Oral argument — Midwest Ozone Group v. EPA

Case No. 21-1146, Midwest Ozone Group Petitioner v. Environmental Protection Agency, and Michael S. Regan, Administrator, United States Environmental Protection Agency. Mr. Flannery for the petitioner, Ms. Coleman for the respondents. Good morning, Mr. Flannery. We will hear from you and you and your friends on the other side, and the other participants will have the historic honor of arguing before Judge Childs, our newest colleague, for the first time. Welcome. Good morning, Your Honor. May it please the Court, I'm David Flannery. I'm representing the petitioner, Midwest Ozone Group. Judge Childs, it's my honor to appear first before you today, and the rest of you. I'd like to reserve, Your Honor, three minutes for rebuttal. And let me open by indicating, of course, that this is a case that involves the so-called revised cross-state pollution rule. It is the fourth effort on the part of the Environmental Protection Agency to address the good neighbor provision, so-called good neighbor provision to the Clean Air Act. Significantly, this case is also one that grew out of the remand of the Wisconsin decision in 2019 that addressed the previous rule, the so-called cross-state air pollution rule. We'll talk to you today about our concerns about that rule, and we'll talk about our concerns that EPA made some abbreviated and manipulated decisions that we think were really designed to achieve the timing of the rule. And it did so in a way that it took some shortcuts, both technical and legal, with that, and we'll take our opening time this morning to take you through those. Out of the ones, the issues that we raised in our briefs, we'll point out this morning three of those specific points that we'd like to emphasize. Of course, we'll open to any questions that you might have about any of that as we proceed. The first point that I'd like to raise deals with the so-called harmonization provision of the Act, and the concern that EPA, when it dealt with this rule, never dealt with the mandate of this court that yet harmonized the obligations of the upwind states and the downwind states as it achieved the result that it was seeking. EPA instead used 2021, the attainment year involved, but it did not seek to reconcile the obligations of upwind and downwind states, and I'll have more for that for you in a moment. The result is it shifts the burden of compliance from the downwind states to the upwind states. We'll also, in our second point, raise with you that even though all of the rules that have been before the court on this have used air quality modeling, to be sure, a very scientific approach to that, all prior rules have used air quality modeling to address these future year issues. In this case, EPA did not do that. It defaulted to a linear interpolation rule, did so to save time, and did so even though this court in its prior decisions have said that if EPA is to do an analysis of these good neighbor provisions, the court expects it to engage in an analysis that is time-sensitive. And in fact, we see a result here that involves this linear interpolation methodology driven by a district court order of the Southern District of New York. But hadn't the EPA actually, I believe it was 2023 or 2016, used these photochemical projections at some point in time, and then even in using this interpolation for 2021, it brought over some of those estimations to come up with the new formula? And Your Honor, we'll go into some detail on that. You're absolutely right. Rather than to model 2021, which is what had been done in every other rule that had been before the court, it took modeling it already had. It happened to have modeling for 2023. It happened to have modeling for 2016. It plotted those, and it literally, Your Honor, drew a straight line to see whether the 2023 data and the 2016 data intersected 2021 on the time axis. We do not believe this was a linear. We do not believe you can extrapolate that data in a linear fashion. And I'll have more for that for you in just a moment. But you are also criticizing their method, but did you actually perform the photochemical modeling yourselves to suggest that the numbers would come different, as you indicated? We do point out that if you conducted other analysis, as EPA did, EPA, for example, looked at, and this is all about power plant emissions and how does power plant emissions affect air quality? And so if you look at what you think is a realistic picture of power plant emissions in 2021, EPA estimated those, and it estimated those as 124,000 tons. But EPA never took that independent estimate and compared that back, what it would have gotten out of that linear interpolation methodology that we think was inaccurate. And it turns out that if you make that comparison, as we do in our brief, it turns out that their methodology overpredicts those emissions by 24 percent. 