# Oral argument — State of Texas v. EPA > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/argument%3A81966 ## Record - **Collection:** Oral argument - **Document type:** Oral argument - **Published:** July 7, 2022 ## Text Good morning, ladies and gentlemen. We have three cases to be argued today, and I would note before we begin that Judge King is participating via Zoom. This worked splendidly yesterday, and so she is able to hear you, and you can see her, and she will ask questions if she has some, so technology can be very helpful. The first case for today is 2017-60088, State of Texas v. EPA, et al. You may proceed. Thank you, Your Honor. May it please the Court, Stephen Jeddare for Luminant Generation Company. I will address Petitioner's argument that EPA's non-attainment designation for Rusk in Panola Counties was contrary to the plain language of the statute. Can you move the . . . is the microphone near you? Is that better? Yeah, just try to keep your voice up, please. Okay, thank you. I will address Petitioner's argument that EPA's non-attainment designation was contrary to the plain language of the statute. Counsel for the State of Texas will address the argument that EPA did not treat like cases alike. The starting point for the Court's review is the statutory language. A non-attainment designation is any area that does not meet the NAAQS. An unclassifiable designation is any area that cannot be classified on the basis of available information as meeting or not meeting the NAAQS. Now, this Court gave us direction on how to read these provisions in the Bexar County ozone case in 2020. The statute uses concrete terms, the Court wrote. An area is only non-attainment when it does not meet the NAAQS. And that determination, the Court said, must be based on fact and not on supposition. So the question for the Court here is whether EPA acted in 2016 based on fact or on supposition. We think the record here clearly shows EPA acted on speculation and supposition. It didn't have the facts, and it didn't do the work to earn the deference for its conclusions. Now, there's no dispute that EPA based its decision in 2016 only on Sierra Club's modeling, and that modeling had at least eight fundamental errors. That's in the appendix 1650. Basic things were missing or wrong. No buildings, wrong operational data, the wrong land surface data where it was supposed to be predicting the emissions. This is a computer program that is supposed to simulate reality, but Sierra Club's attempt got it all wrong. Could Luminate have gone through the Appendix W procedure in Section 3.2.2 to get the AECOM model? You probably have different ways of saying these acronyms than I just reading them. Approved? Well, we think we gave the government everything that it needed to go through that process. That's a process that EPA goes through. It's an internal review process. There is no role for a third party in Section 3.2.2, which is the section that they say we should have complied with. It's a provision that their regional administrator and their model clearinghouse needed to comply with. And the problem is they use that as an excuse not to get down to the details of our modeling and to reject it. They said in the record, without the Appendix W review, which is their responsibility, the validity of the model and the results for Luminate are not known. So they're the user in 3.1.2a. That's right. EPA is the user, and they didn't use the modeling correctly. That's correct. They're the ones using the model for their nonattainment designation. Now, what that section says is that the user is to follow the procedures in 3.2.2, but there's nothing in 3.2.2 for Luminate to do. The only third party referenced in there is an applicant for a PSD permit, which does need to get preapproval, but that is not the situation here. So the absence of that requirement in this situation obviously leads to the conclusion that EPA is the user. If they didn't use the correct modeling, what should happen? What would you seek? In this case, the court should vacate the nonattainment designation. It needs to reset the table, and then EPA can go about assessing the data. The problem that we have here, the reason that you need facts on the front end is so the state can do its job as the primary enforcer and setter of emission limits under the statute. Without those facts, the state can't do its job. The ball was thrown to the state in 2016, and the clock started ticking on regulatory sanctions, which are now in place as a result of the unlawful nonattainment designation. EPA claims it doesn't even have to review the state's plan, that it can issue a federal plan instead. That's at 85 FEDREG 48112. We think now that the monitoring data is in, the state has collected the data in accordance with the data requirements rule. It has used the data to develop an agreed order with Lumina for Martin Lake that is now in place. The problem is that leaving the 2016 designation in place doesn't allow that state plan to move forward with the state as the primary enforcer. Now