, 481 (E.D. Mich. 2013) (internal quotations omitted). There is a five- factor test used to determine whether the strike the improper evidence. Those factors are: (1) the surprise to the party against whom the evidence would be offered; (2) the ability of that party to cure the surprise; (3) the extent to which allowing the evidence would disrupt the trial; (4) the importance of the evidence; and (5) the nondisclosing party’s explanation for its failure to disclose the evidence.

16 Burks v. Washington, 2023 WL 5228906, at *3 (E.D. Mich. Aug. 15, 2023) (quoting Howe v. City of Akron, 801 F.3d 718, 748 (6th Cir. 2015)).

The government asserts that the factors weigh against exclusion of the disclosure. They insist that there is no surprise to Defendants, that Defendants can cure any surprise by deposing Galinsky, the evidence is important, and the

government’s explanation for their failure is reasonable. (ECF No. 148, PageID.5198). Defendants insist there is surprise—the disclosure does not explain how or why Galinsky reached her conclusions. They argue that even if the government

provides that information, they would be unable to adequately respond now that discovery is closed and dispositive motion briefing is finalized. (ECF No. 152). Some factors weigh in the government’s favor. Given that this issue is now

limited to the third disclosure, and the third disclosure does not state mitigation opinions different from those in the second disclosure, the surprise to Defendants is lessened. That said, Galinsky rebuts an expert in the third disclosure unaddressed in her disclosures—Mr. van der Poel. Defendants have not had a chance to probe

those rebuttal opinions. Nor does it appear that allowing testimony from the third disclosure would disrupt the trial. And it appears that the third disclosure is important to the government’s case now that Benson and Dr. Sahu have been

17 stricken, but with the second disclosure available for testimony at trial, the importance of the third disclosure is somewhat lessened.

Some factors weigh in Defendants’ favor. Because of the timing of the disclosure and motion, Defendants cannot cure the surprise. Their only option to address Galinsky’s testimony is through cross-examination at trial. And the

government does not have the most reasonable explanation for failing to disclose the report sooner. True, it did not know that its experts Benson and Dr. Sahu would be stricken until early December 2024. But nothing in the record suggests that Galinsky could not have included the opinions in the second disclosure rather

than creating a third disclosure later—the experts she rebuts submitted their reports in October 2024. (ECF No. 149). The United States does not contend that Galinsky required more than 30 days to form rebuttal opinions. Their only basis

for providing the third disclosure in late December 2024 is that their other experts were stricken. As the Court noted in the Order striking their reports, the government should not have stood idly by during conversations and litigation about Benson and Dr. Sahu; it should have gotten the information contained in the

third disclosure much sooner, even if it would only serve as a back-up to its other experts.

18 Because there is surprise for which there is no time to cure, because the government has no good reason for not disclosing the opinions in the third

disclosure sooner, and because the government has Galinsky’s first and second disclosures which largely cover what is addressed in the third disclosure, the third disclosure is STRICKEN IN PART. Given the significant overlap in the opinions

offered, or summarized to be offered, in all three disclosures, the effect of this Order is to preclude Galinsky from offering testimony in rebuttal to Defendants’ expert Mr. van der Poel. She rebuts Saini’s opinions in both the second and third disclosures. It would likely be incredibly difficult to parse permissive rebuttal to

Saini if Galinsky was limited to rebuttal stated in the second disclosure because of the overlap between the two disclosures. And since Saini is rebutted in the second disclosure, Defendants cannot claim serious surprise. So only the portions of the

third disclosure that addresses or is raised in response to van der Poel’s expert opinion are stricken. Because there is no unwarranted “surprise” from the second disclosure and continued rebuttal of Saini’s report, the Court will not order Galinsky to sit for a

deposition. Defendants have enough information to prepare for Galinsky’s trial testimony and adequately cross-examine her. 3. Galinsky’s Qualification and Reliability

19 Defendants argue that Galinsky only has cursory experience with BACT and/or LAER analyses. (ECF No. 156, PageID.5410-11). Thus they insist that her

opinions on BACT and/or LAER and the effect of compliance are unreliable because they are not based on any specialized knowledge. The United States contends that she is sufficiently qualified having done BACT and/or LAER

analyses 14 times. And much like Defendants’ expert Saini, Galinsky’s job entails determining what BACT and/or LAER are for specific situations. (ECF No. 186, PageID.13372). The government says that her disclosures are detailed enough to explain how she reached her BACT and LAER opinions. (Id. at PageID.13375-

76). Federal Rule of Evidence 702 requires the trial judge to perform a “gatekeeping role” when considering the admissibility of expert testimony.

Daubert v. Merrell Dow Pharm. Inc., 509 U.S. 579, 597 (1993). The United States Supreme Court has established that Rule 702 requires district courts to ensure that expert testimony “rests on a reliable foundation and is relevant to the task at hand.” Id.; see also Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152 (1999)

(expanding Daubert’s analysis of expert scientific testimony to cover expert testimony based on “technical” and “other specialized knowledge”). The gatekeeping role progresses in three steps. First, the witness must be qualified

20 according to his or her “knowledge, skill, experience, training, or education.” In re Scrap Metal Antitrust Litig., 527 F.3d 517, 529 (6th Cir. 2008) (quoting Fed. R.

Evid. 702). Second, the expert’s testimony must be relevant in that it will help “the trier of fact to understand the evidence or to determine a fact in issue.” Id. Third, the testimony must be reliable. To determine whether expert testimony is

“reliable,” the court’s role, and the offering party’s responsibility, “is to make certain that an expert . . . employs in the courtroom the same level of intellectual rigor that characterizes the practice of an expert in the relevant field.” Kumho Tire Co., 526 U.S. at 152. The proponent of the testimony bears the burden to prove by

a preponderance of the evidence that the testimony is reliable. Wellman v. Norfolk & W. Ry. Co., 98 F. Supp. 2d 919, 923 (S.D. Ohio 2000) (citing In re Paoli R.R. Yard PCB Litig., 35 F.3d 717, 744 (3d Cir. 1994)). Courts need not admit opinions

or conclusions that are “connected to existing data only by the ipse dixit of the expert.” See Gen. Elec. Co. v. Joiner, 522 U.S. 136, 146 (1997). As explained in the earlier Order on other motions to exclude experts, “a proffering party can qualify their expert with reference to h[er] ‘knowledge, skill,

experience, training or education.’” Surles ex rel. Johnson v. Greyhound Lines, Inc., 474 F.3d 288, 293 (6th Cir. 2007).

21 Galinsky is sufficiently qualified. She has been an environmental engineer in the Air Enforcement and Compliance Assurance Branch of the EPA since July

2009. (ECF No. 186-1, PageID.13379; ECF No. 9-2, PageID.99). Her role involves supporting EPA enforcement actions for violations of the Clean Air Act. (ECF No. 9-2, PageID.100 at ¶ 5). Her job requires that she achieve emissions

reductions at sites that are at least equivalent to BACT and/or LAER, so it is in the normal course for her to determine what BACT or LAER would likely be. (ECF No. 186-1, PageID.13380-81, at ¶ 4). In doing that work, she has “evaluated new and existing control technologies for appropriateness in controlling a source’s

emissions, reviewed the RACT/BACT/LAER clearinghouse, reviewed permits to determine the control technologies used and limits met by other similar sources, reviewed state and federal regulations for relevant emission limits, communicated

with vendors about the suitable applications for certain control technologies,” among other things. (Id. at PageID.13381, ¶ 4). She has conducted around 50 New Source Review compliance investigations and has evaluated BACT and/or LAER for “at least” 14 facilities in her 15-year career. (Id. at PageID.13382, ¶ 7).