24 percent at a time when the downwind monitors, the critical downwind monitors in Connecticut that drive this rule for all but one state. For 11 out of 12 states, these Connecticut monitors are pivotal in the logic on this. And it turns out those monitors are above the standard by only 4 percent and not the 24 percent. So I understand that you disagree with the EPA's chosen methodology, but what are your arguments that their methodology is arbitrary and capricious, right? EPA's method, in this case, facing statutory and court deadlines. What about their method was arbitrary and capricious? Because that's the standard we have to apply here. Correct, Your Honor, and this court has pointed out in its previous rulings in this case, in these very rules, the Appalachian Power decision in 1999 and in 2001, pointed out that, yes, there will be deference. The court will extend to EPA on these technical issues. But these technical issues have to be vetted back against realistic data. Now, and in this case, where this exists, this court has said in the past that it approves what EPA has done, but only because it has used air quality modeling, modeling the year involved. It's not done here. This court has said, and it did so in its Maryland decision in 2020 most recently, that it expects that when EPA does this analysis, that it will approach this using a time-intensive analytical process. So it is expected of the court, expected at EPA, to conduct the detailed modeling and not take the shortcut. And in this case, they admit they took the shortcut, simply because a district court order had said, we'd like you to do it more quickly. I think that focusing on the district court deadline, I mean, that overlooks that the district court's deadline is set in part because of the statutory deadlines, which are a very important part of the Clean Air Act. It's not just an arbitrary date that they picked. That takes me to my harmonization point that answers your question about the statutory deadline. This case is all about interpreting the provisions of 110A2D of the Clean Air Act, which says that an upwind state may not significantly contribute to downwind non-attainment. But it doesn't say by the next attainment date. The language says you have to determine when the upwind state makes that obligation go away by looking to see what is consistent with what is being done with Title I, the rest of the act, the non-attainment provisions of the act. So there's no specific date that's set there. This court in the Wisconsin remand said, yes, the next analytical, the next attainment date is 2021. But what is said in that opinion is you've got to reconcile what the upwind states and downwind states are doing at that 2021 date. You've got to ask the question, has EPA done something in the downwind states to delay their obligation to comply with the requirements of that attainment date? And if so, the Wisconsin court specifically said EPA may delay the responsibility for the upwind states if that's what's needed to reconcile and harmonize the two obligations. But hasn't the EPA indicated its rationale so that we have a very well-reasoned methodology as to why it chose this particular alternative? But EPA did not offer an explanation for why it didn't look at the delay and the opposition of the controls of the downwind states. But it responded to all that criticism, you know, during this whole process with respect to the rule. I don't think so, Your Honor. I think EPA never addressed the question of the delay that appeared with the downwind states that would have kept those downwind states being in non-attainment for years after the 2021 date. Even though we explained to them, and they do, that delay occurred, but they never acknowledged that that delay had to be reconciled with the burden they were putting on the upwind states. And this court specifically mandated that they do that. That's literally the ruling out of the Wisconsin case. So it's a two-step process. Yes, they have an attainment date of 2021. But when they get there, our reading of the Wisconsin agreement is that you've got to ask yourself a question. Why have the downwind states done if EPA has, in fact, approved a delay in those downwind state controls that affects the rule? And EPA has done that. And there are citations for that in the record of this, where they came back and approved those delays under the non-attainment provisions of the Act, not the good neighbor provisions. And that's the point. The good neighbor provisions are driven off the deadline that EPA sets for the downwind states coming into compliance. And EPA knew that delay had occurred, but it never took that into account when it made its decision to impose these controls on the upwind states. With those errors, and with the failure to align these two obligations with the decision to shortcut the analytical process to use this linear interpolation, rather than air quality modeling, which this court has, the only technique this court has ever sanctioned for EPA to use on it, we have a situation where, in our view, we have to turn to the question of vacating this rule. We know this court has looked at the vacature issue, and in its earlier cases has said that in order to make that decision, you have to look at the seriousness of the discrepancies that are in the rule, and you have to look at whether there are adverse consequences related to that. We've already talked about the serious errors that are here. Just a word