EPA says that we were close enough and we should get deference because modeling is complicated. But EPA didn't do the work to support its conclusions. It outsourced that work to Sierra Club. The score was clear in the white line investment case last year. Reliance on expertise and experience is no substitute for reasoned decision making. We think the most egregious omission was EPA's failure to compare Sierra Club's modeling to the actual single concrete fact that was available in 2016. That was monitoring data from the Longview monitor. Do you argue that because they outsourced this to Sierra Club that they're not entitled to deference because they're not using their own technical expertise? That's right. They did not exercise that technical expertise here. It may be possible. So there's no technical expertise to defer to because all they did was use a third party to go out and gather it, and they used that without being critical of it. That's exactly right. They identified the errors in it, but then they didn't actually correct the errors. It wasn't even until 2021 that EPA claims to have gained certainty about the modeling. Of course, we'd continue to dispute that, but that's not before the court. What about the part that EPA is just stuck because it couldn't wait and it's in a time crunch? Is that your argument or is that the state's or somebody else's? EPA clearly misunderstood the law and its discretion. We think EPA was required to designate it unclassifiable, but even if there was some discretion there, the agency misunderstood its discretion. We commented in 2016 that EPA should have designated it unclassifiable in light of the conflicting available information and allowed for the collection of the data to come in and then classify it as non-attainment or attainment, which, as my colleague will point out, EPA did in other areas. But EPA said in the appendix at 1330 that it didn't have the discretion to do that. It couldn't wait. It couldn't designate it as unclassified. Is that a three-year deadline? Is that right? Or what was the deadline? It had to be done within three years. The monitoring takes place for three years. That's how long the monitoring is. So it said it couldn't wait, and if it were true that it couldn't wait, then does that make the consent decree okay? Because they're desperate and they have to do something? The consent decree, the court out in California was very clear that we're not telling EPA what it has to do. What they did in the record here was essentially use it as an excuse to say we have to pick between attainment and non-attainment. We can't admit that there's a conflicting information or the absence of information, and we can't collect the data in the future. And that is exactly what it did in other areas. Why is the consent decree in California if this is about Rusk and Panola counties? Yeah, that's a great question, Your Honor. Am I wrong? Y'all did not participate in this. My client did not. The state of Texas did challenge the consent decree in that case and said that EPA should have been required to issue the designation as of the statutory deadline and not some later deadline. Everyone concedes that as of the statutory deadline, there was no modeling at all. Did that include other counties? I'm sorry, did it include other counties? I believe there were other . . . It was regional in a broad sense. I think there were counties from all over the country in that case. So it wasn't really regional. So why was it in California if it's national? That's a great question, Your Honor. Because we have a lot of litigation about that. Do you know the answers to any of these questions? I do not. I know that this case is proper in this court and that EPA has conceded venue in this court. Thank you. Okay. Thank you. May it please the Court, Ryan Bosch on behalf of the state of Texas and the Texas Commission on Environmental Equality. My colleague has explained why EPA got its modeling analysis wrong. I'd like to present a more purely legal argument for vacater, namely that EPA failed to explain why it was crediting Sierra Club's modeling here even though that modeling contained materially identical errors that caused EPA to reject it in other counties. Counsel, that's how you divided it up. The Luminant counsel sort of handed off back to the table this issue of the consent decree. Are you prepared to address that? It seems to me that the Texas brief is arguing that certain rights of Texas and other states, I guess, were preserved for further arguments, both in the consent decree and maybe in what the Ninth Circuit said. What's your position on how much the Ninth Circuit, the consent decree in the Northern District of California, controls what the EPA had to do as to the issues before us? EPA did enter into a binding consent decree that established a timetable in California, established a timetable for when it was obligated. What was preserved for Texas and maybe other interveners in your interpretation? At a bare minimum, that consent decree did not govern what designations EPA