All that said, in response to an interrogatory, the United States explained that Galinsky never personally conducted a BACT or LAER analysis for purposes of New Source Review permitting, but she has conducted similar analyses to

22 determine appropriate pollution control technology and emissions limits. (ECF No. 148-5, PageID.5238). Defendants make note of the fact that she never

personally conducted a BACT or LAER analysis. (ECF No. 156, PageID.5416). This is not unusual, according to the government, because Michigan makes the final BACT and/or LAER determination, not the EPA or Galinsky.3 (ECF No.

186, PageID.13372). So that Galinsky did not personally conduct the BACT or LAER analyses for this case does not mean that BACT and LAER play no role in her usual job responsibilities. Defendants also note that Galinsky has no formal education or expertise

relating to BACT and/or LAER and has authored no peer-reviewed papers relating to the same. (ECF No. 156, PageID.5415). While Galinsky might not have taken any courses on BACT or LAER analysis, her education and training as an

environmental engineer sufficiently prepares her to determine what mitigation actions are necessary to bring a facility into compliance with an environmental statute. Moreover, because “[e]xpert opinion may be based entirely on experience,” Innovation Ventures, LLC v. NVE, Inc., 90 F. Supp. 3d 703, 724 (E.D.

3 Defendants’ expert Saini similarly has not personally conducted a BACT or LAER analysis yet opines on BACT and LAER here. At the hearing, Defendants defended Saini’s opinion by noting that he is a retained expert and retained experts can opine outside the area of their direct job responsibility, unlike non-retained experts. As explained here, BACT and LAER are not outside the scope of Galinsky’s work in supporting enforcement actions. 23 Mich. 2015), that Galinsky has authored no papers on this subject is not reason to disqualify her.

In all, Galinsky’s opinions are based on her specialized knowledge as an environmental engineer who focuses on air quality enforcement cases. Her education and work experience determined what is required to bring a facility into

compliance, i.e., BACT and/or LAER, and is sufficient to help the Court in assessing or weighing the evidence. Defendants also attack the reliability of Galinsky’s opinions because of the lack of analysis showing how she reached her opinions based on the materials

reviewed. (ECF No. 156, PageID.5418-19). Galinsky is a non-retained expert, so her (2)(C) reports are sufficient. The rule requires the disclosure to give only the subject matter on which the witness

will present evidence and “a summary of the facts and opinions to which the witness is expected to testify.” Fed. R. Civ. P. 26(a)(2)(C)(i-ii). Only reports required under (2)(b) must include a complete statement of the opinions and the bases for them, any exhibits that would be used, and the facts or data considered.

Galinsky’s second disclosure satisfies Rule (2)(C). In summary, Defendants’ motion at ECF No. 145 to exclude Galinsky’s second and third disclosures is GRANTED IN PART, their motion to exclude at

24 ECF No. 156 is DENIED. Galinsky’s third disclosure is stricken as to her rebuttal of Mr. van der Poel’s opinions.

B. Defendants’ Motion to Partially Exclude the Testimony of US’s Expert Lyle Chinkin (ECF No. 164)

Lyle Chinkin is the government’s “air modeling” expert. He opines on the impacts of EES Coke’s allegedly excessive sulfur dioxide (SO ) emissions on air 2 quality. Before discussing the specifics of his opinion, brief scientific background is helpful to understand his opinions and the challenges to them. Sulfur dioxide is a “primary pollutant.” (ECF No. 164-2, PageID.6175, Chinkin Report). Primary pollutants react with other compounds to produce secondary pollutants such as sulfate PM . Secondary pollutants, i.e., particulate matter, can travel long 2.5 distances before they are removed from the atmosphere. (Id. at PageID.6176). Sulfur dioxide can convert to PM in a couple of hours in a hot, humid 2.5 atmosphere. Sulfur dioxide takes longer to convert, if at all, in a cold, dry atmosphere. (ECF No. 164-3, PageID.6319-20, Chinkin Deposition). Inhalation of PM can have negative health consequences; the parties dispute what 2.5 concentration of PM and the duration of exposure is necessary before negative 2.5 health consequences are realized. Chinkin used two computer air modeling programs to demonstrate the air quality impacts from SO emissions from EES Coke—the Comprehensive Air 2 25 Quality Model with Extensions (“CAMx”) and the American Meteorological Society/Environmental Protection Agency Regulatory Model Improvement

Committee Dispersion Model (“AERMOD”). (ECF No. 164-2, PageID.6173). Only his use of and opinions drawn from the CAMx program are contested. CAMx is a photochemical grid model that simulates emission, transport,

diffusion, chemical transformation, and removal of gaseous and particle pollutants in the atmosphere. (Id. at PageID.6187). CAMx outputs are used in later analyses to quantify “the impact of excess SO2 emissions from EES Coke Battery on ambient PM concentrations in the region” surrounding the EES Coke facility. 2.5 (Id. at PageID.6169). The opinions Chinkin drew from CAMx include that EES Coke’s emissions had local and long-range air quality impacts on PM ; that if EES Coke reduced 2.5 emissions, there will be local and long-range air quality benefits; and that the populations impacted by the alleged excess SO emissions are also impacted by 2 other local and regional sources. (Id. at PageID.6169-70). Chinkin also provided a rebuttal report containing opinions that Defendants characterize as new and untimely opinions that they seek to strike. 1. CAMx Opinions

26 Defendants argue that Chinkin’s CAMx opinions should be stricken because they are unreliable for four reasons: (1) the CAMx program, like other modeling

programs, has inherent uncertainties, (2) CAMx overpredicted the PM 2.5 concentration compared to actual concentration data, (3) Chinkin did not properly calibrate the model, and (4) Chinkin did not validate the results despite knowing of the inherent uncertainties. (ECF No. 164). Some of the more specific issues that fall within those four categories are that Chinkin relied on 2016 meteorological data that lead to inaccuracies in calculating concentrations in 2019, that the Model Performance Evaluation he conducted to test his results was insufficient to test the

model’s accuracy, and the differences in concentrations shown in the model were so small that their size led to further errors in the modeling. The first main argument tests the reliability of CAMx generally, not

Chinkin’s use of the program. As mentioned, “[f]our inquiries guide the reliability analysis: Is the technique testable? Has it been subjected to peer review? What is the error rate and are there standards for lowering it? Is the technique generally accepted in the relevant scientific community?” United States v. Gissantaner, 990

F.3d 457, 463 (6th Cir. 2021) (citing Daubert, 509 U.S. at 585-89). CAMx is a reliable air modeling program used by air quality scientists. The program is testable and has been peer-reviewed “extensively.” Defendants do not

27 challenge Chinkin’s statement that the program has been often used to estimate PM concentrations. (ECF No. 164-2, PageID.6201). Nor do Defendants 2.5 challenge that CAMx is a generally accepted program for estimating concentrations of pollutants in the atmosphere. “What matters is whether the relevant scientific community accepts the software.” Gissantaner, 990 F.3d at 466

(citing Daubert, 509 U.S. at 594). Once that is determined to be the case, cross- examination “is the place to go for accuracy.” Id. So any inherent uncertainty in using a modeling program can be addressed in cross-examination. Chinkin then needed to have “reliably applied” principles and methods to

CAMx to satisfy Fed. R. Evid. 702. This is where Defendants’ three remaining areas of concern are directed. Defendants contend that the CAMx program overpredicted particulate

concentrations, so it is unreliable. The purported overprediction comes to light when looking at the results from the Model Performance Evaluation. This evaluation involves comparing results using base numbers provided by the EPA (the results that “overpredicted” PM ) and results using actual concentrations 2.5 from fixed air monitoring locations. (ECF No. 164-2, PageID.6207). Chinkin explains that differences between base or modeled results and actual concentrations does not suggest that the test is unreliable. He explains that the

28 base test calculates average air pollutant concentrations in a select area encompassing a 12-square-kilometer grid (the entire continental United States is

broken down by 12-square-kilometer grids for air quality modeling). The test using actual numbers draws those numbers from a fixed point in different areas and may represent concentrations from somewhat different time periods. Thus,

modeled outputs will never exactly match measured concentrations at an air quality monitor. (Id.). Knowing that these modeled differences will exist, the scientific community recommends the following error performance criteria for 24-hour PM 2.5 concentrations: a normalized mean base of less than or equal to plus-or-minus 30- 65% and a normalized mean error of less than or equal to plus-or-minus 50-115%. (Id. at PageID.6208). So long as the means are within these ranges, the base case

model is considered reliable or at least unproblematic. Defendants look at some of the calculations to suggest that CAMx overpredicted PM 29 out of 30 times. (ECF No. 164, PageID.6139-40). For 2.5 example, the observed mean of PM concentration (i.e., the actual value taken 2.5 from a monitoring location) in Michigan in Spring was 7.6 micrograms per cubic meter. The mean modeled for that same time was 8.3 micrograms per cubic meter.