about the consequences of vacature. If any, we don't think there would be. This rule never changes into the downwind non-attainment, so if this rule is vacated, it won't affect downwind air quality. Two, there are other mechanisms that are out there that are available to petitioners that have that problem, and for those reasons, Your Honor, I think we'll find that you'll conclude that there really is no adverse consequences to vacating that. I see I'm into my rebuttal time, and we'll pause there. All right, thank you. We'll hear from you on rebuttal, and now we'll hear from Ms. Coleman on behalf of the EPA. May it please the Court, my name is Chloe Coleman on behalf of the United States Environmental Protection Agency. With me at counsel's table are Agency Counsel Rosemary Kavan and Daniel Schramm. In light of what the Court has just heard, I'd like to use my time today to provide a little clarity, if I can, on petitioners' two primary allegations. First, that the need to complete rulemaking in 2021 prompted EPA to unreasonably alter its methodology, and second, that EPA should not have imposed upwind reductions in 2021 because emission reduction efforts in downwind states would continue past that date. Both of these challenges confuse the nature of EPA's technical and policy choices here and misapply the opinion in Wisconsin, so I'd like to briefly address them. With regard to petitioners' contention that EPA used an inferior methodology to project downwind air quality in 2021, petitioners acknowledge that EPA did not have time to build and run a full photochemical grid model for the 2021 analytic year, if it wished to complete rulemaking by the next attainment deadline. So EPA devised here an alternative methodology that could be executed in the time available. It ran photochemical grid modeling for the 2020 analytic year, which was the year for which it had a full suite of data, and then it used those projections to derive 2021 values, essentially by walking back its projections to estimate how much the modeled air quality improvement in 2023 would be actually seen by 2021. That last step to interpolate 2021 values from 2023 data, however, was a relatively small mathematical adjustment to a very detailed multi-step methodology built on both models and actual measured data. And on top of that, EPA performed two additional analyses to check its work, both of which used 2021-specific data, and both of which reached the same regulatory result here, showing that all 12 of the regulated upwind states would be linked to at least one downwind air quality problem in 2021. And this confirmed that the conclusions EPA reached through its primary methodology were well-founded, notwithstanding the use of that limited interpolation method between 2023 and 2021. And Ms. Coleman, you started out your argument about the deadlines, and then one of which was court-imposed. So tell me a little bit about how the attainment deadlines are determined and whether or not they could have postponed. So, Your Honor, the attainment deadlines are set by a statutory schedule. So in the first instance, that schedule appears in Sections 120, or pardon me, 181, which is 7511 of the Clean Air Act. And that sets a schedule that's actually key to a 1990 schedule. It was written at a time when EPA was anticipating the first round of these ozone standards. But EPA then has sort of a regulatory structure that's built on top of that that says, OK, well, we'll take that schedule, and obviously 1990 is not going to work anymore. But all of those deadlines in the 1990 schedule were cued from the moment at which a non-attainment area is sort of classified as having a problem. And can it be postponed at all? No. Well, so there is one provision. There is a provision 7511.85 that allows EPA to do limited extensions of those downwind attainment deadlines. But it's very specific and it's very narrow. It's something EPA can only do in a circumstance where essentially the downwind air quality area has shown that it's really on the knife's edge of attainment. And so, you know, if only it has just a little bit more time, rather than sort of imposing this cascade of consequences that come with being bumped up to a new classification because you failed to meet the last deadline, it gives EPA a very small, sort of narrow authority to give them just one extra year to show, you know, that they're really almost there. And so we shouldn't be triggering all those consequences. But by the court imposing the deadline, did you feel forced to this particular methodology? So I think there are a couple of things going on here, Your Honor. One is I'd like to note that that extension provision in A5 was never used here. So when petitioners talk about giving parallel extensions to New York and to upwind states, New York's downwind attainment deadline was actually never extended. They were just judged to have failed to attain the July 2021 attainment deadline because that was the deadline. So there isn't an issue here of parallel extensions. But with specific reference to your question about the deadline, you know, we have this district court proceeding that's saying, you know, looking at the