would make for certain counties. So Texas and Luminant had full discretion to argue that an unclassifiable designation was appropriate or an attainment designation was appropriate. The consent decree simply didn't touch that issue. But you're not saying that they shouldn't have entered into it at all I think there is a question how a district court in California was able to enter a consent decree that governed the entire country, especially given the venue and jurisdictional provisions under the Clean Air Act. We're very familiar with regional versus national. We have a whole doctrine on it. The D.C. Circuit has a whole doctrine on it. I don't understand how they could do a national consent decree on this in California. I think that's certainly a fair point, Your Honor. A district court in California essentially set deadlines for EPA that would govern the entire country. We advanced the argument in those proceedings. Did you appeal all the way through the Ninth Circuit and the Supreme Court to say this shouldn't have been there or you just let it go? We did appeal to the Ninth Circuit. One of our primary arguments there was that EPA was obligated to issue its designation determinations by the statutory deadline, which was 2013, and base its designations on information that it had in 2013 because it promulgated its SO2 NAC standard in 2010, so three years later, 2013. We argued it was obligated to make the designation determinations as of its data in 2013, and as of 2013, it didn't have Sierra Cubs modeling. It had nothing. So it would have had to issue an unclassifiable designation. But you didn't argue the jurisdiction? I don't believe that we did, Your Honor. Well, then we can't. You don't preserve your arguments like that. Right, and we recognize that at this point the consent decree that it entered into in the Ninth Circuit governed the timetable that it was obligated to promulgate its designation determinations during. But you say it doesn't govern whether or not it was acceptable, the Sierra Club modeling. It certainly doesn't govern that question, Your Honor. Judge Elrod, your 2020 Bexar County opinion recognized that EPA's NACs evaluations present an opportunity for abusive discretion. Indeed, courts have repeatedly concluded that EPA has abused its discretion by failing to explain its disparate treatments of counties under the NACs, including in the Catawba County case, which was discussed in the briefing, and the D.C. Circuit's recently decided Clean Air Wisconsin case. The EPA did so again here. So EPA tellingly tries to hide behind procedural objections here, Your Honors, and those arguments are wrong for the reasons that we've explained in our briefing. But I like to focus principally on EPA's merits argument, which boils down to the idea that different counties and different power plants have different features, and so it's proper to treat them differently. But that bare truism does not save EPA here. This court recognized last year in the University of Texas case that an administrative agency cannot hide behind the fact-intensive nature of penalty adjudications to ignore irrational distinctions between like cases. It must instead explain at a minimum why specific differentiating facts are material. Distill down what the state of Texas believes was arbitrary and capricious here. Distill it down. So EPA identified errors in Sierra Club's modeling in other counties and concluded that those errors were fatal for Sierra Club's modeling in those counties and identified that the same errors were present here and it nevertheless used Sierra Club's modeling. Now, it may be possible in the abstract, Your Honors. It may be possible for EPA to say, well, these counties differ to such a great degree that the errors there were fatal, but they weren't fatal here for some reason. But EPA simply didn't do that. EPA was obligated to do that on the front end, but DOJ instead is doing it on the back end for EPA. All the discussion in DOJ's brief about disparate treatment on the merits is just DOJ's lawyers explaining why there are some differences in those counties that aren't present here. EPA has to make those decisions in the first instance. What does the state of Texas believe should happen? The state of Texas believes that vacater is appropriate, Your Honor. Vacater is the default, of course. Do they go back and do the work now? And can the EPA do the work to prove this up? I think EPA would have discretion how it acts in a post-vacater world, Your Honor. They could make an attainment or non-attainment designation based on the monitoring that is now in place. Of course, Texas relied on EPA's representation in 2010. This is in our briefing in 2010, that if there were no monitoring in place, EPA would presumably designate the county as unclassifiable. So the counties, is it the state's position that they're at all times in attainment? The state's position is that EPA should have waited for Texas' monitoring and relied on that monitoring. Right, but now that you have the monitoring. The monitoring does show