29 (Id. at PageID.6139). Of the 30 rows of data Defendants provided, the mean modeled exceeded the mean observed 29 times.

The government points out some issues with Defendants’ argument. For one thing, there is more data than the 30 rows Defendants provided. (See ECF No. 187, PageID.13404-08). But more importantly, the government tells us that the

focus is not on whether the mean modeled equals or exceeds the mean observed. The focus is instead on the normalized mean bias and error data, and the numbers here are within the recommended ranges. (Id. at PageID.13403-04). The disputes about the differences in concentration and that those

differences are within the generally accepted error rate are the kinds of issues more appropriately raised in cross-examination, rather than reason to strike an expert report. The government offered a reasonable explanation for the numbers

Defendants suggest show unreliability. Because the error rates are within the norm, the Court does not find a reason to strike the testimony. Defendants take issue with using 2016 meteorological data to estimate PM 2.5 concentrations in 2019. According to Defendants, 2016 was the warmest year in an eight-year period, while 2019 was the coldest. Sulfur dioxide converts slowly, if at all, in cold weather. (ECF No. 164, PageID.6148-50).

30 Chinkin explained that he used meteorological data from 2016 because the weather that year was representative, or average, for weather from 2016 to 2021.

(ECF No. 164-2, PageID.6216-17). Though 2019 was, on average, a cooler year than 2016, on average, this is less of a reliability issue and more fodder for cross- examination.

Finally, Defendants note that the modeling results are “miniscule.” (ECF No. 164, PageID.6150). They argue that the smaller the modeled annual average impact of emissions, the greater the chances for uncertainty and overpredictions. (Id.). They note that the modeled outputs are less than the EPA’s Significant

Impact Levels measurement, suggesting that the predicted PM concentrations 2.5 would have no impact on health outcomes. Because of the small numbers, Defendants insist that Chinkin was required to compare the results to real-world data. (Id. at PageID.6151). The government counters that there is no safe level of PM , so the facility’s pollution levels are not small. (ECF No. 187, 2.5 PageID.13413). This final argument is not well taken. To begin, Defendants cite no support for the notion that the small modeled numbers lend themselves to further uncertainty and overprediction to such an extent that the opinion must be stricken as unreliable. Second, it remains unclear why the measurements having

31 purportedly little effect on air quality suggests that Chinkin’s use of CAMx was unreliable. Instead, these issues go to the weight of Chinkin’s opinions, not the

reliability. This is not the first time some of Defendants’ arguments have been raised against the CAMx modeling program. In United States v. Ameren Missouri, 421 F.

Supp. 3d 729, 786 (E.D. Mo. 2019), the EPA relied in part on Chinkin’s opinions drawn from CAMx results. The defendant raised arguments similar to those presented here to attack the reliability of Chinkin’s results. That court did not find the arguments persuasive.

One of the arguments addressed was that predicted atmospheric concentrations of particulate matter smaller than the EPA’s Significant Impact Levels are meaningless. Id. at 786 (affirmed in relevant part, 9 F.4th 989 (8th

Cir.)). The court relied on evidence in the record to explain that the Significant Impact Levels are used in the permitting process to determine whether creation or modification of a source will lease to air quality violations. Id. at 787. The court dispensed with this argument by noting that the EPA relies on modeled

concentrations below the Significant Impact Level in calculating health benefits. Essentially, it appears that the court determined that comparison to the Significant

32 Impact Levels is not a meaningful measure of an air quality model’s accuracy or reliability. This Court agrees.

Two other arguments mirror Defendants—(1) Chinkin’s use of 2011 meteorology data as representative of other years makes the opinion unreliable and (2) the difference between the 12-square-kilometer grid estimates and the observed

atmospheric levels taken from air monitors render the opinion unreliable. Id. As to the first of these arguments, the court agreed that the model results would have been more precise if the model was used to predict concentration levels for every year in contention. But the court found the EPA’s decision to rely on a

representative year reasonable given the expense of running the model for every year. The court also noted that the defendant did not present sufficient evidence to show that using a representative year made the results unreliable. Id. at 788. As to

the second, the court found more persuasive the fact that models will not perfectly match monitoring data, that monitors provide a measurement of air quality at its location while the modeling program provide average air quality concentration values for a 12-square-kilometer grid. Id. at 789.

The Court’s analysis and conclusions match the Ameren court’s analysis and conclusions. None of the arguments Defendants presented persuasively establish that Chinkin’s CAMx model results are unreliable. Chinkin relied on an air quality

33 monitoring system widely used in the field of environmental science, and he used the model in the manner typically used. Issues with the data he chose to input or

the results go more towards the weight to give the evidence, not the reliability of the evidence. 2. Rebuttal Opinions

Defendants want to strike the portion of Chinkin’s rebuttal report where he opines on the nexus between air quality and emissions from facilities near EES Coke. Chinkin touched on this nexus opinion in his initial report in a section titled, “Analysis of Benefits from Control of Other Sources.” (ECF No. 164-2,

PageID.6235). There, Chinkin explains that based on his AERMOD and CAMx modeling described in the report and the location of other sources of SO emissions 2 in the region, he also analyzed whether emissions controls at those other sources would benefit the same population impacted by EES Coke’s emissions. He considered five other sources including the DTE Monroe Power Plant. He said, Based on my knowledge and experience, the meteorological conditions in the region, and the proximity of the other sources to the EES Coke Battery, it is highly likely that the impacted populations and regions overlap and that there is a nexus between the populations harmed by excess emissions from the EES Coke Battery and those that would benefit from controls at these other sources.

(Id.). 34 For his rebuttal report, Chinkin ran HYSPLIT, a different modelling program, to support the nexus opinion and included opinions about the new

model’s results. He wrote that he “conducted a wind trajectory analysis using the [HYSPLIT] model to explore whether there is a nexus between air quality impacts from illegal excess SO emissions at EES Coke Battery and air quality benefits that 2 would be realized if emission controls were applied at the DTE Monroe plant.” (ECF No. 187-3, PageID.13578). Though included in the rebuttal report, the nexus opinion and HYSPLIT analysis are not offered in direct rebuttal to a defense expert.

The government says that the “rebuttal” opinions (1) are disclosed in the initial report where Chinkin mentioned emissions from the Monroe Plant and (2) provide the “full scope” of the opinions. (ECF No. 187, PageID.13413).

Rule 26(a) contemplates two kinds of expert reports that can be submitted after the initial report: rebuttal reports and supplemental reports. The rule defines a rebuttal report as one where “the evidence is intended solely to contradict or rebut evidence on the same subject matter identified by another party[’s]” expert. Fed.

R. Civ. P. 26(a)(2)(D)(ii). As for supplemental reports, Rule 26(e)(1)(A) requires an expert to supplement their report if they learn that that the report is incomplete or incorrect.