situation created by the Wisconsin opinion, looking at the determination in Wisconsin that the act has to be read to require these eliminated reductions to come before the next attainment deadline. And the district court just essentially put that into practice. It asked EPA, you know, what are you doing here? How long is it going to take? You know, how do we make sure this all gets done by 2021? And I think as our brief showed, you know, ultimately that district court deadline is irrelevant in the sense that if EPA was going to complete rulemaking in time for those reductions to be in place by the 2021 deadline, it had to, you know, issue a rule in the spring regardless. So the fact that we sort of had on top of that a district court deadline specifically stating March 15, really sort of, you know, EPA was going to have to do it then anyway. So, you know, I think it's hard to see that that district court deadline was really driving things here. But it's also hard to see how, you know, petitioner's argument comes down to suggesting we should be violating, you know, two federal court orders instead of just one. So, you know, I think the sort of relevant thing here with respect to this first question on interpolation is also sort of playing out what petitioners are saying, because they've essentially said we cannot reach a reasoned result here without a full grid model for 2021. That's what we just heard counsel say. But petitioners don't dispute that modeling in that preferred manner would have pushed EPA past this 2021 deadline in contravention of Wisconsin. And so instead, they suggest that EPA shouldn't have claimed that that deadline was impossible. But this court has repeated on numerous occasions that the pursuit of methodological perfection is not a basis to delay statutory obligations, including in Wisconsin itself, where the court explained that compliance with the statutory mandate is only excused where, quote, the scientific uncertainty is so profound it precludes EPA from making a reasoned judgment. EPA's technical approach here balanced exactitude and timeliness as it was required to do to meet that deadline. And the record demonstrates that it did so reasonably. Can I just ask as a practical matter, how often does EPA fail to meet its nonattainment deadlines? Is that a common occurrence or? With respect to the nonattainment deadlines or its sort of approval actions? You know, I mean, there's a long history here of these approval actions taking a long time. Obviously, we're in a 2022 courtroom for a 2008 standard. And, you know, that reflects a lot of complexity in the statute. So, you know, the nonattainment deadlines are obviously obligations states bear rather than EPA. But as you suggest, I mean, there's a complicated process here of submissions to EPA that EPA then has to approve. And if EPA disapproves, it submits its own plans in the state's place. And so, yes, there is often delay in sort of how those rulemakings and approvals happen, obviously. Does EPA sometimes have to answer suits for mandamus for delay in this process? There are unreasonable delay suits. That's what that district court proceeding was that set the March 15th deadline. It's essentially saying, you know, EPA, you can't wait on this obligation. We have a decision in Wisconsin. You have an obligation under the good name. Right. So the agency sort of has it from both sides. Yeah, of course. And so several of these suits, as you can see in our brief, are, you know, New York v. EPA, because it's, you know, it's oftentimes a downwind state or Maryland EPA, you know, asking EPA to make sure that these obligations are being met on a schedule. You know, one of the things that creates the difficulty here, as you can imagine, and thinking about in particular the New York City attainment area, that's the focus of the briefs here, is that state may sort of start at a moderate, you know, non-attainment or moderate attainment deadline. It has an obligation to, you know, complete certain things by 2018. It's submitting plans with respect to the requirements that attach to moderate non-attainment. And then while EPA is working on, you know, those obligations, all of a sudden it's a non-attainment for moderate. So now it's serious. And so a new set of obligations attached. And so there's this sort of ladder that's constantly happening. And I think that's really at the heart of what makes petitioner's harmonization argument so difficult, because there isn't a single moment in time when EPA's obligation or EPA's obligation to look at upwind and downwind obligations as coinciding, where it says, oh, look, you know, these parties are going to do this and these parties are going to do that. It's all going to be equal and it's all going to be harmonized. Because downwind obligations, as you know, as we know here with New York, you know, if they're continuing to struggle with their quality, they're going to continue to take additional measures. So the peaker rule at issue here is New York making an effort to add additional controls because it's continuing to struggle here. But we know under the act that that doesn't actually allow upwind states to abandon their