exceedances of the SO2 NACs. EPA is not relying on that after-the-fact evidence to determine that it's not attainable. But they could still do this with the monitoring data after the fact. They would have discretion to do that, Your Honor, but there's a key distinction here between what's happened and what should happen. If it went back in a post-vacater world and made a non-attainment designation then, the clock would restart on penalties, on EPA's timing to issue a federal implementation plan if it was unhappy with the state's implementation plan. We'd have time to put those materials together, put together a state implementation plan, without the threat of sanctions hanging over us. Right now the threat of sanctions does hang over us because EPA relied on Sierra Club's modeling. If I may, Your Honors, I'd like to identify just one other county that I think really illustrates the point that EPA simply had no basis to rely on Sierra Club's modeling here. This is Lancaster County, Nebraska. This is at Appendix 409. EPA said the Sierra Club did not include the building dimension information and thus did not address the effects of building downwash. Without actually including downwash in the modeling, it is impossible to characterize the design value impacts from downwash for this source. Impossible. That's Appendix 409, describing Nebraska. We know essentially the same exact thing happened here because they admitted it in their proposed error correction. They said the treatment of building downwash, again specifically flagged in the error correction, was not addressed by Sierra Club's modeling. But EPA nevertheless concluded here, it didn't explain how, but it nevertheless concluded here that the modeling was okay. I think that's impermissible under basic administrative law principles, Your Honor. I recognize that my time is about to expire. I'd be happy to address any other questions. A quick question. Is EPA relying upon deference like Kaiser deference for the interpretation of the word user in your understanding of this briefing? I don't understand it to be. I didn't see that. No, I don't understand it to be relying on that deference, Your Honor. Okay. Thank you. Plain meaning. That's right, Your Honor. Okay. Can we start there? Yes, Your Honor. Is the EPA relying on plain meaning for user or are you relying on Kaiser deference? I don't think you briefed it. We think it's plain meaning, but we also believe this Court has said that to the extent that there is any ambiguity there, EPA gets deference on that. Not if you don't brief it, right? Well, we believe it's plain meaning. Okay. And it's within EPA's discretion. Let me start with the State's argument and the other counties. This Court has made very clear that EPA should not look to other jurisdictions and petitioners should not look to other jurisdictions in trying to assess the reasonableness and the rationality of the findings that EPA makes in a specific location here with the Rusk-Panola counties. The Court said in Texas v. EPA, quote, given significant differences among the counties, a direct one-to-one comparison, including the methods used to measure such data, could be inappropriate or even illogical. But that's a different thing altogether because the issue is the method was similar across the board and it was thrown out in other counties but not thrown out in this one, and there's no reasoning given for why it was thrown out here. So that sentence doesn't actually help you at all, does it? Because we're talking about whether the Sierra Club's methodology that was thrown out in other areas should also be thrown out here because it's faulty, allegedly. Sorry. The Sierra Club methodology was not thrown out at all. They used the ACOM model, which is the preferred model, which has been tested 17 times. But it has lots of errors in it. It had errors the way it was applied in those particular jurisdictions, and those were not present here, for instance. And this Court said . . . So you believe that it had no errors here and that's why it was used here. Is that the argument? EPA analyzed it very . . . scrutinized every aspect of it, went through every aspect of the way that model . . . and it's not Sierra Club's model. It's the AECOM model. They just happened to apply it. It's the AROG model. Went through every aspect . . . Did EPA outsource this to the Sierra Club, as was argued by opposing counsel? Do you dispute that? They didn't outsource it. They didn't have a contract. They asked for comments. The comments included the modeling, and Sierra Club does modeling in a number of these cases. And EPA didn't do its own modeling? EPA did not do its own modeling, but it analyzed every aspect of their modeling and it analyzed every aspect of Luminance's modeling. Did it analyze every aspect of the same types of errors that were found? Absolutely. Absolutely. EPA went through Sierra Club's modeling. First, it noted that it followed the technical assistance document and it used the default regulatory options. And it used the AROG model, which is the accepted