35 In briefing, the government maintains that the nexus opinion is properly included in the rebuttal report but did not show how the nexus opinion is used to

rebut a defense expert. Indeed, counsel for the government confirmed that the two experts Chinkin purports to rebut did not address a connection between heath outcomes from emissions at other sources. Thus, the Court cannot agree that the

nexus opinion is a rebuttal opinion. (See ECF No. 187-3, PageID.13578-81). To be appropriate, then, it must be a supplement. But Chinkin is not supplementing an incomplete or incorrect opinion in the initial report. Instead, as the government says, the nexus opinion is a mere expansion on what he said in his

initial report. Rule 26 contemplates that an expert may “supplement, elaborate upon, [and] explain” his conclusions in the report at trial. Thompson v. Doane Pet Care Co.,

470 F.3d 1201, 1203 (6th Cir. 2006). There is a difference between elaborating on a fully stated and fully supported opinion written in a report and conducting a new analysis to provide new support for a briefly stated opinion in the initial report. To bring that point home, Rule 26 requires a retained expert to provide a complete

statement of all their opinions and the bases and reasons for them. Fed. R. Civ. P. 26(a)(2)(B)(i). The rule does not contemplate the expert adding a new basis for an opinion first stated in the initial report.

36 In the Court’s view, the nexus opinion and HYSPLIT discussion in the rebuttal report are improper additions to the rebuttal report. The HYSPLIT

analysis should have been included in the initial report. Because it wasn’t, the HYSPLIT analysis and opinions flowing from it are STRICKEN. 3. Defendants’ Reply Brief Argument

Defendants’ reply brief make an argument not included in the opening brief—Chinkin’s testimony should be stricken because he relied in part on expert opinion from Dr. Sahu, and Dr. Sahu’s report and testimony were stricken. Among Dr. Sahu’s opinions was that if EES Coke had followed BACT or LAER, it would

have reduced its SO emissions by 95%. (See ECF No. 164, PageID.6142, n. 5). 2 One of the three CAMx tests Chinkin ran used that 95% figure as a control case. (See ECF No. 187, PageID.13396). The government moved to strike this new argument. (ECF No. 221). Defendants contend that they did not have the opportunity to argue about Chinkin’s opinions via Dr. Sahu’s opinion before the reply because the Court had not yet stricken Dr. Sahu’s opinion when the reply was due. (ECF No. 234,

PageID.16629). Dr. Sahu reviewed confidential settlement documents in the course of his work as a retained expert, so they moved to exclude the report as a whole on July

37 29, 2024. (ECF No. 120). The Court granted the motion in part on December 2, 2024, striking Dr. Sahu’s opinion as it related to remedies but not mitigation.

(ECF No. 134). Specifically, the Court struck Sections X through XII of Dr. Sahu’s report. Dr. Sahu discussed his 95%-reduction opinion beginning in Section XI, a section that was stricken in December 2024.4 (See ECF No. 122-5,

PageID.4430). On March 7, 2025, the Court struck Sahu’s opinion in its entirety. (ECF No. 191). Defendants filed their timely reply four days later. Though the report was not stricken in its entirety until just before the reply

was due, the portion of Dr. Sahu’s report that provides the bases for and discussion of the 95%-reduction opinion had been stricken since December 2024, and magistrate judge orders are effective upon filing unless they are stayed. The

Court’s December 2024 Order was not stayed, so the 95% opinion was stricken before Defendants moved to strike Chinkin’s report. This means that they could

4 Defendants assert that the opinion was in Section VIII of Sahu’s report, the portion not stricken in December 2024. (ECF No. 234, PageID.16629). They are incorrect. Section VIII, “Emissions Data Provided to Modeling Expert,” merely explains that he provided the 95%- reduction figure to Chinkin for his air modeling. Dr. Sahu was careful to say that he discusses that figure later in the report. (ECF No. 122-5, PageID.4418). Put another way, Section VIII merely explains what data Sahu gave to Chinkin. This section is not a discussion of how Sahu reaches 95%, it is not opinion. That discussion, the analysis and opinion, come later in Sections XI and XII. Moreover, even if it could be said that there were ambiguity about whether the 95% opinion is raised in Section VIII or a later, stricken section as well, since the opinion is clearly addressed in the stricken sections, the argument was available to Defendants, no matter what was happening with Section VIII. 38 have and should have argued that striking Dr. Sahu’s 95% opinion meant that Chinkin’s reliance on the opinion was tainted.

Given the timeline detailed here, Defendants have offered no good reason why they could not make this argument in the opening brief on Chinkin. Arguments raised for the first time in reply briefs are typically waived. Sanborn v.

Parker, 629 F.3d 554, 579 (6th Cir. 2010) (holding arguments raised for first time in reply brief are waived); Scottsdale Ins. Co. v. Flowers, 513 F.3d 546, 553 (6th Cir. 2008) (same). The Court deems them waived here. The government’s motion to strike this argument is GRANTED.

C. Defendants’ Motion to Exclude US’s Expert Dr. Joel Schwartz (ECF No. 166)

The United States retained Dr. Joel Schwartz “to evaluate the particulate air pollution from the EES Coke Battery Facility, quantify the harm from the pollutants, identify and discuss the health impacts caused by the EES Coke Battery emissions, and provide a comparison of these impacts if proper controls are installed at the EES Coke Facility to eliminate the excess emissions.” (ECF No. 166-3, PageID.7862). Dr. Schwartz is an environmental epidemiologist and professor in the Department of Environmental Health and Epidemiology at the

Harvard School of Public Health. (Id. at PageID.7863). The health effects of air pollution is a “major focus” of his research and work, which includes over 900 39 peer-reviewed papers examining air pollution and airborne particles. (Id.). His work “also includes atmospheric modeling of pollution, exposure modeling, and

risk assessment and benefit-cost analysis.” (Id.). 1. Initial Report Defendants moved to strike the expert report in whole, yet at the hearing

they confirmed that they only want the portion of his opinion addressing any health impacts from the EES Coke facility, the opinions on “quantitative risk assessment,” stricken. They are not challenging Dr. Schwartz’s opinions about health impacts from SO or PM generally. 2 2.5 For his qualitative risk assessment, Dr. Schwartz used the EPA’s Benefits Mapping Analysis Program (“BenMAP”). There are two versions of BenMAP— the Community Edition and the Cloud (online) version. (ECF No. 166,

pageID.7569). Schwartz cited the Community Edition in his initial report, but he used the Cloud version. (See ECF No. 166-2, PageID.7608-09). Among the conclusions drawn from the BenMAP modeling are that EES Coke’s SO emissions in 2019 caused 26 excess deaths and the emissions from 2 2019 to 2022 caused 98 excess deaths. (ECF No. 166-3, PageID.8030). Dr. Schwartz also quantified health impacts from emissions.

40 BenMAP is similar to CAMx in that much of data is pre-loaded into the program. BenMap contains data on population, demographics, and incidence and

prevalence rates of health outcomes that PM causes in the same 12-square- 2.5 kilometers grids for which CAMx estimates exposure to particulate matter. The user will add air quality data. For his analysis, Dr. Schwartz input the CAMx data provided by Chinkin. The CAMx program calculated how much of EES Coke’s SO emissions converted to PM , the concentration of PM , and where that 2 2.5 2.5 particulate matter traveled in the atmosphere. BenMAP compiles all the data and estimates associated health effects such as premature deaths, acute respiratory symptoms, asthma, and Alzheimer’s disease in the 12-square-kilometer grids. Defendants lodge two sets of arguments for striking Dr. Schwartz’s report. The first set of arguments addresses the reliability of the BenMAP program.