responsibility. And we know that for a number of reasons, not least of which is the court's recent opinion in the 2020 Maryland EPA case, where the court was actually looking at the question of if we have upwind, you know, in that case the marginal amount attainment date, which is the first in the schedule. But we know at the same time that downwind reductions aren't actually going to occur, you know, comparable downwind reductions are going to occur for three more years. Is that a problem of harmonization? And the court said no, that it was perfectly acceptable for upwind states to have to deal with their contribution, notwithstanding that downwind actions were going to come a little bit later. And I think at the heart of that is that the Wisconsin court was not talking about aligning upwind and downwind efforts. It was talking about aligning upwind reductions with downwind deadlines. Because the downwind deadlines are when the statute sort of triggers a bunch of consequences for downwind areas that upwind states simply don't bear. So, you know, Wisconsin was talking about this idea of, you know, should we be requiring downwind states then have to make deeper cuts to try and, you know, prevent triggering these consequences or do we just ask downwind states to bear the risk of those consequences, notwithstanding, you know, that they're due in part to inaction by others. And so the Wisconsin court ultimately said, no, we're not going to let upwind states sort of push these obligations off on downwind states. We're going to make sure that upwind states are actually answering the question of whether they bear some responsibility here and taking responsibility for that and not to the point of attainment. You know, upwind states are not required to do everything downwind states are. They're only required to address their significant contributions. And that's what the rule did here. If we do allow the EPA to use this interpolation method, but you have used the other methods in the past, are we then either encouraging or allowing you to continue to be more creative along the way? In other words, you've been set on a particular type of methodology, but now you're using something else. And part of it is based on the deadlines that you've had to deal with. But I'm just concerned about whether the court in the next case, you all come up with something else. Then how are you buying us into that process? Your Honor, I think it's important to recognize that EPA did use the photochemical grid model here. It did everything that it did for the update rule, for the cross-state rule before that. It used that same model. It did that same methodology, the four-step methodology that's been approved by the Supreme Court. So there wasn't a deviation here that says we're going to throw that in the trash and just do something really simple. EPA basically did that entire sort of stack of complicated, you know, relative response factors and all of these things to get to an answer for 2023. And then because of the timeframe, it had to walk that back for 2021. So in a circumstance where EPA had enough time to conduct that model, you know, and build all these new inventories and do all that new work for 2021, it would have done that. And so I don't think there's a concern here that EPA is just sort of jaunting off in a new direction to just simplify its process. You know, it did something very specific here to account for the deadlines in question. But, you know, these are all very complicated projection methodologies that are ultimately about, you know, trying to predict the future. And so this court has been very deferential to EPA's modeling choices, recognizing that that prediction process is, you know, very difficult. And so I don't think, you know, there would be a concern in the future if EPA were to deviate from aspects of how it does that modeling. EPA routinely sort of shifts some aspects of that, you know, which, how are we going to build this particular inventory? You know, what source are we going to use for this? There's always, I think, a little bit of movement in there. And over time, EPA has chosen different modeling tools for different elements. So it's really not as static, I think, as petitioners suggest. But in any case, you know, we think those are areas for very significant deference from this court, precisely because they are so technical and because there's so much uncertainty already built into, you know, the project here of projecting future air quality. Even given the deference, though, what's your response to the petitioner's argument that it's nonetheless arbitrary and capricious, particularly given the check that they did on this that reveals that the model is inaccurate? But, Your Honor, as I mentioned, EPA did two additional confirmatory analyses once it even got through that first set of very complicated modeling. And those confirmatory analyses, you know, checked exactly what petitioners were asking about. You know, if we assume that the amount of tons from the power plant sector might actually be higher than we projected, lower than we projected, is there a concern here? And that's what the sensitivity analysis is. It asks, you