model. It found the assumptions that it used very conservative. For instance, it used a very low background SO2 there. It looked to the lowest monitor, background monitor, in all of Texas and used that. That was a conservative assumption that they put in. It excluded nearby sources. That was another conservative assumption that they put in, which was all confirmed by later modeling. Then it looked at all of Luminance's critiques of that modeling. First of all, Luminance just assumed an unsupportable 50% reduction because they used the AROG model. There's no basis for that, and EPA examined their criticisms of it and found that fundamentally flawed. That model is used throughout the United States, but here Texas decided to just cut it by 50%. Secondly, a major flaw of Texas's criticisms is that they based it on an earlier model, an earlier analysis done by Sierra Club, which was based on allowable emissions. Allowable emissions are what's in a permit when the plant operates at 100% of the time. That's not how plants operate. The modeling that EPA relied upon by Sierra Club was based on actual emissions. That's obviously much lower than what's allowable. Texas never analyzed that, and they went through a sensitivity analysis, too. So EPA looked at all of the criticisms of Texas. For instance, at the time Sierra Club did its modeling, they simply lacked the data on building downwash and stack parameters. So EPA said, well, what is the effect of that? Well, the effect turned out to be that if that had been included, the concentration would have ended up being higher. So again, EPA and Sierra Club took a conservative view of that. Did EPA itself have gone through the Appendix W procedure in Section 3.2.2? The Appendix W procedure is a detailed procedure, and it requires the user to get approval. Right. Could EPA have gone through the Appendix W procedure? Texas had six years to submit that. The way the Appendix W procedure works is— Okay. I'm sorry. Could EPA have gone through the Appendix W procedure itself as a user under the provision, or is it the EPA's position that it could not, it was forbidden from going through the Appendix W procedure? I assume there may be circumstances where EPA looks to use a new model, but EPA was not the model user here. EPA is the reviewer of the model. EPA has a modeling committee that works with the region that looks at all these factors, that looks at peer review and analysis of the model to determine whether it's usable. And this model had serious flaws. Before you get into that, let me ask you, is there anything in the record indicating how this has been handled in the past? There certainly is a strong disagreement between the two sides here about how this modeling should have been reviewed by the clearinghouse. Is there anything in the record about how that's been done for other models? Does EPA sometimes submit a model for review, and sometimes it's an outside commenter? Is there anything in the record about that? There's nothing in the record about that because petitioners didn't raise that as a claim here. What we can go on as far as what the common practice has been. I don't find from the regulation itself much clarity. The common practice is if someone wants to submit a new model. But is that in the record? I'm not asking. We have differences of opinion on how this is supposed to work. Are you about to tell me what's in the record on how it usually works? No, I don't believe there's anything in the record laying out the procedure. In the record of this case, EPA has to do this for jurisdictions all over the United States, and it doesn't put it in the record of each of those. Oh, and by the way, if you want to use a new model, we're going to repeat what's in Appendix W and put it in there. There's an obligation to know the regulations, know the requirements, and if you want to go ahead and submit a new model to base it on. Does EPA have a duty under its guideline to pick the best model available? No. What the guidelines say is the AIRMOD model is preferred, and if you want to use a different model, you have to establish through this process that the model is at least just as good as the AIRMOD model. Well, that's what the problem is. So you say it doesn't have to pick the best one. It has to pick one that's at least as good as the others, so it needs to pick an equivalent model. Would that be accurate? Yes. Okay. Section 1.0E of the guidelines says that in all cases, the model or technique applied to a given situation should be the one that provides the most accurate representation. So that would seem to say that it should use the best model. And in this case, it was clear that the state's model, Luminance's model, was not the best. For instance, they applied a blanket 50% reduction to all SO2 concentrations. They did that based on temperatures and wind speeds, and what EPA studied that and looked at that and found there's no basis for that. So the position that it got put through the 3.2.2 procedure or not, or just got a quick look? Because if EPA says I need to pick the