Defendants argue that BenMAP is not a generally accepted tool for quantitative risk assessment and the data pre-loaded into the program (population and specific pollutants) is inaccurate. Neither of these arguments is well taken. Defendants insist that BenMAP is only meant to assess health benefits to a

community if pollution controls are in place, not to assess health risks or impacts to a community based on the existing pollutants. This argument is a matter of semantics. According to the BenMAP manual, the test is designed to do what

41 Schwartz used it to do: “quantif[y] the impact of changes in air pollutant concentrations in terms of changes in premature mortality and morbidity in

exposed populations; it also enables users to express those impacts in monetary terms.” (ECF No. 166-5, PageID.8343, BenMAP User Manual). Not only does the manual describe the program in terms that comports with Schwartz’s use, but

the court in Ameren also noted that BenMAP is “a reliable and peer-reviewed EPA risk assessment tool.” 421 F. Supp. 3d at 787. Defendants’ citation to their expert, Dr. Kelly, to support the notion that this use of BenMAP is not generally accepted carries no weight since Dr. Kelly cites no sources herself. (See ECF No.

166-8, PageID.8804, 8826). They point to no court finding BenMAP generally or Schwartz’s use specifically to be unsupported in the scientific community. The pre-loaded population and pollutant data does not render BenMAP an

unreliable tool. The population estimates in BenMAP derive from the 2010 Census population counts. (ECF No. 166-5, PageID.8362; 166-2, PageID.7718). Future population estimates derive from a 2015 source, Woods and Poole, that estimates

population through 2050. (ECF No. 166-5, PageID.8362). This method of estimating population does not account for actual changes in population, which leads Defendants to argue that the population projections are overpredicted. (ECF

42 No. 166, PageID.7581). The Science Advisory Board, an advisory committee to the EPA, issued a report about ways to improve BenMAP dated January 17, 2024.

(ECF No. 166-9). Before discussing improvement recommendations, the report first recognized that “BenMAP and the TSD are essential tools in estimating the scale of health benefits and their associated monetary values in EPA’s regulatory

analysis, in particular Clean Air Act regulations reducing fine particles (PM2.5) and ground-level ozone.” The Advisory Board discussed the population estimation that Defendants use to claim unreliability. The Advisory Board said that “scenario analysis is a more appropriate way to analyze future population, rather than making

deterministic predictions” using Woods and Poole. (Id. at PageID.9026). The data for scenario analysis are freely available and transparently documented, unlike the Woods and Poole data which is “closed, inscrutable.” (Id.).

Though population estimates may be more accurate using data other than that derived from the Woods and Poole method, the report does not suggest or establish that BenMAP’s reliance on the 2010 Census data and Woods and Poole for later years makes the program unreliable. Indeed, the Advisory Board made

the point of noting that BenMAP was an essential tool in that kind of analysis despite the less-than-perfect population data in the program.

43 The second data point challenged is that BenMAP does not distinguish between the varieties of PM (sulfates, nitrates, nickel, lead, etc.). (ECF No. 166, 2.5 PageID.7582). Chinkin’s CAMx modeling analyzed the chemical transformation of SO to PM , and Schwartz used that as the input for air quality data. But the 2 2.5 beta coefficient he used for BenMAP did not account for varying levels of toxicity of the different varieties of PM . In other words, when Schwartz opined that there 2.5 were around 98 premature deaths over a three-year period, those premature deaths are not attributed solely to SO .5 Because Schwartz cannot say that the 98 deaths 2 are attributable to EES Coke’s emissions, Defendants insist that the opinion must be excluded. (ECF No. 166, PageID.7583). That BenMAP does not distinguish between deaths caused by one pollutant from another pollutant does not make the program unreliable. A court in Illinois addressed a nearly identical argument—that Dr. Schwartz did not consider the specific type of particulate matter at issue in the litigation. N.R.D.C. v. Ill. Power Res. Generating, LLC, 2018 WL 5777476, at *3-4 (C.D. Ill. Nov. 2, 2018). The court concluded that this challenge, and others like it, is about the assumptions

made rather than the underlying methodology. “The reliability of data and

5 Defendants assert that Schwartz testified that “sulfate PM2.5 is not nearly as toxic as other specimens of PM2.5.” (Id.). In reality, he said that industry particles and biomass particles are “averagely toxic,” and traffic particles are “more toxic.” (ECF No. 166-2, PageID.7696). 44 assumptions used in applying a methodology is tested by the adversarial process and determined by the [trier of fact]; the court's [gatekeeping] role is generally

limited to assessing the reliability of the methodology—the framework—of the expert’s analysis.” Id. (quoting Manpower, Inc. v. Ins. Co. of Penn., 732 F.3d 796, 808 (7th Cir. 2013). Thus, the court admitted the expert testimony. This Court

agrees that Defendants’ arguments about the data in BenMAP are more appropriately an avenue for cross-examination to attack the weight to be given to the results, rather than an issue with the reliability of the framework of Dr. Schwartz’s analysis. As stated above, “[w]hat matters is whether the relevant

scientific community accepts the software.” Gissantaner, 990 F.3d at 466 (citing Daubert, 509 U.S. at 594). Once that is determined to be the case, cross- examination “is the place to go for accuracy.” Id.

The second set of arguments attacks the reliability of Dr. Schwartz’s use of the program. The arguments are: (1) Dr. Schwartz’s opinions are not based on reliable principles because he used the wrong formula and cited the wrong version of the program, (2) the opinions are unreliable because Dr. Schwartz does not

understand the calculations BenMAP performs, (3) Dr. Schwartz failed to review Chinkin’s CAMx work, and (4) Dr. Schwartz improperly extrapolated the 2019

45 results for years 2020-2022. The Court does not find any of these arguments persuasive.

Dr. Schwartz did not use the wrong formula in BenMAP. During questioning by Defendants’ counsel, Schwartz admitted that the formula written in the initial report was a typo. (ECF No. 166-2, PageID.7666-67). Schwartz used

the correct formula written in his rebuttal report for his analysis. And while Dr. Schwartz said in the initial report that he used the Community Edition of BenMAP, in fact he used the Cloud version. That he wrote the wrong version in his report does not make his Cloud version opinions unreliable. Defendants say their expert,

Dr. Kelly, could not replicate Schwartz’s results using the BenMAP Community Edition, but point to pages of her report that do not address an inability to replicate results. (See. ECF No. 166-8, PageID.8832, 8846). It is unclear how an inability

to replicate results using a version of the program Schwartz did not use would cast doubt on the reliability of Schwartz’s opinions—Schwartz testified that he used the online version, not the Community Edition. Defendants inflate the degree to which Dr. Schwartz does not understand

how BenMAP works. They point to instances in which he did not know the answer to a few questions about the program and cite cases in which expert testimony was stricken because the expert could not explain underlying data. (ECF

46 No. 166, PageID.7579). The deposition excerpts Defendants cite to establish that Schwartz “does not understand the calculations BenMAP performs” do not show

that Schwartz does not understand what BenMAP does, or the data BenMAP relies on. In the six instances in which he answered “I don’t know,” he was not admitting lack of knowledge of the program or underlying data, like the stricken

experts in the cases Defendants cite. He was either admitting lack of knowledge about data points he did not need or were unnecessary for his purposes, or lack of knowledge on some relatively small aspect of the data or the program that can be used to attack the weight of his opinion.

Schwartz did not review the data Chinkin used for his CAMx modeling. (ECF No. 166-2, PageID.7714). That is not, however, a basis to conclude that Schwartz’s BenMAP opinions are unreliable. This was not a case in which the

expert reviewed no underlying data used to form his opinion. That is what distinguishes this case from the ones Defendants cite.6 In Orthofix Inc. v.