know, how sensitive is our result to some of these fluctuations? If we were wrong by 20,000 tons, would we get a different result? And EPA actually used numbers that are essentially equivalent to what petitioners used or noted in their brief. So, you know, EPA was using a sensitivity analysis that looked at 124,000 tons. I believe the petitioner's brief is citing a number of 121,000 tons. And no matter what EPA did sort of within that range, it continued to get the same regulatory result, which is 12 upwind states linked to at least one downwind non-air quality area. And so, you know, it did that work to ask that question. Can we be confident that what we've done here, you know, isn't producing something that's deviating too far in one direction or another? And it showed that it wasn't. And I would note one additional thing in that respect, Your Honor, which is the Supreme Court and EME Homer required EPA to do something called an over-control analysis. And that hasn't been an issue in this case. But what the over-control analysis actually asks is once you reduce emissions by the entire budget that we've, you know, done here, once you apply this rule and you get to a budget in this case of about 107,000 tons, the over-control analysis, it asks, you know, do we still see air quality problems? And the Supreme Court was trying to have the agency answer the question of, you know, has it done too much? But I raised it here because even under a budget of 170 or 107,000 tons or so, much lower than the number petitioners suggest has made our model arbitrary and capricious, these non-attainment and maintenance receptors, these air quality areas, still have air quality problems. So we know that there's a problem sort of no matter what that tonnage is. EPA has shown that it has very robust projections here. If the Court has no other questions, I would just close by noting that, you know, consistent with the opinion and remand in Wisconsin, EPA appropriately balanced necessary expedition and methodological precision in crafting a rule that ensured that upwind emissions significantly contributing to downwind pollution would be eliminated in time for the July 2021 deadline. EPA repeatedly checked that those regulatory conclusions were supported by the record and explained its conclusions on the record. None of petitioners' contentions overcome the deference EPA's policy and technical choices are due here, so the revised rule should be upheld. Thank you. Mr. Flannery, we'll give him his full three minutes for rebuttal. Let me begin by responding to the point that the government has made about not having time to address this issue. The two courts involved in this, this court and the district court in New York, both included their orders, ordering the March 15 deadline for EPA to file the district court and your order in the Wisconsin remand, included in both of those cases, that if additional time was needed by EPA in order to accomplish the objective here, the parties ought to approach the court again. This court had made it clear that it expects the analysis done to support this rule to be a time-intensive analysis. You understood that? Those were the words of this court. And yet, EPA never approached either court. It never approached the district court. It never approached this court to explain there's a compromise we have to do. We have a deadline. We read it as a deadline. But if we try to adhere to that deadline, then we have to quit doing the modeling that we had done and all of the other rulemaking we'd ever done for you. We never did that. On the question of the deadline involved, the government would have you believe that there was only one mechanism in the Act by which they could move that non-attainment deadline for the downwind states, that one-year extension. In fact, in this case, under Section 172, part of the non-attainment provisions of the Act, they approved an extension of the compliance obligations of the downwind states that delayed for 2027, from a 2021 attainment year, delayed until 2027, New York and Connecticut getting the benefit of emission reductions that would have brought all of those monitors into attainment. EPA did that. That's formally approved and total registers went to public notice. No appeal that I know of. That's final action. And remember, under the good neighbor provisions under 110A2D, it doesn't set a deadline there. It just says that the good neighbor provisions for upwind states have to be implemented in a way that is consistent with what EPA is doing with the downwind states. And I submit to you what they did here is not consistent with that in any way. In addition, let me mention that the check that EPA said it did with its additional modeling, that they did do other analysis. They looked at 2020 in many different ways. They got comfort out of the analysis. But as we pointed out in our brief, there's no reason to be comfortable with that because if you compare that back against what they did with their linear interpolation, it is significantly in error. Your Honor, my time is up. We encourage you to grant our petition. Thank you. Thank you, counsel on both sides. We'll take the matter under advisement.