best model available under 1.0E of the guideline that has the most accurate representation, and then Luminance says, hey, here's a better model, why doesn't EPA then have a duty to put the guideline through the Section 3.2.2 process to see, to test it, to see if it's a better model? First of all, it had a deadline to meet. Texas had six years, six years, to submit their model for approval through this process. It didn't. It didn't even submit the model until after EPA issued its proposed nonattainment recommendation. So did you compare through the 3.2.2 process the Luminance model or not, factually? What is the record? Factually, effectively, yes. Effectively you did or you didn't? Effectively it did. It did. EPA did. It went through the model. With the 3.2.2 process or some other quick look? I don't know what you mean by a quick look. It's scientists in its analysis looked at the model carefully. With the 3.2.2 process or some other method? Not necessarily through the 3.2.2 process. But it looked at, it had its scientists look at the use of the model for this specific. And that's all in the record of how it was left? That is all in the record. As I said, at R328 pages 42 to 45 it found no basis for a blanket, a blanket 50% reduction at the start of the modeling. It found that what Texas did in running this alternative model was not use the actual emissions. The actual emissions were available. They were present. Instead they used a scaling process which misrepresented the actual emissions. And third, and probably most importantly, they used this model and put into it an automatic 30 to 40% reduction based on future actions which this court has said you cannot do. You cannot base your predictions, your analysis of compliance based on future actions. This court has said it's as of today. And that was a basis of it. EPA found other problems with the modeling. So if someone said the agency declined without explanation to go through the approval process to see if Luminance proposed models with the potential refinements could be used, would that be a true or false statement? I would say that that would be a false statement because notwithstanding the fact that it was not properly submitted, EPA reviewed it anyway. Okay. When you say we're in a deadline, do you mean the deadline of the requirement that it be done, the statutory deadline, I mean the regulatory deadline, or are you saying because of this consent decree? It's actually both. It starts obviously with a statutory deadline. EPA missed that deadline, and so the court issued a ‑‑ Why was this ‑‑ what was going on with the regional thing, including things way out of the region in a California district court? Well, as you know from the venue provision, we took the position that this was a national rule because it's a set. You took a national rule and tried to take it to D.C. and California kept it? No. Just as this court determined that this was local, there that ‑‑ Did you all argue that that was national in that case, or did you all say, okay, this is regional? No. There were various states involved in that case, including Texas. What did the EPA argue in that case about national and going to D.C. versus regional and staying in California? I don't think we took any specific position in that case. You all come pretty hard sometimes on that position, and you all didn't even take a position on that, even though it included things all over the country, in Pennsylvania and in Texas? Well, there were multiple states involved in that case, but it was filed in that court. And you all didn't even file anything to object to that? I don't believe we did, Your Honor. I don't believe we did. As to the situation in the other counties, as I said, this court has already said that you don't look to the other counties. Is that your position, that it was regional? Yes, we accept the court's position that this was regional, that these four are regional. Ultimately, Texas is here today saying you have to look to other jurisdictions, and that's the exact opposite of what they told this court in this case. They said, quote, the designations are based on the weight of evidence for each area, including available air quality monitoring and modeling, and inquiry petitions maintained as highly specific and particularized to this jurisdiction. So they told this court you cannot look to other jurisdictions, and now they're standing here today saying that's exactly what you're supposed to do. Counsel, let me ask you about something a bit different than what we've been discussing. In 2017, you can correct me on the date, new rulemaking was underway. 