6 Interestingly, Defendants cite a different case in support of their expert, Mr. Leahy, addressed below, that cuts against their position here. The case is In re Keurig Green Mountain Single-Serve Coffee Antitrust Litig., 2025 WL 354671, at *48 (S.D.N.Y. Jan. 30, 2025). There, the court suggested that it would not be problematic for an expert to rely on work from another expert without overseeing the other expert’s work, verifying the data behind it, or ensuring that the other expert’s method was properly implemented. Id. at n. 36. So in line with Defendants’ reliance on that case, it is not inherently problematic that Dr. Schwartz did not oversee Chinkin’s CAMx modeling, did not verify Chinkin’s data, and did not ensure that Chinkin’s method was properly implemented. 47 Lemanski, the certified public accountant expert relied entirely on spreadsheets summarizing Orthofix’s sales average to reach conclusions on damages. 2015 WL

129990115, at *1 (E.D. Mich. Sept. 29, 2015). The expert never reviewed the data underlying the spreadsheet, and declined to review underlying documents after an error in the spreadsheet was revealed. Id. at *4. Since the court could not have

confidence that the opinion was based on sufficient facts or data, the opinion was stricken. Id. at *5. In Ask Chemicals, LP v. Computer Packages, Inc., 593 F. App’x 506, 510 (6th Cir. 2014), the “wholesale adoption of Plaintiff’s estimates, without revealing or apparently even evaluating the bases for those estimates, goes

beyond relying on facts or data and instead cloaks unexamined assumptions in the authority of expert analysis.” So the court upheld striking the expert. Id. at 511. In both cases, the expert was provided all the information by a person connected to

a party and based their opinion solely on reviewing that information taking it at face value. Here, Schwartz did not adopt wholesale unexamined assumptions from a party. Instead, he adopted the particulate matter concentrations from Chinkin’s CAMx modeling. Chinkin’s data and methodology are fully stated in his report, so

it is not like Defendants’ cases in which the data had undergone no review before an expert’s adoption. Because Dr. Schwartz was not adopting wholesale untested

48 data, his failure to review the CAMx results does not diminish the reliability of his opinions.

The last argument is that Schwartz improperly extrapolated BenMAP results for years beyond his 2019 estimates. In doing so, Schwartz failed to account for population changes after 2019. (ECF No. 166, PageID.7585-86). By choosing to

extrapolate the 2019 information for later years, Defendants assert that the opinions about the later years are speculative. (Id. at PageID.7586). The government’s response is to say that, according to Schwartz, population changes in future years would have a de minimis impact on the number of deaths and health

impacts, so the extrapolated estimate is reliable. (ECF No. 190, PageID.14143). At the hearing, the government argued that SO and PM have a linear 2 2.5 relationship; their concentrations rise and fall together, so it is reliable to scale them together. Defendants have not persuasively shown why scaling or extrapolating 2019 results for the next three years is so unreliable that the opinions on those years must be stricken. It is not as though the data supporting the opinions is entirely

lacking a factual basis or foundation, nor does it appear that Dr. Schwartz’s method of extrapolation is untestable. Because the underlying methodology is reliable, the Court will not strike the opinions for years 2020-2022.

49 The Court is not insensitive to Defendants’ concerns, but overall their arguments address the weight to be given to Dr. Schwartz’s testimony, not the

reliability of his methods. In balancing the arguments and the Court’s gatekeeper function, the Court is also mindful that trial is proceeding before the District Judge, not a jury. “The ‘gatekeeper’ doctrine was designed to protect juries and is largely

irrelevant in the context of a bench trial.” Deal v. Hamilton Cnty. Bd. of Educ., 392 F.3d 840, 852 (6th Cir. 2004). There is no concern about the Court’s ability to parse reliable from unreliable, credible from uncredible during the trial. 2. Rebuttal Report

Defendants argue that opinions regarding deaths and health impacts from 2023 to 2028 and the related social cost in the rebuttal report should be stricken because they are untimely new opinions. (ECF No. 166, PageID.7587). They note

that the initial report was limited to health impacts from 2019 to 2022. For the first time in the rebuttal report, Dr. Schwartz opines on likely health impacts from the next six years, but in doing so he is not rebutting another expert’s report. (Id. at PageID.7588). They also contend that extrapolating to these years is inappropriate.

(Id.). The government says that it is not new for Dr. Schwartz to opine on the harm for additional years because in his initial report he said that the harm and associated economic impact will continue as long as Defendants omit excess

50 pollutant. (ECF No. 190, PageID.14146). It insists Schwartz was permissibly updating his initial opinions.

For the same reason Chinkin’s HYSPLIT opinion in his rebuttal report was stricken, Dr. Schwartz’s 2023-2028 opinions are STRICKEN. As discussed above, “[t]he purpose of rebuttal reports is not to ‘advance new arguments or new

evidence outside the scope of the opposing expert’s testimony.’” Ohio A. Philip Randolph Inst. v. Smith, 2019 WL 428371, at *2 (S.D. Ohio Feb. 4, 2019) (quoting Sinomax USA, Inc. v. Am. Signature Inc., 2022 WL 7180339, at *2 (S.D. Ohio Sept. 30, 2022)). Otherwise a party could surprise its opponent by submitting new

opinions under the guise of a “rebuttal” that its opponent could not rebut or probe through discovery. The government provided no good reason why Dr. Schwartz did not or could not include the later years in his initial report.

Thus, the motion to strike the initial report is DENIED, but it is GRANTED with respect to the additional opinions in the rebuttal report. 3. Motion to Strike Reply (ECF No. 221) Finally, the government moved to strike the portion of Defendants’ reply

brief that argues that Schwartz’s opinions should be stricken because they rely on Chinkin’s CAMx data which in turn relied in part on information from Dr. Sahu.

51 (ECF No. 221). That motion was granted, above. It is GRANTED here, too, for the same reasons given above.

D. United States’ Motion to Exclude Dr. Kelly (ECF No. 165) Dr. Kathryn Kelly is an environmental toxicologist who provided a rebuttal to the United States’ expert, Dr. Schwartz’s opinions concerning the impact on

human health from EES Coke’s emissions. The United States seeks to exclude Dr. Kelly as an expert based on a perceived lack of scientific support for Dr. Kelly’s opinions. (ECF No. 165). Essentially, the government argues that Dr. Kelly’s methodology and

conclusions rest on junk science, so they are unreliable and must be stricken. It raises a multitude of concerns about her report, but all the contentions flow from this theme: “there is over two decades of consensus in the scientific community

linking exposure of PM to these health effects. Dr. Kelly stands against this 2.5 scientific consensus.” (ECF No. 165, PageID.7293-94, 7298-7300). For the reasons more fully explained below, the Court will not strike Dr. Kelly’s opinions. The government’s arguments more appropriately attack the weight to be given to the opinions rather than the reliability and admissibility of the opinions. But for the main contention that Dr. Kelly opposes scientific consensus concerning health effects of PM , the government appears to 2.5 52 misinterpret Dr. Kelly’s report. They describe her report as asserting that there are no harmful effects from PM inhalation. As Defendants note, “[t]he scientific 2.5 community is divided on the health impacts of minor elevations in PM levels. 2.5 And that is the subject on which Dr. Kelly is opining.” (ECF No. 188, PageID.13609) (citations omitted). Dr. Kelly does not dispute that SO and PM 2 2.5 in high concentrations over a period of time could result in adverse health effects. (Id.). Given that Dr. Kelly’s report focuses on “minor” amounts of PM , it is not 2.5 as clear as the United States suggests that she contradicts scientific consensus through junk science. Though the government contends that PM is harmful at 2.5 any concentration, neither the government nor Dr. Schwartz point to scientific studies looking specifically at the effects of SO emitted in the amounts alleged 2 from EES Coke. Dr. Kelly, on the other hand, cites several sources calling into question some of the research on PM . Those sources are not peer-reviewed, a 2.5 point the government raises to show unreliability. But peer review and publication do “not necessarily correlate with reliability.” Daubert, 509 U.S. 593. That is why publication is just one of the factors courts can consider is determining reliability. Scientific methodology is another factor to be considered. Since peer review and publication are relevant but not dispositive, it is the government’s burden to 53 show why the non-peer-reviewed studies supporting Dr. Kelly, and her report itself, are based on unreliable methods. The government, however, did not address