2019, a suggestion from that initial comment period, a need for error correction. Then you come up in 2021 with a response to all that. There's an argument by Texas, a chinnery problem, that the only defense you can make to the decision is the original information that you had, the original rationale for your designation. What's your response to that? Is there any case law, as you had mentioned, that deals with agency reconsiderations and using new evidence that comes in through notice and comment? Let me make it clear, Your Honor. We are not relying on that later evidence to defend the 2016 designation. Is that an indication you think there's a risk in relying on that? No. Legally? No. I'm not trying to trap you. No, but I want to make it absolutely clear. We are relying on the evidence that was before EPA in 2016 to support this. Now, they came in and asked for a reconsideration. And you know what? If that reconsideration had shown the opposite, that, oh, my God, we missed something, Sierra Club made a big mistake that actually has an effect, they'd be up here saying you have to consider that in analyzing whether EPA made the correct decision in 2016. They asked us to reconsider it, and we did it. It turned out that it established that EPA was absolutely correct, that any concerns it had about the data and the input that Sierra Club used turned out to be not concerns at all, and it also turned out that they used even more conservative assumptions than EPA outlined to begin with. But, again, we are not relying upon that. The 2016 determination stands on its own. EPA analyzed every aspect, every single criticism. What the 2019 process did is Sierra Club came back and said, okay, we'll take their criticisms, we'll stick them in the model, and see if it changes anything. It didn't change anything. In fact, it turned out that it's even more conservative. And as to the differences to the other counties, I see my time is up. If there are any other questions. I think we have your argument. Okay. Thank you, Your Honor. May it please the Court, Lisa Perfetto for Sierra Club. This case is straightforward. Martin Lake is the single largest emitter of sulfur dioxide in the country, and air quality surrounding the plant did not and does not meet the sulfur dioxide standard. This Court should uphold EPA's nonattainment designation for three reasons. First, EPA carefully examined all available information, applied its technical expertise, and found that the evidence conclusively demonstrated nonattainment. That determination is entitled to deference. Second, given that finding, EPA lacked discretion to delay issuing a nonattainment designation. And finally, EPA applied the same multifactor weight of the evidence test to all areas designated under the 2010 standard, reaching different designations based on the data for each area. So regarding the information available for Martin Lake, EPA reviewed Sierra Club's modeling, which applied EPA's preferred air mod modeling protocol, and found that with concentrations at least 14% above the limit and likely higher, it reliably demonstrated nonattainment, particularly given its use of conservative assumptions that EPA found would actually underestimate concentrations. Contrary to petitioner's claim, EPA also conducted a searching review of Lumen's modeling but determined that it was unreliable. It assumed hypothetical future reductions in emissions that never occurred, and it relied on unapproved and unsubstantiated alterations to EPA's preferred modeling, which EPA determined unpredictably distorted the results. And EPA also... Have you yourself gone through the Appendix W procedure? So the Appendix W procedure requires that the user of the model submit it for review. Here, obviously, the user is Lumen, and EPA is the reviewer. That's obvious, frankly. EPA could certainly be the user of the model because it's trying to use the model to determine what the right answer is. Certainly, if EPA had been kind of providing its own modeling and analysis, here... But it often will outsource, for lack of a better word, use the modeling from groups that come in and give modeling. Yes, and the distinction is that the Appendix W process, it's something that has to happen before you use the modeling in a regulatory context. And so it basically is a pre-approval process that has to go to the regional office, and that didn't happen in this case. Lumen would have to provide a kind of detailed statistical performance evaluation, which it didn't provide, and without that information, you know, it was not pre-approved. But nonetheless, as we've heard, you know, EPA did do a very searching review of the modeling, detailing the problems with its alterations, its use of future emissions, its use of unapproved beta options. And so the record does show conclusively why it wasn't reliable. Okay, so when the opposing counsel says the EPA did not perform analysis, so we can't say whether the modeling was close enough or not, is that just wrong, or did it perform the analysis that it needed to perform? What's your position on that? EPA took a close look at all of the modeling that was provided, determined that Luminance was not reliable, determined that the modeling submitted by Sierra Club that used EPA's preferred modeling protocol was reliable, and while there were alterations they would make, that would just show greater nonattainment because the particular issues were leading to a significant underestimate of actual concentrations. And, in fact, that's what we see in the reconsideration proceeding that petitioners sought there. The sensitivity modeling shows that, indeed, making those tweaks just shows a vast increase of nonattainment. And that's the distinction between what we have here and in the other counties that petitioners mentioned. Did they model the three Luminance plants or not? Did it get a complete modeling of that? EPA relied on the modeling that was submitted in the process leading up to the designation. Okay. Anything else? Because I asked you a bunch of questions. I see that I've run out of time. Okay. Thank you. I think I just heard government counsel say that EPA followed the Appendix W process. That's completely contradictory to the record. If you look at the appendix on page 1352, EPA said that the Appendix W process was not initiated or completed in the modeling of the three Texas plants. So I'm not sure what he thinks they did there. Luminance submitted — I couldn't tell if it was some kind of — you didn't need to do that because you could tell on a more quick analysis with your experts that it wasn't necessary. I think their position is there's some kind of screening protocol that they can follow and that they can say, well, we just don't have to do it that way. But what they said in the record was that without the review process — this is in Appendix 1352 — without the Appendix W process, the validity of the models and the resulting concentrations is not known. So there's no screening process in the thing. Appendix W is the screening process. What if your model is so bad that you can just look at it for 10 minutes and say, this is not going to work at all? Well, I'm sure they could call the model clearinghouse and get that taken care of. Sometimes those things can be resolved in a matter of days, as I understand it. Counsel, let me ask you this. It seems to me you have a position, and I'm wondering which one is the more reasonable one, on how to interpret the user in this circumstance. It does seem to me there's some practicality, that if there's some deadlines for notice and comment, you want to introduce a new model in your comment of what would work better, that it's awful late in the day for that to be landing on the desk of the EPA, would you have been, I mean, is your position you're prohibited to initiate that process and could not have done it earlier under your interpretation, or each? It does seem to me that it would have made a lot of sense, maybe not legally required, but it would have made a lot of sense for that to have been introduced by you to get that pre-whatever, cleared, pre-screened or something. And then it would be fully understood immediately upon arrival. I appreciate that. There is a practicality to it, and here is the practical situation that was there. We submitted everything they needed during the public comment period. There was a 30-day public comment period in March of 2016. We didn't know what they were going to do or what they were going to use. They didn't act, and they're saying we need pre-approval. We asked, this is the model we want to use. They didn't act for another nine months. You asked? We submitted our proposal in March 2016 and said this is the approach you ought to use. You weren't saying you made some preliminary contact with EPA and asked what you should do. That's not what you were saying. I think if you look at our submittal in the appendix at 685 to 690, you'll see that we're saying this is the approach that should be used. And we did that during the only public comment period that there was. It was a 30-day public comment period. They had nine months to review it. And here's the real crux of it, Your Honor. This is all post hoc. EPA never said in December 16 what lumen it didn't submit, what particular information it didn't submit, and it never said it couldn't go through the Appendix W process. It just said it didn't. And there was nothing in the rules that said before we could make that comment in March 2016 that we had to go through any particular process. I would also say that their screening analysis that they did was barely that. What they said in the record on 1292 and 1352 was that our refinements seem large, but that they didn't know without the Appendix W process. Also, we didn't just rely on future case modeling. We submitted both current and future scenarios. That's at 690 in the record. And our merits arguments are consistent with our venue arguments. You can have a determination be fact-specific and local, but EPA is still required to treat like cases alike. I would point the panel to the regional haystack decision, 829.423, where this court held that just because EPA action may have precedential value in terms of establishing the agency's practice, it does not convert each individual decision into one of nationwide scope and effect. Thank you. Thank you. This case is submitted. Thank you very much to all counsel. We appreciate your arguments today. --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/argument%3A81966. Public record. Not legal advice.