the methodology. For instance, one source that Dr. Kelly addresses, Shifting Sands, is a review by two toxicologists of PM research literature that documents 2.5 biostatistical and methodological errors in PM research, including that conducted 2.5 by Dr. Schwartz. (ECF No. 165-5, PageID.7381). Though not peer-reviewed, this paper is based on publicly available reports and several other research papers. (ECF No. 188, PageID.13603). Dr. Kelly explains that there is no peer-reviewed study finding significant health impacts at the small levels of PM near the 2.5 facility. (ECF No. 188, PageID.13611). The government did persuade the Court that reliance on sources such as these is problematic because it has given the Court no cause to question the methodology in these studies. Defendants’ arguments against Mr. Chinkin’s and Dr. Schwartz’s opinions is somewhat helpful in support of Dr. Kelly’s opinions—the EPA established a Significant Impact Level for PM where emissions below that level do not require 2.5 additional action from the source. (ECF No. 165-5, PageID.7365-66). EES Coke’s alleged PM contribution is below that amount. While recognizing that 2.5 the Significant Impact Level is a permitting tool, not a health standard, still it is notable that the EPA is less concerned about EES Coke’s alleged emissions in the 54 permitting process. This suggests that Dr. Kelly’s opinions about health impacts from low levels of PM are perhaps not so far away from scientific consensus. 2.5 Among the more detailed arguments is that Dr. Kelly criticizes scientific studies supported by universities, organizations, or government agencies. For instance, she considers the American Lung Association as “lobbyist,” advocating

on behalf of one position. (ECF No. 165, PageID.7303-04). Her views on consensus-supporting research are not grounds to strike her report. The government asserts that Dr. Kelly was conclusion-driven when she began her work rebutting Dr. Schwartz. She did what she calls a “background

review” on EES Coke and its SO emissions in December 2023, before Dr. 2 Schwartz’s opinion was served during April 2024. Her review involved looking at the possibility of adverse effects from SO in general. (ECF No. 165, 2 PageID.7306-07). Whether Dr. Kelly impermissibly made pre-determined decisions about what occurred at EES Coke and effects of the facility’s emissions will be best assessed by the Court during Dr. Kelly’s testimony and cross- examination. The United States relies on United States v. Ameren Missouri, 421 F. Supp. 3d 729 (E.D. Mo. Sept. 30, 2019), in support of Chinkin’s and Schwartz’s reports. Defendants use Ameren to support Dr. Kelly’s report. As here, in that case the

55 United States proffered Chinkin and his CAMx conclusions and Dr. Schwartz’s BenMAP conclusions as expert reports. The defendant offered Dr. Valberg in

rebuttal to Dr. Schwartz. It appears that Dr. Valberg’s discussion and conclusions were similar to Dr. Kelly’s. The court considered Dr. Valberg’s opinions to be the “minority view” contrary to the bulk of scientific literature on PM . Id. at 776. 2.5 The court found Dr. Valberg’s opinions unpersuasive against the other opposing expert evidence. Defendants use this case to show that courts can and do consider “minority view” scientific opinions. Although the Ameren court was not considering a motion to strike Dr.

Valberg’s report, it is notable that the court considered their opinions. The United States’ quarrels with Dr. Kelly’s positions and conclusions do not address her methodology. Instead, the quarrels are of the kind best addressed

through cross-examination to persuade the Court that Dr. Kelly’s opinions carry no weight. But Dr. Kelly’s skepticism of scientific consensus or lack of peer- reviewed support do not, without more, make her an unreliable expert witness. This motion is DENIED.

E. Sierra Club’s Motion to Exclude Dr. Kelly (ECF No. 158) Sierra Club incorporated by reference the United States’ motion to exclude Dr. Kelly. Sierra Club’s motion, however, focuses on Dr. Kelly’s rebuttal to its

56 expert, Dr. Gentile. The Court struck Dr. Gentile’s report. (ECF No. 259). At the hearing, Sierra Club raised issues about Dr. Kelly, such as that her experience as a

toxicologist does not equip her to opine on public health or the Clean Air Act, and that her critique of scientific consensus is inconsistent with the EPA’s integrated scientific assessments program. There is no briefing on these arguments.

According to the EPA’s website, www.epa.gov/isa, the integrated scientific assessments program is intended to “accurately reflect the latest scientific knowledge useful in indicating the kind and extent of identifiable effects on public health and welfare which may be expected from the presence of pollutant(s) in the

ambient air.” https://perma.cc/VPL4-RJFM. The Court will not assess arguments raised for the first time at the hearing. The government’s motion to strike Dr. Kelly’s report and testimony was denied.

Sierra Club’s motion to strike her report and testimony in rebuttal to Dr. Gentile is TERMINATED AS MOOT. F. United States’ Motion to Exclude Leahy’s Opinion (ECF No. 161) Brian Leahy, a meteorologist, is Defendants’ air modeling expert who used

the CAMx program Chinkin used to estimate the concentration of PM 2.5 surrounding the EES Coke facility. He is offered to rebut Chinkin’s CAMx opinions. The United States argues that Leahy is not qualified in air modeling and 57 his opinions are unreliable because he relied on opinions from non-testifying individuals to run CAMx. (ECF No. 161).

The government argues that even if Leahy is qualified, the CAMx opinions should be stricken because they are not his opinions. Rather, they are the opinions of his colleagues Mr. Bennette and Mr. Skoglund. (ECF No. 161, PageID.5967).

It was his colleagues who analyzed and evaluated Chinkin’s CAMx modeling. Defendants say there is nothing wrong with relying on facts, data, and conclusions of colleagues or other experts, but an expert may not simply adopt another expert’s opinions “wholesale.” (ECF No. 189, PageID.13850). They contend that, because

Leahy understands the modeling inputs and independently reviewed all the data and methodologies underlying his report, his opinions are his own. (Id. at PageID.13851-52).

The Court will address the government’s second argument first. “Using staff to complete tasks and gather data relevant to the expert analysis is appropriate.” Bledsoe v. FCA US LLC, 2022 WL 4596156, at *16 (E.D. Mich. Sept. 30, 2022) (citing Chavez v. Carranza, 559 F.3d 486, 497 (6th Cir. 2009)).

But “[e]xperts are indeed not permitted to premise their opinion ‘entirely’ on other experts without undertaking any of the necessary steps to form their own opinion.” In re Flint Water Cases, 2023 WL 6147255, at *3 (E.D. Mich. Sept. 20, 2023).

58 Put another way, “expert witnesses are entitled to rely on facts, opinions and data developed or prepared by another so long as the expert in the end gives his or her

own opinion instead of simply aggregating or reciting the opinions of others.” In re Keurig Green Mountain Single-Serve Coffee Antitrust Litig., 2025 WL 354671, at *48 (S.D.N.Y. Jan. 30, 2025) (citation omitted) (emphasis added).

Defendants cite In re Keurig to support the notion that Leahy’s reliance on his colleagues was permissible. The Court reads the case as support for striking Leahy’s opinion. That court first defined the gravamen of the challenged expert’s opinion, then showed how that expert did not merely adopt another’s opinion,

rather—he used the work from other experts “as a basis for his own, separate opinion.” Id. at *49. The same is not true here. The gravamen of the challenged opinion here is that errors in Chinkin’s

CAMx methodology invalidate his opinions. For instance, the Leahy report says that Chinkin’s CAMx model simulations “did not account for the time-varying emissions at EES Coke, which inevitably must result in model prediction errors.” (ECF No. 189-2, PageID.13917). But to reach these opinions, Leahy’s colleagues

did the work to analyze Chinkin’s methods. We know this because he testified that he asked them to “evaluate the data that was input to the model, how the model was used, to identify uncertainties related to the modeling, and to review the

59 results of the modeling.” (ECF No. 189-2, PageID.3935). Leahy maintains that he had “overall” responsibility for the work and oversaw the work, but he did not do

any of the work himself. (ECF No. 189-3, PageID.5993). He couldn’t—he has never run a CAMx model. (ECF No. 161-2, PageID.5983-84). He partook in drafting the CAMx section in the expert report along with his colleagues, but could

not say exactly what he wrote versus what his colleagues wrote. (Id. at PageID.5991). There are 20 figures included in the report. Leahy created three of them—all three are historical weather maps. (Id.; ECF No. 189-2, PageID.13863). This makes sense considering Leahy is a meteorologist. The remaining figures,

including the CAMx-specific figures, are either taken from Chinkin’s report or created by Leahy’s colleagues. Comparing this to the expert in In re Keurig, it appears more strongly that

Leahy aggregated or recited the opinions of his colleagues in a report he drafted with them. This amounts to impermissible reliance on another non-disclosed expert’s work or adoption of another non-disclosed expert’s opinion. Even if the Court found that the CAMx opinion is Leahy’s opinion, the

Court has concern about his qualification to opine on Chinkin’s use of the CAMx program. In support of its argument that Leahy is unqualified, the government asserts that he has never run the CAMx model or any photochemical grid model.

60 (ECF No. 161, PageID.5963). Instead, he has experience with AERMOD, another program Chinkin ran. The government explains that AERMOD is “simplistic.”

(Id. at PageID.5964). Unlike AERMOD, CAMx “requires particular knowledge in atmospheric chemistry and chemical reactions because it simulates chemical interactions in the atmosphere.” (Id.). Leahy is not equipped with the knowledge

to run CAMx. The government contends that he does not have the education to qualify as an expert in photochemical modeling. The only chemistry class he has taken was in high school; he does not have education in atmospheric chemistry nor has he been trained in CAMx modeling. (Id. at PageID.5965). It notes instances

during Leahy’s deposition where he incorrectly answered questions about CAMx inputs. (Id. at PageID.5965-67). Defendants insist Leahy is sufficiently qualified. He has 34 years of

experience providing air quality impact analyses and he specializes in atmospheric dispersion modeling. (ECF No. 189-2, PageID.13867, Leahy Report). His experience with CAMx, specifically, is in overseeing and reviewing the work of others at his engineering firm who run CAMx models. (ECF No. 189-3,

PageID.13927, Leahy Deposition). Defendants argue that direct experience in the precise subject matter at issue is unnecessary. Rather, a person is qualified as an expert so long as their expertise in the general field is applicable to the issues.

61 (ECF No. 189, PageID.13844-45). Thus, they argue that under Sixth Circuit precedent, Leahy is qualified to offer CAMx opinions even though he has never

personally run the program. Next, they contend that Leahy’s answers to the three questions noted in the government’s brief are not reason to exclude him. First, even if his answers to those questions were incorrect, his expertise in the general

field is sufficient. Second, his response to the question about what “normalized mean bias” is irrelevant because he was not opining on normalized mean bias figures. (Id. at PageID.13848-49). As Defendants note, “[a] party’s expert need not be a ‘blue ribbon

practitioner’ or even have direct experience with the precise subject matter at issue” in order to testify on that issue.” Cap. Mortg. Sols., LLC v. Cincinnati Ins. Co., 2022 WL 16920409, at *4 (E.D. Mich. Nov. 14, 2022) (quoting Amber

Reineck House v. City of Howell, 2021 WL 6881861, at *15 (E.D. Mich. Sept. 29, 2021)). Here, however, Leahy has no experience with the subject matter at issue— CAMx modeling. Leahy’s experience in meteorology and general exposure to CAMx through

his colleagues who run the program certainly give him more knowledge on CAMx than the average person. If he were called to testify about meteorological modeling or atmospheric concentrations of particulate matter generally, including

62 the existence of CAMx, he may be qualified. But the opinions at issue directly attack another person’s use of the model which he has no direct experience in. It

would be like calling an outfielder in professional baseball to testify about pitching techniques. The outfielder knows a lot about pitching and could speak generally to some techniques. But if we want to know the details about finger positioning

during the throw, arm positioning during the throw, and the like, the outfielder is not the qualified person, the pitcher is. See also Berry v. City of Detroit, 25 F.3d 1342, 1350 (6th Cir. 1994). Given the foregoing, the motion to exclude Leahy’s expert report and

testimony as it relates to the CAMx model is GRANTED. G. Defendants’ Renewed Motion to Seal (ECF No. 269) Defendants moved to seal various documents and briefs earlier in the case.

The Court denied those motions without prejudice because they did not seek a narrowly tailored seal in line with Shane Grp., Inc. v. Blue Cross Blue Shield of Mich., 825 F.3d 299, 305 (6th Cir. 2016) and Grae v. Corrections Corp. of Am., -- F.4th --, 2025 WL 1132413 (6th Cir. Apr. 17, 2025). (ECF No. 241). Defendants

are back before the Court with much more detailed support and far more narrowly tailored requests to seal portions of documents. Many of the documents they no longer seek to file under seal. (ECF No. 269; ECF No. 269-2).

63 The Court thoroughly explained the standards relevant to motions to seal at ECF No. 241 and will not repeat them here.

The motion is GRANTED for the reasons given in the motion. Defendants have drastically reduced the number of documents to be sealed and the amount of text to be redacted in the public filings. They adequately justified the reasons for

sealing or redacting those documents. For instance, they seek to redact settlement communications or references to them, and personal telephone numbers. (ECF No. 269-2). Such sensitive and privileged information will not be given to the public. (ECF No. 269, PageID.17727-29).

They also seek to keep under seal some confidential business and financial information. For these, Defendants have sufficiently established that disclosing the full content of these documents would cause competitive harm to one or more of

the parties. (Id. at PageID.17729-32). They seek to redact only those portions necessary to prevent competitors from gaining a commercial advantage. The last category of documents addressed in the motion contain private information of third parties wholly unnecessary to the merits of this litigation. (Id.

at PageID.17732-35). Sierra Club opposes redactions in five of the documents included in Defendants’ motion: (1) Defendants’ response in opposition to Sierra Club’s

64 motion for summary judgment, (2) Mr. Smith’s declaration, (3) a document produced in discovery, (4) amended answers and objections to the government’s

first set of interrogatories, and (5) excerpts of Mr. Patterson’s deposition. (ECF No. 276). Its main concern (the only document with specific argument) is the opposition brief. The Court reviewed the rather limited redaction in the brief that

is on the last two pages. (ECF No. 269-27, ECF No. 270-24). These portions address a party’s financial viability and market share. It is not difficult to see that revealing this information to the public and the party’s competitors could harm the party. Sierra Club’s contention that the redactions are improper lacks merit.

IT IS SO ORDERED. The parties here may object to and seek review of this Order, but are required to file any objections within 14 days of service as provided for in Federal

Rule of Civil Procedure 72(a) and Local Rule 72.1(d). A party may not assign as error any defect in this Order to which timely objection was not made. Fed. R. Civ. P. 72(a). Any objections are required to specify the part of the Order to which the party objects and state the basis of the objection. When an objection is filed to

a magistrate judge’s ruling on a non-dispositive motion, the ruling remains in effect unless it is stayed by the magistrate judge or a district judge. E.D. Mich. Local Rule 72.2.

65 Date: June 9, 2025 s/Curtis Ivy, Jr. Curtis Ivy, Jr. United States Magistrate Judge